diff options
| author | Patrik Nyblom <[email protected]> | 2010-12-03 15:38:14 +0100 |
|---|---|---|
| committer | Patrik Nyblom <[email protected]> | 2010-12-03 15:38:14 +0100 |
| commit | 73b53fe24ebc595a81c4ee023bcfef5a71e59db4 (patch) | |
| tree | ed75b935318dbef275d62ab141a5f59b5b36bce0 /erts | |
| parent | 52b0fbbc4daa4e22fcad6fe39dd8f8cec8bb8d1b (diff) | |
| parent | 159d79b6657e5bb4f3accb6a0d74b74f06da4ba2 (diff) | |
| download | otp-73b53fe24ebc595a81c4ee023bcfef5a71e59db4.tar.gz otp-73b53fe24ebc595a81c4ee023bcfef5a71e59db4.tar.bz2 otp-73b53fe24ebc595a81c4ee023bcfef5a71e59db4.zip | |
Merge branch 'pan/unicode-filenames/OTP-8887' into dev
* pan/unicode-filenames/OTP-8887: (27 commits)
Test and correct filelib and filename
Add documentation to erlang.xml and slight correction to unicode_usage.xml
Add section about Unicode file names to stdlib users guide
Correct bug in file_name_SUITE making it fail on Unix instead of Windows7
Add documentation about raw filenames and Unicode file name translation mode
Make filelib not crash on re codepoints beyond 255 in re when filename is raw
Mend on_load_embedded testcase which did not handle windows links
Correct testcase regarding windows versions supporting soft links.
Teach filelib to use re in unicode mode when filenames are not raw
Treat soft links on Windows correctly in file_name_SUITE
Adapt new soft and hard link routines on Windos to Unicode
Corrected testcases broken by unicode filenames
Update preloaded prim_file
Teach prim_file not to accept atoms and not to throw exceptions
Adapt inet_drv to Visual Studio 2008
Teach spawn_executable about Unicode
Convert filenames read on MacOSX to canonical form
Teach file to accept codepoints beyond 255.
Add testcases
Correct shell utilities to handle unicode and possibly binaries
...
Diffstat (limited to 'erts')
32 files changed, 4855 insertions, 584 deletions
diff --git a/erts/configure.in b/erts/configure.in index 8d629c25ae..762a86ef6f 100644 --- a/erts/configure.in +++ b/erts/configure.in @@ -1479,7 +1479,7 @@ AC_CHECK_HEADERS(fcntl.h limits.h unistd.h syslog.h dlfcn.h ieeefp.h \ sys/ioctl.h sys/time.h sys/uio.h \ sys/socket.h sys/sockio.h sys/socketio.h \ net/errno.h malloc.h mach-o/dyld.h arpa/nameser.h \ - pty.h util.h utmp.h langinfo.h poll.h) + pty.h util.h utmp.h langinfo.h poll.h sdkddkver.h) AC_CHECK_HEADER(sys/resource.h, [AC_DEFINE(HAVE_SYS_RESOURCE_H, 1, diff --git a/erts/doc/src/erl.xml b/erts/doc/src/erl.xml index 9224d73b6f..77bd952d41 100644 --- a/erts/doc/src/erl.xml +++ b/erts/doc/src/erl.xml @@ -550,6 +550,19 @@ <p>Force the <c>compressed</c> option on all ETS tables. Only intended for test and evaluation.</p> </item> + <tag><c><![CDATA[+fnl]]></c></tag> + <item> + <p>The VM works with file names as if they are encoded using the ISO-latin-1 encoding, disallowing Unicode characters with codepoints beyond 255. This is default on operating systems that have transparent file naming, i.e. all Unixes except MacOSX.</p> + </item> + <tag><c><![CDATA[+fnu]]></c></tag> + <item> + <p>The VM works with file names as if they are encoded using UTF-8 (or some other system specific Unicode encoding). This is the default on operating systems that enforce Unicode encoding, i.e. Windows and MacOSX.</p> + <p>By enabling Unicode file name translation on systems where this is not default, you open up to the possibility that some file names can not be interpreted by the VM and therefore will be returned to the program as raw binaries. The option is therefore considered experimental.</p> + </item> + <tag><c><![CDATA[+fna]]></c></tag> + <item> + <p>Selection between <c>+fnl</c> and <c>+fnu</c> is done based on the current locale settings in the OS, meaning that if you have set your terminal for UTF-8 encoding, the filesystem is expected to use the same encoding for filenames (use with care).</p> + </item> <tag><c><![CDATA[+hms Size]]></c></tag> <item> <p>Sets the default heap size of processes to the size diff --git a/erts/doc/src/erlang.xml b/erts/doc/src/erlang.xml index 638f7eef10..78d58a1e56 100644 --- a/erts/doc/src/erlang.xml +++ b/erts/doc/src/erlang.xml @@ -2781,14 +2781,17 @@ os_prompt%</pre> <name>open_port(PortName, PortSettings) -> port()</name> <fsummary>Open a port</fsummary> <type> - <v>PortName = {spawn, Command} | {spawn_driver, Command} | {spawn_executable, Command} | {fd, In, Out}</v> + <v>PortName = {spawn, Command} | {spawn_driver, Command} | {spawn_executable, FileName} | {fd, In, Out}</v> <v> Command = string()</v> + <v> FileName = [ FileNameChar ] | binary()</v> + <v> FileNameChar = int() (1..255 or any Unicode codepoint, see description)</v> <v> In = Out = int()</v> <v>PortSettings = [Opt]</v> - <v> Opt = {packet, N} | stream | {line, L} | {cd, Dir} | {env, Env} | {args, [ string() ]} | {arg0, string()} | exit_status | use_stdio | nouse_stdio | stderr_to_stdout | in | out | binary | eof</v> + <v> Opt = {packet, N} | stream | {line, L} | {cd, Dir} | {env, Env} | {args, [ ArgString ]} | {arg0, ArgString} | exit_status | use_stdio | nouse_stdio | stderr_to_stdout | in | out | binary | eof</v> <v> N = 1 | 2 | 4</v> <v> L = int()</v> <v> Dir = string()</v> + <v> ArgString = [ FileNameChar ] | binary()</v> <v> Env = [{Name, Val}]</v> <v> Name = string()</v> <v> Val = string() | false</v> @@ -2851,7 +2854,26 @@ os_prompt%</pre> executed, the appropriate command interpreter will implicitly be invoked, but there will still be no command argument expansion or implicit PATH search.</p> - + + <p>The name of the executable as well as the arguments + given in <c>args</c> and <c>arg0</c> is subject to + Unicode file name translation if the system is running + in Unicode file name mode. To avoid + translation or force i.e. UTF-8, supply the executable + and/or arguments as a binary in the correct + encoding. See the <seealso + marker="kernel:file">file</seealso> module, the + <seealso marker="kernel:file#native_name_encoding/0"> + file:native_name_encoding/0</seealso> function and the + <seealso marker="stdlib:unicode_usage">stdlib users guide + </seealso> for details.</p> + + <note>The characters in the name (if given as a list) + can only be > 255 if the Erlang VM is started in + Unicode file name translation mode, otherwise the name + of the executable is limited to the ISO-latin-1 + character set.</note> + <p>If the <c>Command</c> cannot be run, an error exception, with the posix error code as the reason, is raised. The error reason may differ between operating @@ -2954,6 +2976,21 @@ os_prompt%</pre> should not be given in this list. The proper executable name will automatically be used as argv[0] where applicable.</p> + <p>When the Erlang VM is running in Unicode file name + mode, the arguments can contain any Unicode characters and + will be translated into whatever is appropriate on the + underlying OS, which means UTF-8 for all platforms except + Windows, which has other (more transparent) ways of + dealing with Unicode arguments to programs. To avoid + Unicode translation of arguments, they can be supplied as + binaries in whatever encoding is deemed appropriate.</p> + + <note>The characters in the arguments (if given as a + list of characters) can only be > 255 if the Erlang + VM is started in Unicode file name mode, + otherwise the arguments are limited to the + ISO-latin-1 character set.</note> + <p>If one, for any reason, wants to explicitly set the program name in the argument vector, the <c>arg0</c> option can be used.</p> @@ -2969,6 +3006,9 @@ os_prompt%</pre> responds to this is highly system dependent and no specific effect is guaranteed.</p> + <p>The unicode file name translation rules of the + <c>args</c> option apply to this option as well.</p> + </item> <tag><c>exit_status</c></tag> diff --git a/erts/emulator/Makefile.in b/erts/emulator/Makefile.in index 4ed0ccabc6..6c33e2ca16 100644 --- a/erts/emulator/Makefile.in +++ b/erts/emulator/Makefile.in @@ -796,7 +796,8 @@ endif OS_OBJS += $(OBJDIR)/erl_mseg.o \ $(OBJDIR)/erl_$(ERLANG_OSTYPE)_sys_ddll.o \ - $(OBJDIR)/erl_mtrace_sys_wrap.o + $(OBJDIR)/erl_mtrace_sys_wrap.o \ + $(OBJDIR)/erl_sys_common_misc.o HIPE_x86_OS_OBJS=$(HIPE_x86_$(OPSYS)_OBJS) HIPE_x86_OBJS=$(OBJDIR)/hipe_x86.o $(OBJDIR)/hipe_x86_glue.o $(OBJDIR)/hipe_x86_bifs.o $(OBJDIR)/hipe_x86_signal.o $(OBJDIR)/hipe_x86_stack.o $(HIPE_x86_OS_OBJS) diff --git a/erts/emulator/beam/bif.tab b/erts/emulator/beam/bif.tab index 0674aae77f..60b4b1946b 100644 --- a/erts/emulator/beam/bif.tab +++ b/erts/emulator/beam/bif.tab @@ -795,6 +795,13 @@ bif erlang:nif_error/1 bif erlang:nif_error/2 # +# Helpers for unicode filenames +# +bif prim_file:internal_name2native/1 +bif prim_file:internal_native2name/1 +bif prim_file:internal_normalize_utf8/1 +bif file:native_name_encoding/0 +# # Obsolete # diff --git a/erts/emulator/beam/binary.c b/erts/emulator/beam/binary.c index 8ee8fbcb29..4be869f269 100644 --- a/erts/emulator/beam/binary.c +++ b/erts/emulator/beam/binary.c @@ -217,8 +217,8 @@ erts_get_aligned_binary_bytes_extra(Eterm bin, byte** base_ptr, ErtsAlcType_t al return bytes; } -static Eterm -bin_bytes_to_list(Eterm previous, Eterm* hp, byte* bytes, Uint size, Uint bitoffs) +Eterm +erts_bin_bytes_to_list(Eterm previous, Eterm* hp, byte* bytes, Uint size, Uint bitoffs) { if (bitoffs == 0) { while (size) { @@ -263,7 +263,7 @@ BIF_RETTYPE binary_to_list_1(BIF_ALIST_1) Eterm* hp = HAlloc(BIF_P, 2 * size); byte* bytes = binary_bytes(real_bin)+offset; - BIF_RET(bin_bytes_to_list(NIL, hp, bytes, size, bitoffs)); + BIF_RET(erts_bin_bytes_to_list(NIL, hp, bytes, size, bitoffs)); } error: @@ -295,7 +295,7 @@ BIF_RETTYPE binary_to_list_3(BIF_ALIST_3) } i = stop-start+1; hp = HAlloc(BIF_P, 2*i); - BIF_RET(bin_bytes_to_list(NIL, hp, bytes+start-1, i, bitoffs)); + BIF_RET(erts_bin_bytes_to_list(NIL, hp, bytes+start-1, i, bitoffs)); error: BIF_ERROR(BIF_P, BADARG); @@ -339,7 +339,7 @@ BIF_RETTYPE bitstring_to_list_1(BIF_ALIST_1) previous = CONS(hp, make_binary(last), previous); hp += 2; } - BIF_RET(bin_bytes_to_list(previous, hp, bytes, size, bitoffs)); + BIF_RET(erts_bin_bytes_to_list(previous, hp, bytes, size, bitoffs)); } diff --git a/erts/emulator/beam/erl_bif_binary.c b/erts/emulator/beam/erl_bif_binary.c index b6a445c55c..684fa5d12f 100644 --- a/erts/emulator/beam/erl_bif_binary.c +++ b/erts/emulator/beam/erl_bif_binary.c @@ -1477,7 +1477,7 @@ BIF_RETTYPE binary_matches_3(BIF_ALIST_3) goto badarg; } if (hsend == 0) { - BIF_RET(am_nomatch); + BIF_RET(NIL); } if (is_tuple(BIF_ARG_2)) { tp = tuple_val(BIF_ARG_2); diff --git a/erts/emulator/beam/erl_bif_port.c b/erts/emulator/beam/erl_bif_port.c index 378c5e73fd..fbc92b9730 100644 --- a/erts/emulator/beam/erl_bif_port.c +++ b/erts/emulator/beam/erl_bif_port.c @@ -610,6 +610,7 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) int binary_io; int soft_eof; Sint linebuf; + Eterm edir = NIL; byte dir[MAXPATHLEN]; /* These are the defaults */ @@ -686,19 +687,10 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) } else if (option == am_arg0) { char *a0; - int n; - if (is_nil(*tp)) { - n = 0; - } else if( (n = is_string(*tp)) == 0) { + + if ((a0 = erts_convert_filename_to_native(*tp, ERTS_ALC_T_TMP, 1)) == NULL) { goto badarg; } - a0 = (char *) erts_alloc(ERTS_ALC_T_TMP, - (n + 1) * sizeof(byte)); - if (intlist_to_buf(*tp, a0, n) != n) { - erl_exit(1, "%s:%d: Internal error\n", - __FILE__, __LINE__); - } - a0[n] = '\0'; if (opts.argv == NULL) { opts.argv = erts_alloc(ERTS_ALC_T_TMP, 2 * sizeof(char **)); @@ -711,22 +703,7 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) opts.argv[0] = a0; } } else if (option == am_cd) { - Eterm iolist; - DeclareTmpHeap(heap,4,p); - int r; - - UseTmpHeap(4,p); - heap[0] = *tp; - heap[1] = make_list(heap+2); - heap[2] = make_small(0); - heap[3] = NIL; - iolist = make_list(heap); - r = io_list_to_buf(iolist, (char*) dir, MAXPATHLEN); - UnUseTmpHeap(4,p); - if (r < 0) { - goto badarg; - } - opts.wd = (char *) dir; + edir = *tp; } else { goto badarg; } @@ -838,19 +815,7 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) goto badarg; } name = tp[1]; - if (is_atom(name)) { - name_buf = (char *) erts_alloc(ERTS_ALC_T_TMP, - atom_tab(atom_val(name))->len+1); - sys_memcpy((void *) name_buf, - (void *) atom_tab(atom_val(name))->name, - atom_tab(atom_val(name))->len); - name_buf[atom_tab(atom_val(name))->len] = '\0'; - } else if ((i = is_string(name))) { - name_buf = (char *) erts_alloc(ERTS_ALC_T_TMP, i + 1); - if (intlist_to_buf(name, name_buf, i) != i) - erl_exit(1, "%s:%d: Internal error\n", __FILE__, __LINE__); - name_buf[i] = '\0'; - } else { + if ((name_buf = erts_convert_filename_to_native(name,ERTS_ALC_T_TMP,0)) == NULL) { goto badarg; } opts.spawn_type = ERTS_SPAWN_EXECUTABLE; @@ -892,7 +857,33 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) /* Argument vector only if explicit spawn_executable */ goto badarg; } - + + if (edir != NIL) { + /* A working directory is expressed differently if spawn_executable, i.e. Unicode is handles + for spawn_executable... */ + if (opts.spawn_type != ERTS_SPAWN_EXECUTABLE) { + Eterm iolist; + DeclareTmpHeap(heap,4,p); + int r; + + UseTmpHeap(4,p); + heap[0] = edir; + heap[1] = make_list(heap+2); + heap[2] = make_small(0); + heap[3] = NIL; + iolist = make_list(heap); + r = io_list_to_buf(iolist, (char*) dir, MAXPATHLEN); + UnUseTmpHeap(4,p); + if (r < 0) { + goto badarg; + } + opts.wd = (char *) dir; + } else { + if ((opts.wd = erts_convert_filename_to_native(edir,ERTS_ALC_T_TMP,0)) == NULL) { + goto badarg; + } + } + } if (driver != &spawn_driver && opts.exit_status) { goto badarg; @@ -941,6 +932,9 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) if (opts.argv) { free_args(opts.argv); } + if (opts.wd && opts.wd != ((char *)dir)) { + erts_free(ERTS_ALC_T_TMP, (void *) opts.wd); + } return port_num; badarg: @@ -950,6 +944,7 @@ open_port(Process* p, Eterm name, Eterm settings, int *err_nump) #undef OPEN_PORT_ERROR } +/* Arguments can be given i unicode and as raw binaries, convert filename is used to convert */ static char **convert_args(Eterm l) { char **pp; @@ -966,22 +961,14 @@ static char **convert_args(Eterm l) pp[i++] = erts_default_arg0; while (is_list(l)) { str = CAR(list_val(l)); - - if (is_nil(str)) { - n = 0; - } else if( (n = is_string(str)) == 0) { - /* Not a string... */ + if ((b = erts_convert_filename_to_native(str,ERTS_ALC_T_TMP,1)) == NULL) { int j; for (j = 1; j < i; ++j) erts_free(ERTS_ALC_T_TMP, pp[j]); erts_free(ERTS_ALC_T_TMP, pp); return NULL; - } - b = (char *) erts_alloc(ERTS_ALC_T_TMP, (n + 1) * sizeof(byte)); - pp[i++] = (char *) b; - if (intlist_to_buf(str, b, n) != n) - erl_exit(1, "%s:%d: Internal error\n", __FILE__, __LINE__); - b[n] = '\0'; + } + pp[i++] = b; l = CDR(list_val(l)); } pp[i] = NULL; diff --git a/erts/emulator/beam/erl_binary.h b/erts/emulator/beam/erl_binary.h index a569fe2e85..bdf0fe23fc 100644 --- a/erts/emulator/beam/erl_binary.h +++ b/erts/emulator/beam/erl_binary.h @@ -152,6 +152,8 @@ do { \ void erts_init_binary(void); byte* erts_get_aligned_binary_bytes_extra(Eterm, byte**, ErtsAlcType_t, unsigned extra); +/* Used by unicode module */ +Eterm erts_bin_bytes_to_list(Eterm previous, Eterm* hp, byte* bytes, Uint size, Uint bitoffs); /* * Common implementation for erlang:list_to_binary/1 and binary:list_to_bin/1 diff --git a/erts/emulator/beam/erl_init.c b/erts/emulator/beam/erl_init.c index a9f4f041ac..464ee750f7 100644 --- a/erts/emulator/beam/erl_init.c +++ b/erts/emulator/beam/erl_init.c @@ -251,6 +251,7 @@ erl_init(int ncpu) erts_init_monitors(); erts_init_gc(); init_time(); + erts_init_sys_common_misc(); erts_init_process(ncpu); erts_init_scheduling(use_multi_run_queue, no_schedulers, @@ -907,7 +908,27 @@ erl_start(int argc, char **argv) VERBOSE(DEBUG_SYSTEM, ("using display items %d\n",display_items)); break; - + case 'f': + if (!strncmp(argv[i],"-fn",3)) { + arg = get_arg(argv[i]+3, argv[i+1], &i); + switch (*arg) { + case 'u': + erts_set_user_requested_filename_encoding(ERL_FILENAME_UTF8); + break; + case 'l': + erts_set_user_requested_filename_encoding(ERL_FILENAME_LATIN1); + break; + case 'a': + erts_set_user_requested_filename_encoding(ERL_FILENAME_UNKNOWN); + default: + erts_fprintf(stderr, "bad filename encoding %s, can be (l,u or a)\n", arg); + erts_usage(); + } + break; + } else { + erts_fprintf(stderr, "%s unknown flag %s\n", argv[0], argv[i]); + erts_usage(); + } case 'l': display_loads++; break; diff --git a/erts/emulator/beam/erl_unicode.c b/erts/emulator/beam/erl_unicode.c index d01a3661f9..72207df621 100644 --- a/erts/emulator/beam/erl_unicode.c +++ b/erts/emulator/beam/erl_unicode.c @@ -30,6 +30,8 @@ #include "big.h" #include "erl_unicode.h" +#include "erl_unicode_normalize.h" + typedef struct _restart_context { byte *bytes; @@ -54,13 +56,6 @@ static BIF_RETTYPE finalize_list_to_list(Process *p, Uint num_resulting_chars, int state, int left, Eterm tail); -static int analyze_utf8(byte *source, Uint size, - byte **err_pos, Uint *num_chars, int *left); -#define UTF8_OK 0 -#define UTF8_INCOMPLETE 1 -#define UTF8_ERROR 2 -#define UTF8_ANALYZE_MORE 3 - static BIF_RETTYPE characters_to_utf8_trap(BIF_ALIST_3); static BIF_RETTYPE characters_to_list_trap_1(BIF_ALIST_3); static BIF_RETTYPE characters_to_list_trap_2(BIF_ALIST_3); @@ -463,7 +458,7 @@ L_Again: /* Restart with sublist, old listend was pushed on stack */ } objp = list_val(ioterm); obj = CAR(objp); - if (!is_byte(obj)) + if (!is_small(obj)) break; } } else if (is_nil(obj)) { @@ -970,11 +965,11 @@ static int is_valid_utf8(Eterm orig_bin) bytes = erts_get_aligned_binary_bytes(orig_bin, &temp_alloc); } size = binary_size(orig_bin); - ret = analyze_utf8(bytes, + ret = erts_analyze_utf8(bytes, size, &endpos,&numchar,NULL); erts_free_aligned_binary_bytes(temp_alloc); - return (ret == UTF8_OK); + return (ret == ERTS_UTF8_OK); } BIF_RETTYPE unicode_characters_to_binary_2(BIF_ALIST_2) @@ -1084,14 +1079,14 @@ static BIF_RETTYPE build_list_return(Process *p, byte *bytes, int pos, Uint char hp += 2; rest_term = CONS(hp,leftover_bin,rest_term); } - BIF_RET(finalize_list_to_list(p, bytes, rest_term, 0U, pos, characters, UTF8_ERROR, left, NIL)); + BIF_RET(finalize_list_to_list(p, bytes, rest_term, 0U, pos, characters, ERTS_UTF8_ERROR, left, NIL)); } else if (rest_term == NIL && num_leftovers != 0) { Eterm leftover_bin = new_binary(p, leftover, num_leftovers); if (check_leftovers(leftover,num_leftovers) != 0) { - BIF_RET(finalize_list_to_list(p, bytes, leftover_bin, 0U, pos, characters, UTF8_ERROR, + BIF_RET(finalize_list_to_list(p, bytes, leftover_bin, 0U, pos, characters, ERTS_UTF8_ERROR, left, NIL)); } else { - BIF_RET(finalize_list_to_list(p, bytes, leftover_bin, 0U, pos, characters, UTF8_INCOMPLETE, + BIF_RET(finalize_list_to_list(p, bytes, leftover_bin, 0U, pos, characters, ERTS_UTF8_INCOMPLETE, left, NIL)); } } else { /* All OK */ @@ -1107,11 +1102,11 @@ static BIF_RETTYPE build_list_return(Process *p, byte *bytes, int pos, Uint char rc.num_processed_bytes = 0; /* not used */ rc.num_bytes_to_process = pos; rc.num_resulting_chars = characters; - rc.state = UTF8_OK; /* not used */ + rc.state = ERTS_UTF8_OK; /* not used */ BIF_TRAP3(&characters_to_list_trap_1_exp, p, make_magic_bin_for_restart(p,&rc), rest_term, latin1); } else { /* Success */ - BIF_RET(finalize_list_to_list(p, bytes, NIL, 0U, pos, characters, UTF8_OK, left, NIL)); + BIF_RET(finalize_list_to_list(p, bytes, NIL, 0U, pos, characters, ERTS_UTF8_OK, left, NIL)); } } } @@ -1205,7 +1200,7 @@ BIF_RETTYPE unicode_characters_to_list_2(BIF_ALIST_2) * When input to characters_to_list is a plain binary and the format is 'unicode', we do * a faster analyze and size count with this function. */ -static int analyze_utf8(byte *source, Uint size, +int erts_analyze_utf8(byte *source, Uint size, byte **err_pos, Uint *num_chars, int *left) { *err_pos = source; @@ -1216,60 +1211,60 @@ static int analyze_utf8(byte *source, Uint size, --size; } else if (((*source) & ((byte) 0xE0)) == 0xC0) { if (size < 2) { - return UTF8_INCOMPLETE; + return ERTS_UTF8_INCOMPLETE; } if (((source[1] & ((byte) 0xC0)) != 0x80) || ((*source) < 0xC2) /* overlong */) { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } source += 2; size -= 2; } else if (((*source) & ((byte) 0xF0)) == 0xE0) { if (size < 3) { - return UTF8_INCOMPLETE; + return ERTS_UTF8_INCOMPLETE; } if (((source[1] & ((byte) 0xC0)) != 0x80) || ((source[2] & ((byte) 0xC0)) != 0x80) || (((*source) == 0xE0) && (source[1] < 0xA0)) /* overlong */ ) { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } if ((((*source) & ((byte) 0xF)) == 0xD) && ((source[1] & 0x20) != 0)) { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } if (((*source) == 0xEF) && (source[1] == 0xBF) && ((source[2] == 0xBE) || (source[2] == 0xBF))) { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } source += 3; size -= 3; } else if (((*source) & ((byte) 0xF8)) == 0xF0) { if (size < 4) { - return UTF8_INCOMPLETE; + return ERTS_UTF8_INCOMPLETE; } if (((source[1] & ((byte) 0xC0)) != 0x80) || ((source[2] & ((byte) 0xC0)) != 0x80) || ((source[3] & ((byte) 0xC0)) != 0x80) || (((*source) == 0xF0) && (source[1] < 0x90)) /* overlong */) { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } if ((((*source) & ((byte)0x7)) > 0x4U) || ((((*source) & ((byte)0x7)) == 0x4U) && ((source[1] & ((byte)0x3F)) > 0xFU))) { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } source += 4; size -= 4; } else { - return UTF8_ERROR; + return ERTS_UTF8_ERROR; } ++(*num_chars); *err_pos = source; if (left && --(*left) <= 0) { - return UTF8_ANALYZE_MORE; + return ERTS_UTF8_ANALYZE_MORE; } } - return UTF8_OK; + return ERTS_UTF8_OK; } /* @@ -1304,7 +1299,7 @@ static Eterm do_utf8_to_list(Process *p, Uint num, byte *bytes, Uint sz, } else if (((*source) & ((byte) 0xE0)) == 0xC0) { unipoint = (((Uint) ((*source) & ((byte) 0x1F))) << 6) | - ((Uint) (source[1] & ((byte) 0x3F))); + ((Uint) (source[1] & ((byte) 0x3F))); } else if (((*source) & ((byte) 0xF0)) == 0xE0) { unipoint = (((Uint) ((*source) & ((byte) 0xF))) << 12) | @@ -1330,6 +1325,216 @@ static Eterm do_utf8_to_list(Process *p, Uint num, byte *bytes, Uint sz, return ret; } +static int is_candidate(Uint cp) +{ + int index,pos; + if (cp < 768) return 0; + if (cp > 4023) { + if (cp == 12441 || cp == 12442) return 1; + return 0; + } + index = cp / 32 - COMP_CANDIDATE_MAP_OFFSET; + pos = cp % 32; + return !!(comp_candidate_map[index] & (1UL << pos)); +} + +static int hashsearch(int *htab, int htab_size, CompEntry *cv, Uint16 c) +{ + int bucket = c % htab_size; + while (htab[bucket] != -1 && cv[htab[bucket]].c != c) + bucket = (bucket + 1) % htab_size; + return htab[bucket]; +} + +#define TRANSLATE_NO 0 +#define TRANSLATE_MAYBE -1 + +/* The s array is reversed */ +static int translate(Uint16 *s, int slen, Uint16 *res) +{ + /* Go backwards through buffer and match against tree */ + int pos = 0; + CompEntry *cv = compose_tab; + int *hc = hash_compose_tab; + int cvs = compose_tab_size; + int x; + while (pos < slen) { + x = hashsearch(hc,cvs*HASH_SIZE_FACTOR,cv,s[pos]); + if (x < 0) { + return TRANSLATE_NO; + } + if (cv[x].res) { + *res = cv[x].res; + return pos; + } + cvs = cv[x].num_subs; + hc = cv[x].hash; + cv = cv[x].subs; + ++pos; + } + return TRANSLATE_MAYBE; +} + +static void handle_first_norm(Uint16 *savepoints, int *numpointsp, Uint unipoint) +{ + /*erts_fprintf(stderr,"CP = %d, numpoints = %d\n",(int) unipoint,(int) *numpointsp);*/ + *numpointsp = 1; + savepoints[0] = (Uint16) unipoint; +} + +static void cleanup_norm(Eterm **hpp, Uint16 *savepoints, int numpoints, Eterm *retp) +{ + Eterm *hp = *hpp; + int res,i; + Uint16 newpoint; + Eterm ret = *retp; + + ret = CONS(hp,make_small((Uint) savepoints[0]),ret); + hp += 2; + + for (i = 1;i < numpoints;) { + if(!is_candidate(savepoints[i]) || + ((res = translate(savepoints+i,numpoints - i, &newpoint)) <= 0)) { + ret = CONS(hp,make_small((Uint) savepoints[i]),ret); + hp += 2; + ++i; + } else { + ret = CONS(hp,make_small((Uint) newpoint),ret); + hp += 2; + i += res; + } + } + *retp = ret; +} + +static void handle_potential_norm(Eterm **hpp, Uint16 *savepoints, int *numpointsp, Uint unipoint, Eterm *retp) +{ + Eterm *hp = *hpp; + int numpoints = *numpointsp; + int res,i; + Uint16 newpoint; + Eterm ret = *retp; + + /* erts_fprintf(stderr,"CP = %d, numpoints = %d\n",(int) unipoint,(int) numpoints);*/ + if ((unipoint >> 16) == 0) { /* otherwise we're done here */ + savepoints[numpoints++] = (Uint16) unipoint; + res = translate(savepoints,numpoints,&newpoint); + if (res == TRANSLATE_NO) { + ret = CONS(hp,make_small((Uint) savepoints[0]),ret); + hp += 2; + for (i = 1;i < numpoints;) { + if(!is_candidate(savepoints[i]) || + ((res = translate(savepoints+i,numpoints - i, &newpoint)) == 0)) { + ret = CONS(hp,make_small((Uint) savepoints[i]),ret); + hp += 2; + ++i; + } else if (res > 0) { + ret = CONS(hp,make_small((Uint) newpoint),ret); + hp += 2; + i += res; + } else { /* res < 0 */ + /* A "maybe", means we are not done yet */ + int j = 0; + while (i < numpoints) { + savepoints[j++] = savepoints[i++]; + } + numpoints = j; + goto breakaway; + } + } + numpoints = 0; + breakaway: + ; + } else if (res > 0) { + numpoints = 0; + ret = CONS(hp,make_small((Uint) newpoint),ret); + hp += 2; + } /* < 0 means go on */ + } else { + /* Unconditional rollup, this character is larger than 16 bit */ + ret = CONS(hp,make_small((Uint) savepoints[0]),ret); + hp += 2; + + for (i = 1;i < numpoints;) { + if(!is_candidate(savepoints[i]) || + ((res = translate(savepoints+i,numpoints - i, &newpoint)) <= 0)) { + ret = CONS(hp,make_small((Uint) savepoints[i]),ret); + hp += 2; + ++i; + } else { + ret = CONS(hp,make_small((Uint) newpoint),ret); + hp += 2; + i += res; + } + } + ret = CONS(hp,make_small(unipoint),ret); + hp += 2; + numpoints = 0; + } + *hpp = hp; + *numpointsp = numpoints; + *retp = ret; +} + +static Eterm do_utf8_to_list_normalize(Process *p, Uint num, byte *bytes, Uint sz) +{ + Eterm *hp,*hp_end; + Eterm ret; + byte *source; + Uint unipoint; + Uint16 savepoints[4]; + int numpoints = 0; + + ASSERT(num > 0); + + hp = HAlloc(p,num * 2); /* May be to much */ + hp_end = hp + num * 2; + ret = NIL; + source = bytes + sz; + while(--source >= bytes) { + if (((*source) & ((byte) 0x80)) == 0) { + unipoint = (Uint) *source; + } else if (((*source) & ((byte) 0xE0)) == 0xC0) { + unipoint = + (((Uint) ((*source) & ((byte) 0x1F))) << 6) | + ((Uint) (source[1] & ((byte) 0x3F))); + } else if (((*source) & ((byte) 0xF0)) == 0xE0) { + unipoint = + (((Uint) ((*source) & ((byte) 0xF))) << 12) | + (((Uint) (source[1] & ((byte) 0x3F))) << 6) | + ((Uint) (source[2] & ((byte) 0x3F))); + } else if (((*source) & ((byte) 0xF8)) == 0xF0) { + unipoint = + (((Uint) ((*source) & ((byte) 0x7))) << 18) | + (((Uint) (source[1] & ((byte) 0x3F))) << 12) | + (((Uint) (source[2] & ((byte) 0x3F))) << 6) | + ((Uint) (source[3] & ((byte) 0x3F))); + } else { + /* ignore 2#10XXXXXX */ + continue; + } + if (numpoints) { + handle_potential_norm(&hp,savepoints,&numpoints,unipoint,&ret); + continue; + } + /* We are not building up any normalizations yet, look that we shouldn't start... */ + if (is_candidate(unipoint)) { + handle_first_norm(savepoints,&numpoints,unipoint); + continue; + } + ret = CONS(hp,make_small(unipoint),ret); + hp += 2; + } + /* so, we'we looped to the beginning, do we have anything saved? */ + if (numpoints) { + cleanup_norm(&hp,savepoints,numpoints,&ret); + } + if (hp_end != hp) { + HRelease(p,hp_end,hp); + } + return ret; +} + /* * The last step of characters_to_list, build a list from the buffer 'bytes' (created in the same way * as for characters_to_utf8). All sizes are known in advance and most data will be held in a @@ -1378,10 +1583,10 @@ static BIF_RETTYPE finalize_list_to_list(Process *p, */ free_restart(bytes); - if (state == UTF8_INCOMPLETE) { + if (state == ERTS_UTF8_INCOMPLETE) { hp = HAlloc(p,4); ret = TUPLE3(hp,am_incomplete,converted,rest); - } else if (state == UTF8_ERROR) { + } else if (state == ERTS_UTF8_ERROR) { hp = HAlloc(p,4); ret = TUPLE3(hp,am_error,converted,rest); } else { @@ -1408,7 +1613,7 @@ static BIF_RETTYPE characters_to_list_trap_2(BIF_ALIST_3) /* * Hooks into the process of decoding a binary depending on state. - * If last_state is UTF8_ANALYZE_MORE, num_bytes_to_process + * If last_state is ERTS_UTF8_ANALYZE_MORE, num_bytes_to_process * and num_resulting_chars will grow * until we're done analyzing the binary. Then we'll eat * the bytes to process, lowering num_bytes_to_process and num_resulting_chars, @@ -1465,14 +1670,14 @@ static BIF_RETTYPE do_bif_utf8_to_list(Process *p, left = allowed_iterations(p); - if (state == UTF8_ANALYZE_MORE) { - state = analyze_utf8(bytes + num_bytes_to_process, + if (state == ERTS_UTF8_ANALYZE_MORE) { + state = erts_analyze_utf8(bytes + num_bytes_to_process, size - num_bytes_to_process, &endpos,&numchar,&left); cost_to_proc(p,numchar); num_resulting_chars += numchar; num_bytes_to_process = endpos - bytes; - if (state == UTF8_ANALYZE_MORE) { + if (state == ERTS_UTF8_ANALYZE_MORE) { Eterm epos = erts_make_integer(num_bytes_to_process,p); Eterm enumchar = erts_make_integer(num_resulting_chars,p); erts_free_aligned_binary_bytes(temp_alloc); @@ -1528,7 +1733,7 @@ static BIF_RETTYPE do_bif_utf8_to_list(Process *p, ErlSubBin *sb; Eterm orig; Uint offset; - ASSERT(state != UTF8_OK); + ASSERT(state != ERTS_UTF8_OK); hp = HAlloc(p, ERL_SUB_BIN_SIZE); sb = (ErlSubBin *) hp; ERTS_GET_REAL_BIN(orig_bin, orig, offset, bitoffs, bitsize); @@ -1544,14 +1749,14 @@ static BIF_RETTYPE do_bif_utf8_to_list(Process *p, /* Done */ - if (state == UTF8_INCOMPLETE) { + if (state == ERTS_UTF8_INCOMPLETE) { if (check_leftovers(bytes + num_bytes_to_process + num_processed_bytes, b_sz) != 0) { goto error_return; } hp = HAlloc(p,4); ret = TUPLE3(hp,am_incomplete,converted,rest); - } else if (state == UTF8_ERROR) { + } else if (state == ERTS_UTF8_ERROR) { error_return: hp = HAlloc(p,4); ret = TUPLE3(hp,am_error,converted,rest); @@ -1589,7 +1794,7 @@ static BIF_RETTYPE characters_to_list_trap_3(BIF_ALIST_3) 0U, /* nothing processed yet */ num_bytes_to_process, num_resulting_chars, - UTF8_ANALYZE_MORE, /* always this state here */ + ERTS_UTF8_ANALYZE_MORE, /* always this state here */ NIL); /* Nothing built -> no tail yet */ } @@ -1642,7 +1847,7 @@ static BIF_RETTYPE utf8_to_list(BIF_ALIST_1) BIF_ERROR(BIF_P,BADARG); } return do_bif_utf8_to_list(BIF_P, BIF_ARG_1, 0U, 0U, 0U, - UTF8_ANALYZE_MORE,NIL); + ERTS_UTF8_ANALYZE_MORE,NIL); } @@ -1728,8 +1933,8 @@ binary_to_atom(Process* p, Eterm bin, Eterm enc, int must_exist) Uint n; int reds_left = bin_size+1; /* Number of reductions left. */ - if (analyze_utf8(bytes, bin_size, &err_pos, - &n, &reds_left) == UTF8_OK) { + if (erts_analyze_utf8(bytes, bin_size, &err_pos, + &n, &reds_left) == ERTS_UTF8_OK) { /* * Correct UTF-8 encoding, but too many characters to * fit in an atom. @@ -1813,3 +2018,616 @@ BIF_RETTYPE binary_to_existing_atom_2(BIF_ALIST_2) { return binary_to_atom(BIF_P, BIF_ARG_1, BIF_ARG_2, 1); } + +/********************************************************** + * Simpler non-interruptable routines for UTF-8 and + * Windowish UTF-16 (restricted) + **********************************************************/ +/* + * This function is the heart of the Unicode support for + * open_port - spawn_executable. It converts both the name + * of the executable and the arguments according to the same rules + * as for filename conversion. That means as if your arguments are + * to be raw, you supply binaries, else unicode characters are allowed up to + * the encoding maximum (256 of the unicode max). + * Depending on the filename encoding standard, the vector is then + * converted to whatever is used, which might mean win_utf16 if on windows. + * Do not peek into the argument vector or filenam with ordinary + * string routines, that will certainly fail on some OS. + */ + +char *erts_convert_filename_to_native(Eterm name, ErtsAlcType_t alloc_type, int allow_empty) +{ + int encoding = erts_get_native_filename_encoding(); + char* name_buf = NULL; + + if (is_atom(name) || is_list(name) || (allow_empty && is_nil(name))) { + Sint need; + if ((need = erts_native_filename_need(name,encoding)) < 0) { + return NULL; + } + if (encoding == ERL_FILENAME_WIN_WCHAR) { + need += 2; + } else { + ++need; + } + name_buf = (char *) erts_alloc(alloc_type, need); + erts_native_filename_put(name,encoding,(byte *)name_buf); + name_buf[need-1] = 0; + if (encoding == ERL_FILENAME_WIN_WCHAR) { + name_buf[need-2] = 0; + } + } else if (is_binary(name)) { + byte *temp_alloc = NULL; + byte *bytes; + byte *err_pos; + Uint size,num_chars; + + size = binary_size(name); + bytes = erts_get_aligned_binary_bytes(name, &temp_alloc); + if (encoding != ERL_FILENAME_WIN_WCHAR) { + /*Add 0 termination only*/ + name_buf = (char *) erts_alloc(alloc_type, size+1); + memcpy(name_buf,bytes,size); + name_buf[size]=0; + } else if (erts_analyze_utf8(bytes,size,&err_pos,&num_chars,NULL) != ERTS_UTF8_OK || + erts_get_user_requested_filename_encoding() == ERL_FILENAME_LATIN1) { + byte *p; + /* What to do now? Maybe latin1, so just take byte for byte instead */ + name_buf = (char *) erts_alloc(alloc_type, (size+1)*2); + p = (byte *) name_buf; + while (size--) { + *p++ = *bytes++; + *p++ = 0; + } + *p++ = 0; + *p++ = 0; + } else { /* WIN_WCHAR and valid UTF8 */ + name_buf = (char *) erts_alloc(alloc_type, (num_chars+1)*2); + erts_copy_utf8_to_utf16_little((byte *) name_buf, bytes, num_chars); + name_buf[num_chars*2] = 0; + name_buf[num_chars*2+1] = 0; + } + erts_free_aligned_binary_bytes(temp_alloc); + } else { + return NULL; + } + return name_buf; +} + + +Sint erts_native_filename_need(Eterm ioterm, int encoding) +{ + Eterm *objp; + Eterm obj; + DECLARE_ESTACK(stack); + Sint need = 0; + + if (is_atom(ioterm)) { + Atom* ap; + int i; + ap = atom_tab(atom_val(ioterm)); + switch (encoding) { + case ERL_FILENAME_LATIN1: + need = ap->len; + break; + case ERL_FILENAME_UTF8_MAC: + case ERL_FILENAME_UTF8: + for (i = 0; i < ap->len; i++) { + need += (ap->name[i] >= 0x80) ? 2 : 1; + } + break; + case ERL_FILENAME_WIN_WCHAR: + need = 2*(ap->len); + break; + default: + need = -1; + } + DESTROY_ESTACK(stack); + return need; + } + + if (is_nil(ioterm)) { + DESTROY_ESTACK(stack); + return need; + } + if (!is_list(ioterm)) { + DESTROY_ESTACK(stack); + return (Sint) -1; + } + /* OK a list, needs to be processed in order, handling each flat list-level + as they occur, just like io_list_to_binary would */ + ESTACK_PUSH(stack,ioterm); + while (!ESTACK_ISEMPTY(stack)) { + ioterm = ESTACK_POP(stack); + if (is_nil(ioterm)) { + /* ignore empty lists */ + continue; + } + if(is_list(ioterm)) { +L_Again: /* Restart with sublist, old listend was pushed on stack */ + objp = list_val(ioterm); + obj = CAR(objp); + for(;;) { /* loop over one flat list of bytes and binaries + until sublist or list end is encountered */ + if (is_small(obj)) { /* Always small */ + for(;;) { + Uint x = unsigned_val(obj); + switch (encoding) { + case ERL_FILENAME_LATIN1: + if (x > 255) { + DESTROY_ESTACK(stack); + return ((Sint) -1); + } + need += 1; + break; + case ERL_FILENAME_UTF8_MAC: + case ERL_FILENAME_UTF8: + if (x < 0x80) { + need +=1; + } else if (x < 0x800) { + need += 2; + } else if (x < 0x10000) { + if ((x >= 0xD800 && x <= 0xDFFF) || + (x == 0xFFFE) || + (x == 0xFFFF)) { /* Invalid unicode range */ + DESTROY_ESTACK(stack); + return ((Sint) -1); + } + need += 3; + } else if (x < 0x110000) { + need += 4; + } else { + DESTROY_ESTACK(stack); + return ((Sint) -1); + } + break; + case ERL_FILENAME_WIN_WCHAR: + if (x <= 0xffff) { + need += 2; + break; + } /* else fall throug to error */ + default: + DESTROY_ESTACK(stack); + return ((Sint) -1); + } + + /* everything else will give badarg later + in the process, so we dont check */ + ioterm = CDR(objp); + if (!is_list(ioterm)) { + break; + } + objp = list_val(ioterm); + obj = CAR(objp); + if (!is_small(obj)) + break; + } + } else if (is_nil(obj)) { + ioterm = CDR(objp); + if (!is_list(ioterm)) { + break; + } + objp = list_val(ioterm); + obj = CAR(objp); + } else if (is_list(obj)) { + /* push rest of list for later processing, start + again with sublist */ + ESTACK_PUSH(stack,CDR(objp)); + ioterm = obj; + goto L_Again; + } else { + DESTROY_ESTACK(stack); + return ((Sint) -1); + } + if (is_nil(ioterm) || !is_list(ioterm)) { + break; + } + } /* for(;;) */ + } /* is_list(ioterm) */ + + if (!is_list(ioterm) && !is_nil(ioterm)) { + /* inproper list end */ + DESTROY_ESTACK(stack); + return ((Sint) -1); + } + } /* while not estack empty */ + DESTROY_ESTACK(stack); + return need; +} + +void erts_native_filename_put(Eterm ioterm, int encoding, byte *p) +{ + Eterm *objp; + Eterm obj; + DECLARE_ESTACK(stack); + + if (is_atom(ioterm)) { + Atom* ap; + int i; + ap = atom_tab(atom_val(ioterm)); + switch (encoding) { + case ERL_FILENAME_LATIN1: + for (i = 0; i < ap->len; i++) { + *p++ = ap->name[i]; + } + break; + case ERL_FILENAME_UTF8_MAC: + case ERL_FILENAME_UTF8: + for (i = 0; i < ap->len; i++) { + if(ap->name[i] < 0x80) { + *p++ = ap->name[i]; + } else { + *p++ = (((ap->name[i]) >> 6) | ((byte) 0xC0)); + *p++ = (((ap->name[i]) & 0x3F) | ((byte) 0x80)); + } + } + break; + case ERL_FILENAME_WIN_WCHAR: + for (i = 0; i < ap->len; i++) { + /* Little endian */ + *p++ = ap->name[i]; + *p++ = 0; + } + break; + default: + ASSERT(0); + } + DESTROY_ESTACK(stack); + return; + } + + if (is_nil(ioterm)) { + DESTROY_ESTACK(stack); + return; + } + ASSERT(is_list(ioterm)); + /* OK a list, needs to be processed in order, handling each flat list-level + as they occur, just like io_list_to_binary would */ + ESTACK_PUSH(stack,ioterm); + while (!ESTACK_ISEMPTY(stack)) { + ioterm = ESTACK_POP(stack); + if (is_nil(ioterm)) { + /* ignore empty lists */ + continue; + } + if(is_list(ioterm)) { +L_Again: /* Restart with sublist, old listend was pushed on stack */ + objp = list_val(ioterm); + obj = CAR(objp); + for(;;) { /* loop over one flat list of bytes and binaries + until sublist or list end is encountered */ + if (is_small(obj)) { /* Always small */ + for(;;) { + Uint x = unsigned_val(obj); + switch (encoding) { + case ERL_FILENAME_LATIN1: + ASSERT( x < 256); + *p++ = (byte) x; + break; + case ERL_FILENAME_UTF8_MAC: + case ERL_FILENAME_UTF8: + if (x < 0x80) { + *p++ = (byte) x; + } + else if (x < 0x800) { + *p++ = (((byte) (x >> 6)) | + ((byte) 0xC0)); + *p++ = (((byte) (x & 0x3F)) | + ((byte) 0x80)); + } else if (x < 0x10000) { + ASSERT(!((x >= 0xD800 && x <= 0xDFFF) || + (x == 0xFFFE) || + (x == 0xFFFF))); + *p++ = (((byte) (x >> 12)) | + ((byte) 0xE0)); + *p++ = ((((byte) (x >> 6)) & 0x3F) | + ((byte) 0x80)); + *p++ = (((byte) (x & 0x3F)) | + ((byte) 0x80)); + } else { + ASSERT(x < 0x110000); + *p++ = (((byte) (x >> 18)) | + ((byte) 0xF0)); + *p++ = ((((byte) (x >> 12)) & 0x3F) | + ((byte) 0x80)); + *p++ = ((((byte) (x >> 6)) & 0x3F) | + ((byte) 0x80)); + *p++ = (((byte) (x & 0x3F)) | + ((byte) 0x80)); + } + break; + case ERL_FILENAME_WIN_WCHAR: + ASSERT(x <= 0xFFFF); + *p++ = (byte) (x & 0xFFU); + *p++ = (byte) ((x >> 8) & 0xFFU); + break; + default: + ASSERT(0); + } + + /* everything else will give badarg later + in the process, so we dont check */ + ioterm = CDR(objp); + if (!is_list(ioterm)) { + break; + } + objp = list_val(ioterm); + obj = CAR(objp); + if (!is_small(obj)) + break; + } + } else if (is_nil(obj)) { + ioterm = CDR(objp); + if (!is_list(ioterm)) { + break; + } + objp = list_val(ioterm); + obj = CAR(objp); + } else if (is_list(obj)) { + /* push rest of list for later processing, start + again with sublist */ + ESTACK_PUSH(stack,CDR(objp)); + ioterm = obj; + goto L_Again; + } else { + ASSERT(0); + } + if (is_nil(ioterm) || !is_list(ioterm)) { + break; + } + } /* for(;;) */ + } /* is_list(ioterm) */ + + ASSERT(is_list(ioterm) || is_nil(ioterm)); + } /* while not estack empty */ + DESTROY_ESTACK(stack); + return; +} +void erts_copy_utf8_to_utf16_little(byte *target, byte *bytes, int num_chars) +{ + Uint unipoint; + + while (num_chars--) { + if (((*bytes) & ((byte) 0x80)) == 0) { + unipoint = (Uint) *bytes; + ++bytes; + } else if (((*bytes) & ((byte) 0xE0)) == 0xC0) { + unipoint = + (((Uint) ((*bytes) & ((byte) 0x1F))) << 6) | + ((Uint) (bytes[1] & ((byte) 0x3F))); + bytes += 2; + } else if (((*bytes) & ((byte) 0xF0)) == 0xE0) { + unipoint = + (((Uint) ((*bytes) & ((byte) 0xF))) << 12) | + (((Uint) (bytes[1] & ((byte) 0x3F))) << 6) | + ((Uint) (bytes[2] & ((byte) 0x3F))); + bytes +=3; + } else if (((*bytes) & ((byte) 0xF8)) == 0xF0) { + unipoint = + (((Uint) ((*bytes) & ((byte) 0x7))) << 18) | + (((Uint) (bytes[1] & ((byte) 0x3F))) << 12) | + (((Uint) (bytes[2] & ((byte) 0x3F))) << 6) | + ((Uint) (bytes[3] & ((byte) 0x3F))); + bytes += 4; + } else { + erl_exit(1,"Internal unicode error in prim_file:internal_name2native/1"); + } + *target++ = (byte) (unipoint & 0xFF); + *target++ = (byte) ((unipoint >> 8) & 0xFF); + } +} + +/* + * This internal bif converts a filename to whatever format is suitable for the file driver + * It also adds zero termination so that prim_file neednt bother with the character encoding + * of the file driver + */ +BIF_RETTYPE prim_file_internal_name2native_1(BIF_ALIST_1) +{ + int encoding = erts_get_native_filename_encoding(); + Sint need; + Eterm bin_term; + byte* bin_p; + /* Prim file explicitly does not allow atoms, although we could + very well cope with it. Instead of letting 'file' handle them, + it would probably be more efficient to handle them here. Subject to + change in R15. */ + if (is_atom(BIF_ARG_1)) { + BIF_ERROR(BIF_P,BADARG); + } + if (is_binary(BIF_ARG_1)) { + byte *temp_alloc = NULL; + byte *bytes; + byte *err_pos; + Uint size,num_chars; + /* Uninterpreted encoding except if windows widechar, in case we convert from + utf8 to win_wchar */ + size = binary_size(BIF_ARG_1); + bytes = erts_get_aligned_binary_bytes(BIF_ARG_1, &temp_alloc); + if (encoding != ERL_FILENAME_WIN_WCHAR) { + /*Add 0 termination only*/ + bin_term = new_binary(BIF_P, NULL, size+1); + bin_p = binary_bytes(bin_term); + memcpy(bin_p,bytes,size); + bin_p[size]=0; + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(bin_term); + } + /* In a wchar world, the emulator flags only affect how + binaries are interpreted when sent from the user. */ + /* Determine real length and create a new binary */ + if (erts_analyze_utf8(bytes,size,&err_pos,&num_chars,NULL) != ERTS_UTF8_OK || + erts_get_user_requested_filename_encoding() == ERL_FILENAME_LATIN1) { + /* What to do now? Maybe latin1, so just take byte for byte instead */ + bin_term = new_binary(BIF_P, 0, (size+1)*2); + bin_p = binary_bytes(bin_term); + while (size--) { + *bin_p++ = *bytes++; + *bin_p++ = 0; + } + *bin_p++ = 0; + *bin_p++ = 0; + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(bin_term); + } + /* OK, UTF8 ok, number of characters is in num_chars */ + bin_term = new_binary(BIF_P, 0, (num_chars+1)*2); + bin_p = binary_bytes(bin_term); + erts_copy_utf8_to_utf16_little(bin_p, bytes, num_chars); + /* zero termination */ + bin_p[num_chars*2] = 0; + bin_p[num_chars*2+1] = 0; + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(bin_term); + } /* binary */ + + + if ((need = erts_native_filename_need(BIF_ARG_1,encoding)) < 0) { + BIF_ERROR(BIF_P,BADARG); + } + if (encoding == ERL_FILENAME_WIN_WCHAR) { + need += 2; + } else { + ++need; + } + + bin_term = new_binary(BIF_P, 0, need); + bin_p = binary_bytes(bin_term); + erts_native_filename_put(BIF_ARG_1,encoding,bin_p); + bin_p[need-1] = 0; + if (encoding == ERL_FILENAME_WIN_WCHAR) { + bin_p[need-2] = 0; + } + BIF_RET(bin_term); +} + +BIF_RETTYPE prim_file_internal_native2name_1(BIF_ALIST_1) +{ + Eterm real_bin; + Uint offset; + Uint size,num_chars; + Uint bitsize; + Uint bitoffs; + Eterm *hp; + byte *temp_alloc = NULL; + byte *bytes; + byte *err_pos; + Uint num_built; /* characters */ + Uint num_eaten; /* bytes */ + Eterm ret; + int mac = 0; + + if (is_not_binary(BIF_ARG_1)) { + BIF_ERROR(BIF_P,BADARG); + } + size = binary_size(BIF_ARG_1); + ERTS_GET_REAL_BIN(BIF_ARG_1, real_bin, offset, bitoffs, bitsize); + if (bitsize != 0) { + BIF_ERROR(BIF_P,BADARG); + } + if (size == 0) { + BIF_RET(NIL); + } + switch (erts_get_native_filename_encoding()) { + case ERL_FILENAME_LATIN1: + hp = HAlloc(BIF_P, 2 * size); + bytes = binary_bytes(real_bin)+offset; + + BIF_RET(erts_bin_bytes_to_list(NIL, hp, bytes, size, bitoffs)); + case ERL_FILENAME_UTF8_MAC: + mac = 1; + case ERL_FILENAME_UTF8: + bytes = erts_get_aligned_binary_bytes(BIF_ARG_1, &temp_alloc); + if (erts_analyze_utf8(bytes,size,&err_pos,&num_chars,NULL) != ERTS_UTF8_OK) { + erts_free_aligned_binary_bytes(temp_alloc); + goto noconvert; + } + num_built = 0; + num_eaten = 0; + if (mac) { + ret = do_utf8_to_list_normalize(BIF_P, num_chars, bytes, size); + } else { + ret = do_utf8_to_list(BIF_P, num_chars, bytes, size, num_chars, &num_built, &num_eaten, NIL); + } + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(ret); + case ERL_FILENAME_WIN_WCHAR: + bytes = erts_get_aligned_binary_bytes(BIF_ARG_1, &temp_alloc); + if ((size % 2) != 0) { /* Panic fixup to avoid crashing the emulator */ + size--; + hp = HAlloc(BIF_P, size+2); + ret = CONS(hp,make_small((Uint) bytes[size]),NIL); + hp += 2; + } else { + hp = HAlloc(BIF_P, size); + ret = NIL; + } + bytes += size-1; + while (size > 0) { + Uint x = ((Uint) *bytes--) << 8; + x |= ((Uint) *bytes--); + size -= 2; + ret = CONS(hp,make_small(x),ret); + hp += 2; + } + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(ret); + default: + goto noconvert; + } + noconvert: + BIF_RET(BIF_ARG_1); +} + +BIF_RETTYPE prim_file_internal_normalize_utf8_1(BIF_ALIST_1) +{ + Eterm real_bin; + Uint offset; + Uint size,num_chars; + Uint bitsize; + Uint bitoffs; + Eterm ret; + byte *temp_alloc = NULL; + byte *bytes; + byte *err_pos; + + if (is_not_binary(BIF_ARG_1)) { + BIF_ERROR(BIF_P,BADARG); + } + size = binary_size(BIF_ARG_1); + ERTS_GET_REAL_BIN(BIF_ARG_1, real_bin, offset, bitoffs, bitsize); + if (bitsize != 0) { + BIF_ERROR(BIF_P,BADARG); + } + if (size == 0) { + BIF_RET(NIL); + } + bytes = erts_get_aligned_binary_bytes(BIF_ARG_1, &temp_alloc); + if (erts_analyze_utf8(bytes,size,&err_pos,&num_chars,NULL) != ERTS_UTF8_OK) { + erts_free_aligned_binary_bytes(temp_alloc); + BIF_ERROR(BIF_P,BADARG); + } + ret = do_utf8_to_list_normalize(BIF_P, num_chars, bytes, size); + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(ret); +} + +BIF_RETTYPE file_native_name_encoding_0(BIF_ALIST_0) +{ + switch (erts_get_native_filename_encoding()) { + case ERL_FILENAME_LATIN1: + BIF_RET(am_latin1); + case ERL_FILENAME_UTF8_MAC: + case ERL_FILENAME_UTF8: + BIF_RET(am_utf8); + case ERL_FILENAME_WIN_WCHAR: + if (erts_get_user_requested_filename_encoding() == ERL_FILENAME_LATIN1) { + BIF_RET(am_latin1); + } else { + BIF_RET(am_utf8); + } + default: + BIF_RET(am_undefined); + } +} diff --git a/erts/emulator/beam/erl_unicode_normalize.h b/erts/emulator/beam/erl_unicode_normalize.h new file mode 100644 index 0000000000..fb0a111ca2 --- /dev/null +++ b/erts/emulator/beam/erl_unicode_normalize.h @@ -0,0 +1,1687 @@ +/* +* %CopyrightBegin% +* +* Copyright Ericsson AB 1999-2010. All Rights Reserved. +* +* The contents of this file are subject to the Erlang Public License, +* Version 1.1, (the "License"); you may not use this file except in +* compliance with the License. You should have received a copy of the +* Erlang Public License along with this software. If not, it can be +* retrieved online at http://www.erlang.org/. +* +* Software distributed under the License is distributed on an "AS IS" +* basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See +* the License for the specific language governing rights and limitations +* under the License. +* +* %CopyrightEnd% +*/ +/* +* This file is automatically generated by dec.erl, do not edit manually +*/ +#define HASH_SIZE_FACTOR 2 +typedef struct _compose_entry { + Uint16 c; + Uint16 res; + Uint16 num_subs; + struct _compose_entry *subs; + int *hash; +} CompEntry; + +static int compose_tab_size = 61; +static int hash_compose_tab_0_15[12] = +{-1,3,-1,5,-1,0,4,2,-1,1,-1,-1}; /* hash_compose_tab_0_15 */ +static CompEntry compose_tab_0_15[] = { +{65, 7846, 0, NULL, NULL}, +{69, 7872, 0, NULL, NULL}, +{79, 7890, 0, NULL, NULL}, +{97, 7847, 0, NULL, NULL}, +{101, 7873, 0, NULL, NULL}, +{111, 7891, 0, NULL, NULL} +}; /* compose_tab_0_15 */ +static int hash_compose_tab_0_16[8] = +{3,-1,-1,-1,-1,0,2,1}; /* hash_compose_tab_0_16 */ +static CompEntry compose_tab_0_16[] = { +{69, 7700, 0, NULL, NULL}, +{79, 7760, 0, NULL, NULL}, +{101, 7701, 0, NULL, NULL}, +{111, 7761, 0, NULL, NULL} +}; /* compose_tab_0_16 */ +static int hash_compose_tab_0_17[4] = +{-1,0,1,-1}; /* hash_compose_tab_0_17 */ +static CompEntry compose_tab_0_17[] = { +{65, 7856, 0, NULL, NULL}, +{97, 7857, 0, NULL, NULL} +}; /* compose_tab_0_17 */ +static int hash_compose_tab_0_18[8] = +{-1,2,-1,-1,-1,0,1,3}; /* hash_compose_tab_0_18 */ +static CompEntry compose_tab_0_18[] = { +{85, 475, 0, NULL, NULL}, +{117, 476, 0, NULL, NULL}, +{953, 8146, 0, NULL, NULL}, +{965, 8162, 0, NULL, NULL} +}; /* compose_tab_0_18 */ +static int hash_compose_tab_0_19_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_0_19_0 */ +static CompEntry compose_tab_0_19_0[] = { +{913, 8074, 0, NULL, NULL}, +{919, 8090, 0, NULL, NULL}, +{937, 8106, 0, NULL, NULL}, +{945, 8066, 0, NULL, NULL}, +{951, 8082, 0, NULL, NULL}, +{969, 8098, 0, NULL, NULL} +}; /* compose_tab_0_19_0 */ +static int hash_compose_tab_0_19[28] = +{9,10,-1,5,-1,-1,-1,11,-1,-1,-1,-1,-1,6,12,-1,-1,1,13,-1,-1,2,7,3,-1,0,4,8}; /* hash_compose_tab_0_19 */ +static CompEntry compose_tab_0_19[] = { +{837, 0, 6, compose_tab_0_19_0, hash_compose_tab_0_19_0}, +{913, 7946, 0, NULL, NULL}, +{917, 7962, 0, NULL, NULL}, +{919, 7978, 0, NULL, NULL}, +{921, 7994, 0, NULL, NULL}, +{927, 8010, 0, NULL, NULL}, +{937, 8042, 0, NULL, NULL}, +{945, 7938, 0, NULL, NULL}, +{949, 7954, 0, NULL, NULL}, +{951, 7970, 0, NULL, NULL}, +{953, 7986, 0, NULL, NULL}, +{959, 8002, 0, NULL, NULL}, +{965, 8018, 0, NULL, NULL}, +{969, 8034, 0, NULL, NULL} +}; /* compose_tab_0_19 */ +static int hash_compose_tab_0_20_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_0_20_0 */ +static CompEntry compose_tab_0_20_0[] = { +{913, 8075, 0, NULL, NULL}, +{919, 8091, 0, NULL, NULL}, +{937, 8107, 0, NULL, NULL}, +{945, 8067, 0, NULL, NULL}, +{951, 8083, 0, NULL, NULL}, +{969, 8099, 0, NULL, NULL} +}; /* compose_tab_0_20_0 */ +static int hash_compose_tab_0_20[30] = +{-1,-1,-1,6,-1,13,-1,7,-1,14,-1,-1,-1,1,-1,8,-1,2,-1,3,9,4,10,11,-1,-1,-1,0,5, + 12}; /* hash_compose_tab_0_20 */ +static CompEntry compose_tab_0_20[] = { +{837, 0, 6, compose_tab_0_20_0, hash_compose_tab_0_20_0}, +{913, 7947, 0, NULL, NULL}, +{917, 7963, 0, NULL, NULL}, +{919, 7979, 0, NULL, NULL}, +{921, 7995, 0, NULL, NULL}, +{927, 8011, 0, NULL, NULL}, +{933, 8027, 0, NULL, NULL}, +{937, 8043, 0, NULL, NULL}, +{945, 7939, 0, NULL, NULL}, +{949, 7955, 0, NULL, NULL}, +{951, 7971, 0, NULL, NULL}, +{953, 7987, 0, NULL, NULL}, +{959, 8003, 0, NULL, NULL}, +{965, 8019, 0, NULL, NULL}, +{969, 8035, 0, NULL, NULL} +}; /* compose_tab_0_20 */ +static int hash_compose_tab_0_21[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_0_21 */ +static CompEntry compose_tab_0_21[] = { +{79, 7900, 0, NULL, NULL}, +{85, 7914, 0, NULL, NULL}, +{111, 7901, 0, NULL, NULL}, +{117, 7915, 0, NULL, NULL} +}; /* compose_tab_0_21 */ +static int hash_compose_tab_0_22[6] = +{-1,-1,-1,0,1,2}; /* hash_compose_tab_0_22 */ +static CompEntry compose_tab_0_22[] = { +{945, 8114, 0, NULL, NULL}, +{951, 8130, 0, NULL, NULL}, +{969, 8178, 0, NULL, NULL} +}; /* compose_tab_0_22 */ +static int hash_compose_tab_0[78] = +{38,3,29,-1,-1,-1,-1,4,19,5,20,6,14,30,31,21,32,33,37,7,-1,-1,-1,8,34,-1,-1,9, + -1,35,-1,-1,-1,10,36,-1,-1,-1,-1,11,-1,12,-1,13,-1,-1,-1,-1,-1,-1,-1,-1,-1, + -1,-1,23,-1,22,-1,24,-1,25,-1,26,-1,0,-1,-1,15,1,16,27,17,2,18,28,-1,-1}; /* hash_compose_tab_0 */ +static CompEntry compose_tab_0[] = { +{65, 192, 0, NULL, NULL}, +{69, 200, 0, NULL, NULL}, +{73, 204, 0, NULL, NULL}, +{79, 210, 0, NULL, NULL}, +{85, 217, 0, NULL, NULL}, +{87, 7808, 0, NULL, NULL}, +{89, 7922, 0, NULL, NULL}, +{97, 224, 0, NULL, NULL}, +{101, 232, 0, NULL, NULL}, +{105, 236, 0, NULL, NULL}, +{111, 242, 0, NULL, NULL}, +{117, 249, 0, NULL, NULL}, +{119, 7809, 0, NULL, NULL}, +{121, 7923, 0, NULL, NULL}, +{168, 8173, 0, NULL, NULL}, +{770, 0, 6, compose_tab_0_15, hash_compose_tab_0_15}, +{772, 0, 4, compose_tab_0_16, hash_compose_tab_0_16}, +{774, 0, 2, compose_tab_0_17, hash_compose_tab_0_17}, +{776, 0, 4, compose_tab_0_18, hash_compose_tab_0_18}, +{787, 0, 14, compose_tab_0_19, hash_compose_tab_0_19}, +{788, 0, 15, compose_tab_0_20, hash_compose_tab_0_20}, +{795, 0, 4, compose_tab_0_21, hash_compose_tab_0_21}, +{837, 0, 3, compose_tab_0_22, hash_compose_tab_0_22}, +{913, 8122, 0, NULL, NULL}, +{917, 8136, 0, NULL, NULL}, +{919, 8138, 0, NULL, NULL}, +{921, 8154, 0, NULL, NULL}, +{927, 8184, 0, NULL, NULL}, +{933, 8170, 0, NULL, NULL}, +{937, 8186, 0, NULL, NULL}, +{945, 8048, 0, NULL, NULL}, +{949, 8050, 0, NULL, NULL}, +{951, 8052, 0, NULL, NULL}, +{953, 8054, 0, NULL, NULL}, +{959, 8056, 0, NULL, NULL}, +{965, 8058, 0, NULL, NULL}, +{969, 8060, 0, NULL, NULL}, +{8127, 8141, 0, NULL, NULL}, +{8190, 8157, 0, NULL, NULL} +}; /* compose_tab_0 */ +static int hash_compose_tab_1_39[12] = +{-1,3,-1,5,-1,0,4,2,-1,1,-1,-1}; /* hash_compose_tab_1_39 */ +static CompEntry compose_tab_1_39[] = { +{65, 7844, 0, NULL, NULL}, +{69, 7870, 0, NULL, NULL}, +{79, 7888, 0, NULL, NULL}, +{97, 7845, 0, NULL, NULL}, +{101, 7871, 0, NULL, NULL}, +{111, 7889, 0, NULL, NULL} +}; /* compose_tab_1_39 */ +static int hash_compose_tab_1_40[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_1_40 */ +static CompEntry compose_tab_1_40[] = { +{79, 7756, 0, NULL, NULL}, +{85, 7800, 0, NULL, NULL}, +{111, 7757, 0, NULL, NULL}, +{117, 7801, 0, NULL, NULL} +}; /* compose_tab_1_40 */ +static int hash_compose_tab_1_41[8] = +{3,-1,-1,-1,-1,0,2,1}; /* hash_compose_tab_1_41 */ +static CompEntry compose_tab_1_41[] = { +{69, 7702, 0, NULL, NULL}, +{79, 7762, 0, NULL, NULL}, +{101, 7703, 0, NULL, NULL}, +{111, 7763, 0, NULL, NULL} +}; /* compose_tab_1_41 */ +static int hash_compose_tab_1_42[4] = +{-1,0,1,-1}; /* hash_compose_tab_1_42 */ +static CompEntry compose_tab_1_42[] = { +{65, 7854, 0, NULL, NULL}, +{97, 7855, 0, NULL, NULL} +}; /* compose_tab_1_42 */ +static int hash_compose_tab_1_43[12] = +{-1,0,1,-1,-1,4,5,-1,-1,2,3,-1}; /* hash_compose_tab_1_43 */ +static CompEntry compose_tab_1_43[] = { +{73, 7726, 0, NULL, NULL}, +{85, 471, 0, NULL, NULL}, +{105, 7727, 0, NULL, NULL}, +{117, 472, 0, NULL, NULL}, +{953, 8147, 0, NULL, NULL}, +{965, 8163, 0, NULL, NULL} +}; /* compose_tab_1_43 */ +static int hash_compose_tab_1_44[4] = +{-1,0,1,-1}; /* hash_compose_tab_1_44 */ +static CompEntry compose_tab_1_44[] = { +{65, 506, 0, NULL, NULL}, +{97, 507, 0, NULL, NULL} +}; /* compose_tab_1_44 */ +static int hash_compose_tab_1_45_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_1_45_0 */ +static CompEntry compose_tab_1_45_0[] = { +{913, 8076, 0, NULL, NULL}, +{919, 8092, 0, NULL, NULL}, +{937, 8108, 0, NULL, NULL}, +{945, 8068, 0, NULL, NULL}, +{951, 8084, 0, NULL, NULL}, +{969, 8100, 0, NULL, NULL} +}; /* compose_tab_1_45_0 */ +static int hash_compose_tab_1_45[28] = +{9,10,-1,5,-1,-1,-1,11,-1,-1,-1,-1,-1,6,12,-1,-1,1,13,-1,-1,2,7,3,-1,0,4,8}; /* hash_compose_tab_1_45 */ +static CompEntry compose_tab_1_45[] = { +{837, 0, 6, compose_tab_1_45_0, hash_compose_tab_1_45_0}, +{913, 7948, 0, NULL, NULL}, +{917, 7964, 0, NULL, NULL}, +{919, 7980, 0, NULL, NULL}, +{921, 7996, 0, NULL, NULL}, +{927, 8012, 0, NULL, NULL}, +{937, 8044, 0, NULL, NULL}, +{945, 7940, 0, NULL, NULL}, +{949, 7956, 0, NULL, NULL}, +{951, 7972, 0, NULL, NULL}, +{953, 7988, 0, NULL, NULL}, +{959, 8004, 0, NULL, NULL}, +{965, 8020, 0, NULL, NULL}, +{969, 8036, 0, NULL, NULL} +}; /* compose_tab_1_45 */ +static int hash_compose_tab_1_46_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_1_46_0 */ +static CompEntry compose_tab_1_46_0[] = { +{913, 8077, 0, NULL, NULL}, +{919, 8093, 0, NULL, NULL}, +{937, 8109, 0, NULL, NULL}, +{945, 8069, 0, NULL, NULL}, +{951, 8085, 0, NULL, NULL}, +{969, 8101, 0, NULL, NULL} +}; /* compose_tab_1_46_0 */ +static int hash_compose_tab_1_46[30] = +{-1,-1,-1,6,-1,13,-1,7,-1,14,-1,-1,-1,1,-1,8,-1,2,-1,3,9,4,10,11,-1,-1,-1,0,5, + 12}; /* hash_compose_tab_1_46 */ +static CompEntry compose_tab_1_46[] = { +{837, 0, 6, compose_tab_1_46_0, hash_compose_tab_1_46_0}, +{913, 7949, 0, NULL, NULL}, +{917, 7965, 0, NULL, NULL}, +{919, 7981, 0, NULL, NULL}, +{921, 7997, 0, NULL, NULL}, +{927, 8013, 0, NULL, NULL}, +{933, 8029, 0, NULL, NULL}, +{937, 8045, 0, NULL, NULL}, +{945, 7941, 0, NULL, NULL}, +{949, 7957, 0, NULL, NULL}, +{951, 7973, 0, NULL, NULL}, +{953, 7989, 0, NULL, NULL}, +{959, 8005, 0, NULL, NULL}, +{965, 8021, 0, NULL, NULL}, +{969, 8037, 0, NULL, NULL} +}; /* compose_tab_1_46 */ +static int hash_compose_tab_1_47[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_1_47 */ +static CompEntry compose_tab_1_47[] = { +{79, 7898, 0, NULL, NULL}, +{85, 7912, 0, NULL, NULL}, +{111, 7899, 0, NULL, NULL}, +{117, 7913, 0, NULL, NULL} +}; /* compose_tab_1_47 */ +static int hash_compose_tab_1_48[4] = +{1,-1,-1,0}; /* hash_compose_tab_1_48 */ +static CompEntry compose_tab_1_48[] = { +{67, 7688, 0, NULL, NULL}, +{99, 7689, 0, NULL, NULL} +}; /* compose_tab_1_48 */ +static int hash_compose_tab_1_49[6] = +{-1,-1,-1,0,1,2}; /* hash_compose_tab_1_49 */ +static CompEntry compose_tab_1_49[] = { +{945, 8116, 0, NULL, NULL}, +{951, 8132, 0, NULL, NULL}, +{959, 8180, 0, NULL, NULL} +}; /* compose_tab_1_49 */ +static int hash_compose_tab_1[140] = +{-1,-1,-1,-1,-1,-1,-1,68,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, + -1,-1,-1,34,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, + -1,-1,-1,-1,-1,-1,-1,-1,35,-1,-1,-1,-1,64,-1,0,-1,1,-1,2,39,3,40,4,41,5,6,7, + 8,9,10,36,11,12,42,13,43,14,44,15,16,37,45,46,50,47,51,17,52,18,53,19,54,20, + 55,21,56,22,23,24,25,26,27,38,28,29,48,30,57,31,58,32,33,59,60,61,62,65,66, + 63,67,69,-1,-1,-1,-1,-1,49,-1,-1}; /* hash_compose_tab_1 */ +static CompEntry compose_tab_1[] = { +{65, 193, 0, NULL, NULL}, +{67, 262, 0, NULL, NULL}, +{69, 201, 0, NULL, NULL}, +{71, 500, 0, NULL, NULL}, +{73, 205, 0, NULL, NULL}, +{75, 7728, 0, NULL, NULL}, +{76, 313, 0, NULL, NULL}, +{77, 7742, 0, NULL, NULL}, +{78, 323, 0, NULL, NULL}, +{79, 211, 0, NULL, NULL}, +{80, 7764, 0, NULL, NULL}, +{82, 340, 0, NULL, NULL}, +{83, 346, 0, NULL, NULL}, +{85, 218, 0, NULL, NULL}, +{87, 7810, 0, NULL, NULL}, +{89, 221, 0, NULL, NULL}, +{90, 377, 0, NULL, NULL}, +{97, 225, 0, NULL, NULL}, +{99, 263, 0, NULL, NULL}, +{101, 233, 0, NULL, NULL}, +{103, 501, 0, NULL, NULL}, +{105, 237, 0, NULL, NULL}, +{107, 7729, 0, NULL, NULL}, +{108, 314, 0, NULL, NULL}, +{109, 7743, 0, NULL, NULL}, +{110, 324, 0, NULL, NULL}, +{111, 243, 0, NULL, NULL}, +{112, 7765, 0, NULL, NULL}, +{114, 341, 0, NULL, NULL}, +{115, 347, 0, NULL, NULL}, +{117, 250, 0, NULL, NULL}, +{119, 7811, 0, NULL, NULL}, +{121, 253, 0, NULL, NULL}, +{122, 378, 0, NULL, NULL}, +{168, 8174, 0, NULL, NULL}, +{198, 508, 0, NULL, NULL}, +{216, 510, 0, NULL, NULL}, +{230, 509, 0, NULL, NULL}, +{248, 511, 0, NULL, NULL}, +{770, 0, 6, compose_tab_1_39, hash_compose_tab_1_39}, +{771, 0, 4, compose_tab_1_40, hash_compose_tab_1_40}, +{772, 0, 4, compose_tab_1_41, hash_compose_tab_1_41}, +{774, 0, 2, compose_tab_1_42, hash_compose_tab_1_42}, +{776, 0, 6, compose_tab_1_43, hash_compose_tab_1_43}, +{778, 0, 2, compose_tab_1_44, hash_compose_tab_1_44}, +{787, 0, 14, compose_tab_1_45, hash_compose_tab_1_45}, +{788, 0, 15, compose_tab_1_46, hash_compose_tab_1_46}, +{795, 0, 4, compose_tab_1_47, hash_compose_tab_1_47}, +{807, 0, 2, compose_tab_1_48, hash_compose_tab_1_48}, +{837, 0, 3, compose_tab_1_49, hash_compose_tab_1_49}, +{913, 8123, 0, NULL, NULL}, +{917, 8137, 0, NULL, NULL}, +{919, 8139, 0, NULL, NULL}, +{921, 8155, 0, NULL, NULL}, +{927, 8185, 0, NULL, NULL}, +{933, 8171, 0, NULL, NULL}, +{937, 8187, 0, NULL, NULL}, +{945, 8049, 0, NULL, NULL}, +{949, 8051, 0, NULL, NULL}, +{951, 8053, 0, NULL, NULL}, +{953, 8055, 0, NULL, NULL}, +{959, 8057, 0, NULL, NULL}, +{965, 8059, 0, NULL, NULL}, +{969, 8061, 0, NULL, NULL}, +{1043, 1027, 0, NULL, NULL}, +{1050, 1036, 0, NULL, NULL}, +{1075, 1107, 0, NULL, NULL}, +{1082, 1116, 0, NULL, NULL}, +{8127, 8142, 0, NULL, NULL}, +{8190, 8158, 0, NULL, NULL} +}; /* compose_tab_1 */ +static int hash_compose_tab_2_26[12] = +{-1,3,-1,5,-1,0,4,2,-1,1,-1,-1}; /* hash_compose_tab_2_26 */ +static CompEntry compose_tab_2_26[] = { +{65, 7852, 0, NULL, NULL}, +{69, 7878, 0, NULL, NULL}, +{79, 7896, 0, NULL, NULL}, +{97, 7853, 0, NULL, NULL}, +{101, 7879, 0, NULL, NULL}, +{111, 7897, 0, NULL, NULL} +}; /* compose_tab_2_26 */ +static int hash_compose_tab_2[54] = +{-1,-1,-1,20,-1,-1,-1,21,-1,22,-1,0,23,1,24,2,25,3,4,5,6,-1,-1,-1,-1,7,-1,-1, + -1,8,-1,9,-1,10,-1,11,12,-1,-1,-1,-1,-1,-1,13,-1,14,-1,15,26,16,17,18,19,-1}; /* hash_compose_tab_2 */ +static CompEntry compose_tab_2[] = { +{65, 194, 0, NULL, NULL}, +{67, 264, 0, NULL, NULL}, +{69, 202, 0, NULL, NULL}, +{71, 284, 0, NULL, NULL}, +{72, 292, 0, NULL, NULL}, +{73, 206, 0, NULL, NULL}, +{74, 308, 0, NULL, NULL}, +{79, 212, 0, NULL, NULL}, +{83, 348, 0, NULL, NULL}, +{85, 219, 0, NULL, NULL}, +{87, 372, 0, NULL, NULL}, +{89, 374, 0, NULL, NULL}, +{90, 7824, 0, NULL, NULL}, +{97, 226, 0, NULL, NULL}, +{99, 265, 0, NULL, NULL}, +{101, 234, 0, NULL, NULL}, +{103, 285, 0, NULL, NULL}, +{104, 293, 0, NULL, NULL}, +{105, 238, 0, NULL, NULL}, +{106, 309, 0, NULL, NULL}, +{111, 244, 0, NULL, NULL}, +{115, 349, 0, NULL, NULL}, +{117, 251, 0, NULL, NULL}, +{119, 373, 0, NULL, NULL}, +{121, 375, 0, NULL, NULL}, +{122, 7825, 0, NULL, NULL}, +{803, 0, 6, compose_tab_2_26, hash_compose_tab_2_26} +}; /* compose_tab_2 */ +static int hash_compose_tab_3_16[12] = +{-1,3,-1,5,-1,0,4,2,-1,1,-1,-1}; /* hash_compose_tab_3_16 */ +static CompEntry compose_tab_3_16[] = { +{65, 7850, 0, NULL, NULL}, +{69, 7876, 0, NULL, NULL}, +{79, 7894, 0, NULL, NULL}, +{97, 7851, 0, NULL, NULL}, +{101, 7877, 0, NULL, NULL}, +{111, 7895, 0, NULL, NULL} +}; /* compose_tab_3_16 */ +static int hash_compose_tab_3_17[4] = +{-1,0,1,-1}; /* hash_compose_tab_3_17 */ +static CompEntry compose_tab_3_17[] = { +{65, 7860, 0, NULL, NULL}, +{97, 7861, 0, NULL, NULL} +}; /* compose_tab_3_17 */ +static int hash_compose_tab_3_18[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_3_18 */ +static CompEntry compose_tab_3_18[] = { +{79, 7904, 0, NULL, NULL}, +{85, 7918, 0, NULL, NULL}, +{111, 7905, 0, NULL, NULL}, +{117, 7919, 0, NULL, NULL} +}; /* compose_tab_3_18 */ +static int hash_compose_tab_3[38] = +{-1,-1,3,4,13,14,-1,15,-1,5,6,16,-1,7,17,-1,-1,-1,-1,-1,-1,8,-1,-1,-1,9,-1,0, + -1,10,-1,1,-1,-1,11,2,12,18}; /* hash_compose_tab_3 */ +static CompEntry compose_tab_3[] = { +{65, 195, 0, NULL, NULL}, +{69, 7868, 0, NULL, NULL}, +{73, 296, 0, NULL, NULL}, +{78, 209, 0, NULL, NULL}, +{79, 213, 0, NULL, NULL}, +{85, 360, 0, NULL, NULL}, +{86, 7804, 0, NULL, NULL}, +{89, 7928, 0, NULL, NULL}, +{97, 227, 0, NULL, NULL}, +{101, 7869, 0, NULL, NULL}, +{105, 297, 0, NULL, NULL}, +{110, 241, 0, NULL, NULL}, +{111, 245, 0, NULL, NULL}, +{117, 361, 0, NULL, NULL}, +{118, 7805, 0, NULL, NULL}, +{121, 7929, 0, NULL, NULL}, +{770, 0, 6, compose_tab_3_16, hash_compose_tab_3_16}, +{774, 0, 2, compose_tab_3_17, hash_compose_tab_3_17}, +{795, 0, 4, compose_tab_3_18, hash_compose_tab_3_18} +}; /* compose_tab_3 */ +static int hash_compose_tab_4_14[4] = +{-1,0,1,-1}; /* hash_compose_tab_4_14 */ +static CompEntry compose_tab_4_14[] = { +{65, 480, 0, NULL, NULL}, +{97, 481, 0, NULL, NULL} +}; /* compose_tab_4_14 */ +static int hash_compose_tab_4_15[8] = +{-1,0,2,-1,-1,1,3,-1}; /* hash_compose_tab_4_15 */ +static CompEntry compose_tab_4_15[] = { +{65, 478, 0, NULL, NULL}, +{85, 469, 0, NULL, NULL}, +{97, 479, 0, NULL, NULL}, +{117, 470, 0, NULL, NULL} +}; /* compose_tab_4_15 */ +static int hash_compose_tab_4_16[8] = +{-1,-1,1,3,0,2,-1,-1}; /* hash_compose_tab_4_16 */ +static CompEntry compose_tab_4_16[] = { +{76, 7736, 0, NULL, NULL}, +{82, 7772, 0, NULL, NULL}, +{108, 7737, 0, NULL, NULL}, +{114, 7773, 0, NULL, NULL} +}; /* compose_tab_4_16 */ +static int hash_compose_tab_4_17[4] = +{1,-1,-1,0}; /* hash_compose_tab_4_17 */ +static CompEntry compose_tab_4_17[] = { +{79, 492, 0, NULL, NULL}, +{111, 493, 0, NULL, NULL} +}; /* compose_tab_4_17 */ +static int hash_compose_tab_4[56] = +{-1,22,-1,-1,-1,11,13,-1,-1,0,-1,-1,-1,1,23,2,26,3,18,16,-1,-1,-1,4,17,19,-1, + 27,-1,5,12,-1,-1,-1,-1,-1,-1,20,-1,-1,24,6,-1,-1,-1,7,-1,8,14,9,15,21,25,-1, + -1,10}; /* hash_compose_tab_4 */ +static CompEntry compose_tab_4[] = { +{65, 256, 0, NULL, NULL}, +{69, 274, 0, NULL, NULL}, +{71, 7712, 0, NULL, NULL}, +{73, 298, 0, NULL, NULL}, +{79, 332, 0, NULL, NULL}, +{85, 362, 0, NULL, NULL}, +{97, 257, 0, NULL, NULL}, +{101, 275, 0, NULL, NULL}, +{103, 7713, 0, NULL, NULL}, +{105, 299, 0, NULL, NULL}, +{111, 333, 0, NULL, NULL}, +{117, 363, 0, NULL, NULL}, +{198, 482, 0, NULL, NULL}, +{230, 483, 0, NULL, NULL}, +{775, 0, 2, compose_tab_4_14, hash_compose_tab_4_14}, +{776, 0, 4, compose_tab_4_15, hash_compose_tab_4_15}, +{803, 0, 4, compose_tab_4_16, hash_compose_tab_4_16}, +{808, 0, 2, compose_tab_4_17, hash_compose_tab_4_17}, +{913, 8121, 0, NULL, NULL}, +{921, 8153, 0, NULL, NULL}, +{933, 8169, 0, NULL, NULL}, +{945, 8113, 0, NULL, NULL}, +{953, 8145, 0, NULL, NULL}, +{965, 8161, 0, NULL, NULL}, +{1048, 1250, 0, NULL, NULL}, +{1059, 1262, 0, NULL, NULL}, +{1080, 1251, 0, NULL, NULL}, +{1091, 1263, 0, NULL, NULL} +}; /* compose_tab_4 */ +static int hash_compose_tab_5_12[4] = +{-1,0,1,-1}; /* hash_compose_tab_5_12 */ +static CompEntry compose_tab_5_12[] = { +{65, 7862, 0, NULL, NULL}, +{97, 7863, 0, NULL, NULL} +}; /* compose_tab_5_12 */ +static int hash_compose_tab_5_13[4] = +{-1,0,1,-1}; /* hash_compose_tab_5_13 */ +static CompEntry compose_tab_5_13[] = { +{69, 7708, 0, NULL, NULL}, +{101, 7709, 0, NULL, NULL} +}; /* compose_tab_5_13 */ +static int hash_compose_tab_5[60] = +{28,-1,-1,-1,-1,0,19,-1,-1,1,-1,2,29,3,14,-1,-1,-1,-1,4,20,15,-1,12,-1,5,21, + 13,22,23,-1,-1,-1,16,-1,-1,-1,6,-1,24,-1,7,-1,8,-1,9,17,-1,-1,-1,-1,10,25,18, + -1,-1,-1,11,26,27}; /* hash_compose_tab_5 */ +static CompEntry compose_tab_5[] = { +{65, 258, 0, NULL, NULL}, +{69, 276, 0, NULL, NULL}, +{71, 286, 0, NULL, NULL}, +{73, 300, 0, NULL, NULL}, +{79, 334, 0, NULL, NULL}, +{85, 364, 0, NULL, NULL}, +{97, 259, 0, NULL, NULL}, +{101, 277, 0, NULL, NULL}, +{103, 287, 0, NULL, NULL}, +{105, 301, 0, NULL, NULL}, +{111, 335, 0, NULL, NULL}, +{117, 365, 0, NULL, NULL}, +{803, 0, 2, compose_tab_5_12, hash_compose_tab_5_12}, +{807, 0, 2, compose_tab_5_13, hash_compose_tab_5_13}, +{913, 8120, 0, NULL, NULL}, +{921, 8152, 0, NULL, NULL}, +{933, 8168, 0, NULL, NULL}, +{945, 8112, 0, NULL, NULL}, +{953, 8144, 0, NULL, NULL}, +{965, 8160, 0, NULL, NULL}, +{1040, 1232, 0, NULL, NULL}, +{1045, 1238, 0, NULL, NULL}, +{1046, 1217, 0, NULL, NULL}, +{1048, 1049, 0, NULL, NULL}, +{1059, 1038, 0, NULL, NULL}, +{1072, 1233, 0, NULL, NULL}, +{1077, 1239, 0, NULL, NULL}, +{1078, 1218, 0, NULL, NULL}, +{1080, 1081, 0, NULL, NULL}, +{1091, 1118, 0, NULL, NULL} +}; /* compose_tab_5 */ +static int hash_compose_tab_6_36[4] = +{1,-1,-1,0}; /* hash_compose_tab_6_36 */ +static CompEntry compose_tab_6_36[] = { +{83, 7780, 0, NULL, NULL}, +{115, 7781, 0, NULL, NULL} +}; /* compose_tab_6_36 */ +static int hash_compose_tab_6_38[4] = +{1,-1,-1,0}; /* hash_compose_tab_6_38 */ +static CompEntry compose_tab_6_38[] = { +{83, 7782, 0, NULL, NULL}, +{115, 7783, 0, NULL, NULL} +}; /* compose_tab_6_38 */ +static int hash_compose_tab_6_39[4] = +{1,-1,-1,0}; /* hash_compose_tab_6_39 */ +static CompEntry compose_tab_6_39[] = { +{83, 7784, 0, NULL, NULL}, +{115, 7785, 0, NULL, NULL} +}; /* compose_tab_6_39 */ +static int hash_compose_tab_6[80] = +{10,-1,11,12,13,39,-1,14,15,16,17,-1,-1,-1,-1,-1,-1,-1,18,19,20,21,22,23,24, + -1,-1,-1,-1,25,26,-1,27,-1,28,29,30,-1,-1,31,32,33,34,-1,-1,-1,-1,-1,-1,36, + -1,-1,-1,-1,37,-1,-1,-1,-1,-1,38,-1,-1,35,-1,-1,0,1,2,3,4,5,6,7,-1,-1,-1,8,9, + -1}; /* hash_compose_tab_6 */ +static CompEntry compose_tab_6[] = { +{66, 7682, 0, NULL, NULL}, +{67, 266, 0, NULL, NULL}, +{68, 7690, 0, NULL, NULL}, +{69, 278, 0, NULL, NULL}, +{70, 7710, 0, NULL, NULL}, +{71, 288, 0, NULL, NULL}, +{72, 7714, 0, NULL, NULL}, +{73, 304, 0, NULL, NULL}, +{77, 7744, 0, NULL, NULL}, +{78, 7748, 0, NULL, NULL}, +{80, 7766, 0, NULL, NULL}, +{82, 7768, 0, NULL, NULL}, +{83, 7776, 0, NULL, NULL}, +{84, 7786, 0, NULL, NULL}, +{87, 7814, 0, NULL, NULL}, +{88, 7818, 0, NULL, NULL}, +{89, 7822, 0, NULL, NULL}, +{90, 379, 0, NULL, NULL}, +{98, 7683, 0, NULL, NULL}, +{99, 267, 0, NULL, NULL}, +{100, 7691, 0, NULL, NULL}, +{101, 279, 0, NULL, NULL}, +{102, 7711, 0, NULL, NULL}, +{103, 289, 0, NULL, NULL}, +{104, 7715, 0, NULL, NULL}, +{109, 7745, 0, NULL, NULL}, +{110, 7749, 0, NULL, NULL}, +{112, 7767, 0, NULL, NULL}, +{114, 7769, 0, NULL, NULL}, +{115, 7777, 0, NULL, NULL}, +{116, 7787, 0, NULL, NULL}, +{119, 7815, 0, NULL, NULL}, +{120, 7819, 0, NULL, NULL}, +{121, 7823, 0, NULL, NULL}, +{122, 380, 0, NULL, NULL}, +{383, 7835, 0, NULL, NULL}, +{769, 0, 2, compose_tab_6_36, hash_compose_tab_6_36}, +{774, 784, 0, NULL, NULL}, +{780, 0, 2, compose_tab_6_38, hash_compose_tab_6_38}, +{803, 0, 2, compose_tab_6_39, hash_compose_tab_6_39} +}; /* compose_tab_6 */ +static int hash_compose_tab_7_23[4] = +{1,-1,-1,0}; /* hash_compose_tab_7_23 */ +static CompEntry compose_tab_7_23[] = { +{79, 7758, 0, NULL, NULL}, +{111, 7759, 0, NULL, NULL} +}; /* compose_tab_7_23 */ +static int hash_compose_tab_7_24[4] = +{-1,0,1,-1}; /* hash_compose_tab_7_24 */ +static CompEntry compose_tab_7_24[] = { +{85, 7802, 0, NULL, NULL}, +{117, 7803, 0, NULL, NULL} +}; /* compose_tab_7_24 */ +static int hash_compose_tab_7[100] = +{48,10,21,-1,11,12,-1,-1,-1,-1,49,13,-1,-1,-1,20,14,15,-1,16,17,18,25,-1,-1, + -1,-1,-1,-1,22,30,-1,-1,26,-1,-1,-1,-1,-1,-1,31,-1,-1,-1,-1,32,33,34,35,-1, + -1,-1,-1,27,36,-1,-1,-1,-1,37,-1,-1,-1,38,-1,0,28,39,-1,1,-1,23,2,3,24,40,-1, + 41,29,4,42,43,44,-1,-1,5,45,6,7,8,-1,46,-1,-1,-1,47,-1,9,-1,19}; /* hash_compose_tab_7 */ +static CompEntry compose_tab_7[] = { +{65, 196, 0, NULL, NULL}, +{69, 203, 0, NULL, NULL}, +{72, 7718, 0, NULL, NULL}, +{73, 207, 0, NULL, NULL}, +{79, 214, 0, NULL, NULL}, +{85, 220, 0, NULL, NULL}, +{87, 7812, 0, NULL, NULL}, +{88, 7820, 0, NULL, NULL}, +{89, 376, 0, NULL, NULL}, +{97, 228, 0, NULL, NULL}, +{101, 235, 0, NULL, NULL}, +{104, 7719, 0, NULL, NULL}, +{105, 239, 0, NULL, NULL}, +{111, 246, 0, NULL, NULL}, +{116, 7831, 0, NULL, NULL}, +{117, 252, 0, NULL, NULL}, +{119, 7813, 0, NULL, NULL}, +{120, 7821, 0, NULL, NULL}, +{121, 255, 0, NULL, NULL}, +{399, 1242, 0, NULL, NULL}, +{415, 1258, 0, NULL, NULL}, +{601, 1243, 0, NULL, NULL}, +{629, 1259, 0, NULL, NULL}, +{771, 0, 2, compose_tab_7_23, hash_compose_tab_7_23}, +{772, 0, 2, compose_tab_7_24, hash_compose_tab_7_24}, +{921, 938, 0, NULL, NULL}, +{933, 939, 0, NULL, NULL}, +{953, 970, 0, NULL, NULL}, +{965, 971, 0, NULL, NULL}, +{978, 980, 0, NULL, NULL}, +{1030, 1031, 0, NULL, NULL}, +{1040, 1234, 0, NULL, NULL}, +{1045, 1025, 0, NULL, NULL}, +{1046, 1244, 0, NULL, NULL}, +{1047, 1246, 0, NULL, NULL}, +{1048, 1252, 0, NULL, NULL}, +{1054, 1254, 0, NULL, NULL}, +{1059, 1264, 0, NULL, NULL}, +{1063, 1268, 0, NULL, NULL}, +{1067, 1272, 0, NULL, NULL}, +{1072, 1235, 0, NULL, NULL}, +{1077, 1105, 0, NULL, NULL}, +{1078, 1245, 0, NULL, NULL}, +{1079, 1247, 0, NULL, NULL}, +{1080, 1253, 0, NULL, NULL}, +{1086, 1255, 0, NULL, NULL}, +{1091, 1265, 0, NULL, NULL}, +{1095, 1269, 0, NULL, NULL}, +{1099, 1273, 0, NULL, NULL}, +{1110, 1111, 0, NULL, NULL} +}; /* compose_tab_7 */ +static int hash_compose_tab_8_12[12] = +{-1,3,-1,5,-1,0,4,2,-1,1,-1,-1}; /* hash_compose_tab_8_12 */ +static CompEntry compose_tab_8_12[] = { +{65, 7848, 0, NULL, NULL}, +{69, 7874, 0, NULL, NULL}, +{79, 7892, 0, NULL, NULL}, +{97, 7849, 0, NULL, NULL}, +{101, 7875, 0, NULL, NULL}, +{111, 7893, 0, NULL, NULL} +}; /* compose_tab_8_12 */ +static int hash_compose_tab_8_13[4] = +{-1,0,1,-1}; /* hash_compose_tab_8_13 */ +static CompEntry compose_tab_8_13[] = { +{65, 7858, 0, NULL, NULL}, +{97, 7859, 0, NULL, NULL} +}; /* compose_tab_8_13 */ +static int hash_compose_tab_8_14[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_8_14 */ +static CompEntry compose_tab_8_14[] = { +{79, 7902, 0, NULL, NULL}, +{85, 7916, 0, NULL, NULL}, +{111, 7903, 0, NULL, NULL}, +{117, 7917, 0, NULL, NULL} +}; /* compose_tab_8_14 */ +static int hash_compose_tab_8[30] = +{-1,11,-1,-1,-1,0,-1,6,-1,1,-1,7,-1,2,-1,8,14,-1,-1,3,12,9,-1,-1,13,4,-1,10, + -1,5}; /* hash_compose_tab_8 */ +static CompEntry compose_tab_8[] = { +{65, 7842, 0, NULL, NULL}, +{69, 7866, 0, NULL, NULL}, +{73, 7880, 0, NULL, NULL}, +{79, 7886, 0, NULL, NULL}, +{85, 7910, 0, NULL, NULL}, +{89, 7926, 0, NULL, NULL}, +{97, 7843, 0, NULL, NULL}, +{101, 7867, 0, NULL, NULL}, +{105, 7881, 0, NULL, NULL}, +{111, 7887, 0, NULL, NULL}, +{117, 7911, 0, NULL, NULL}, +{121, 7927, 0, NULL, NULL}, +{770, 0, 6, compose_tab_8_12, hash_compose_tab_8_12}, +{774, 0, 2, compose_tab_8_13, hash_compose_tab_8_13}, +{795, 0, 4, compose_tab_8_14, hash_compose_tab_8_14} +}; /* compose_tab_8 */ +static int hash_compose_tab_9[12] = +{-1,1,2,5,-1,0,-1,-1,-1,3,-1,4}; /* hash_compose_tab_9 */ +static CompEntry compose_tab_9[] = { +{65, 197, 0, NULL, NULL}, +{85, 366, 0, NULL, NULL}, +{97, 229, 0, NULL, NULL}, +{117, 367, 0, NULL, NULL}, +{119, 7832, 0, NULL, NULL}, +{121, 7833, 0, NULL, NULL} +}; /* compose_tab_9 */ +static int hash_compose_tab_10[12] = +{-1,1,-1,2,4,-1,-1,0,-1,3,-1,5}; /* hash_compose_tab_10 */ +static CompEntry compose_tab_10[] = { +{79, 336, 0, NULL, NULL}, +{85, 368, 0, NULL, NULL}, +{111, 337, 0, NULL, NULL}, +{117, 369, 0, NULL, NULL}, +{1059, 1266, 0, NULL, NULL}, +{1091, 1267, 0, NULL, NULL} +}; /* compose_tab_10 */ +static int hash_compose_tab_11_33[4] = +{-1,0,1,-1}; /* hash_compose_tab_11_33 */ +static CompEntry compose_tab_11_33[] = { +{85, 473, 0, NULL, NULL}, +{117, 474, 0, NULL, NULL} +}; /* compose_tab_11_33 */ +static int hash_compose_tab_11[68] = +{2,3,-1,4,-1,5,-1,6,7,-1,8,9,-1,-1,10,11,12,13,-1,-1,-1,-1,14,-1,-1,-1,-1,-1, + 33,15,-1,16,17,18,31,19,-1,20,21,22,23,-1,24,25,-1,-1,26,27,28,29,32,-1,-1, + -1,30,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,0,-1,1}; /* hash_compose_tab_11 */ +static CompEntry compose_tab_11[] = { +{65, 461, 0, NULL, NULL}, +{67, 268, 0, NULL, NULL}, +{68, 270, 0, NULL, NULL}, +{69, 282, 0, NULL, NULL}, +{71, 486, 0, NULL, NULL}, +{73, 463, 0, NULL, NULL}, +{75, 488, 0, NULL, NULL}, +{76, 317, 0, NULL, NULL}, +{78, 327, 0, NULL, NULL}, +{79, 465, 0, NULL, NULL}, +{82, 344, 0, NULL, NULL}, +{83, 352, 0, NULL, NULL}, +{84, 356, 0, NULL, NULL}, +{85, 467, 0, NULL, NULL}, +{90, 381, 0, NULL, NULL}, +{97, 462, 0, NULL, NULL}, +{99, 269, 0, NULL, NULL}, +{100, 271, 0, NULL, NULL}, +{101, 283, 0, NULL, NULL}, +{103, 487, 0, NULL, NULL}, +{105, 464, 0, NULL, NULL}, +{106, 496, 0, NULL, NULL}, +{107, 489, 0, NULL, NULL}, +{108, 318, 0, NULL, NULL}, +{110, 328, 0, NULL, NULL}, +{111, 466, 0, NULL, NULL}, +{114, 345, 0, NULL, NULL}, +{115, 353, 0, NULL, NULL}, +{116, 357, 0, NULL, NULL}, +{117, 468, 0, NULL, NULL}, +{122, 382, 0, NULL, NULL}, +{439, 494, 0, NULL, NULL}, +{658, 495, 0, NULL, NULL}, +{776, 0, 2, compose_tab_11_33, hash_compose_tab_11_33} +}; /* compose_tab_11 */ +static int hash_compose_tab_12_1[4] = +{-1,0,1,-1}; /* hash_compose_tab_12_1 */ +static CompEntry compose_tab_12_1[] = { +{953, 912, 0, NULL, NULL}, +{965, 944, 0, NULL, NULL} +}; /* compose_tab_12_1 */ +static int hash_compose_tab_12[34] = +{11,4,12,5,-1,-1,-1,13,-1,6,-1,-1,-1,14,-1,7,-1,15,-1,8,-1,-1,-1,-1,-1,-1,16, + 9,1,2,-1,10,0,3}; /* hash_compose_tab_12 */ +static CompEntry compose_tab_12[] = { +{168, 901, 0, NULL, NULL}, +{776, 0, 2, compose_tab_12_1, hash_compose_tab_12_1}, +{913, 902, 0, NULL, NULL}, +{917, 904, 0, NULL, NULL}, +{919, 905, 0, NULL, NULL}, +{921, 906, 0, NULL, NULL}, +{927, 908, 0, NULL, NULL}, +{933, 910, 0, NULL, NULL}, +{937, 911, 0, NULL, NULL}, +{945, 940, 0, NULL, NULL}, +{949, 941, 0, NULL, NULL}, +{951, 942, 0, NULL, NULL}, +{953, 943, 0, NULL, NULL}, +{959, 972, 0, NULL, NULL}, +{965, 973, 0, NULL, NULL}, +{969, 974, 0, NULL, NULL}, +{978, 979, 0, NULL, NULL} +}; /* compose_tab_12 */ +static int hash_compose_tab_13[28] = +{-1,5,10,-1,-1,11,-1,-1,-1,0,-1,-1,-1,1,6,-1,-1,2,7,-1,12,8,13,3,-1,-1,4,9}; /* hash_compose_tab_13 */ +static CompEntry compose_tab_13[] = { +{65, 512, 0, NULL, NULL}, +{69, 516, 0, NULL, NULL}, +{73, 520, 0, NULL, NULL}, +{79, 524, 0, NULL, NULL}, +{82, 528, 0, NULL, NULL}, +{85, 532, 0, NULL, NULL}, +{97, 513, 0, NULL, NULL}, +{101, 517, 0, NULL, NULL}, +{105, 521, 0, NULL, NULL}, +{111, 525, 0, NULL, NULL}, +{114, 529, 0, NULL, NULL}, +{117, 533, 0, NULL, NULL}, +{1140, 1142, 0, NULL, NULL}, +{1141, 1143, 0, NULL, NULL} +}; /* compose_tab_13 */ +static int hash_compose_tab_14[24] = +{-1,2,6,-1,-1,7,-1,3,-1,8,4,-1,-1,5,-1,9,-1,0,10,-1,-1,1,11,-1}; /* hash_compose_tab_14 */ +static CompEntry compose_tab_14[] = { +{65, 514, 0, NULL, NULL}, +{69, 518, 0, NULL, NULL}, +{73, 522, 0, NULL, NULL}, +{79, 526, 0, NULL, NULL}, +{82, 530, 0, NULL, NULL}, +{85, 534, 0, NULL, NULL}, +{97, 515, 0, NULL, NULL}, +{101, 519, 0, NULL, NULL}, +{105, 523, 0, NULL, NULL}, +{111, 527, 0, NULL, NULL}, +{114, 531, 0, NULL, NULL}, +{117, 535, 0, NULL, NULL} +}; /* compose_tab_14 */ +static int hash_compose_tab_15_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_15_0 */ +static CompEntry compose_tab_15_0[] = { +{913, 8072, 0, NULL, NULL}, +{919, 8088, 0, NULL, NULL}, +{937, 8104, 0, NULL, NULL}, +{945, 8064, 0, NULL, NULL}, +{951, 8080, 0, NULL, NULL}, +{969, 8096, 0, NULL, NULL} +}; /* compose_tab_15_0 */ +static int hash_compose_tab_15[30] = +{-1,12,-1,-1,-1,13,-1,6,-1,14,-1,-1,-1,1,-1,7,-1,2,-1,3,8,4,9,10,-1,-1,-1,0,5, + 11}; /* hash_compose_tab_15 */ +static CompEntry compose_tab_15[] = { +{837, 0, 6, compose_tab_15_0, hash_compose_tab_15_0}, +{913, 7944, 0, NULL, NULL}, +{917, 7960, 0, NULL, NULL}, +{919, 7976, 0, NULL, NULL}, +{921, 7992, 0, NULL, NULL}, +{927, 8008, 0, NULL, NULL}, +{937, 8040, 0, NULL, NULL}, +{945, 7936, 0, NULL, NULL}, +{949, 7952, 0, NULL, NULL}, +{951, 7968, 0, NULL, NULL}, +{953, 7984, 0, NULL, NULL}, +{959, 8000, 0, NULL, NULL}, +{961, 8164, 0, NULL, NULL}, +{965, 8016, 0, NULL, NULL}, +{969, 8032, 0, NULL, NULL} +}; /* compose_tab_15 */ +static int hash_compose_tab_16_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_16_0 */ +static CompEntry compose_tab_16_0[] = { +{913, 8073, 0, NULL, NULL}, +{919, 8089, 0, NULL, NULL}, +{937, 8105, 0, NULL, NULL}, +{945, 8065, 0, NULL, NULL}, +{951, 8081, 0, NULL, NULL}, +{969, 8097, 0, NULL, NULL} +}; /* compose_tab_16_0 */ +static int hash_compose_tab_16[34] = +{11,3,12,4,-1,-1,-1,13,-1,5,14,6,-1,15,-1,7,-1,16,-1,8,-1,0,-1,-1,-1,-1,-1,9, + -1,1,-1,10,-1,2}; /* hash_compose_tab_16 */ +static CompEntry compose_tab_16[] = { +{837, 0, 6, compose_tab_16_0, hash_compose_tab_16_0}, +{913, 7945, 0, NULL, NULL}, +{917, 7961, 0, NULL, NULL}, +{919, 7977, 0, NULL, NULL}, +{921, 7993, 0, NULL, NULL}, +{927, 8009, 0, NULL, NULL}, +{929, 8172, 0, NULL, NULL}, +{933, 8025, 0, NULL, NULL}, +{937, 8041, 0, NULL, NULL}, +{945, 7937, 0, NULL, NULL}, +{949, 7953, 0, NULL, NULL}, +{951, 7969, 0, NULL, NULL}, +{953, 7985, 0, NULL, NULL}, +{959, 8001, 0, NULL, NULL}, +{961, 8165, 0, NULL, NULL}, +{965, 8017, 0, NULL, NULL}, +{969, 8033, 0, NULL, NULL} +}; /* compose_tab_16 */ +static int hash_compose_tab_17[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_17 */ +static CompEntry compose_tab_17[] = { +{79, 416, 0, NULL, NULL}, +{85, 431, 0, NULL, NULL}, +{111, 417, 0, NULL, NULL}, +{117, 432, 0, NULL, NULL} +}; /* compose_tab_17 */ +static int hash_compose_tab_18_38[8] = +{2,-1,-1,-1,-1,1,3,0}; /* hash_compose_tab_18_38 */ +static CompEntry compose_tab_18_38[] = { +{79, 7906, 0, NULL, NULL}, +{85, 7920, 0, NULL, NULL}, +{111, 7907, 0, NULL, NULL}, +{117, 7921, 0, NULL, NULL} +}; /* compose_tab_18_38 */ +static int hash_compose_tab_18[78] = +{9,10,-1,-1,11,12,13,14,15,16,-1,17,18,-1,-1,38,-1,-1,-1,19,20,-1,21,22,-1,-1, + 23,24,-1,25,26,27,28,29,-1,-1,30,31,32,33,34,35,-1,36,37,-1,-1,-1,-1,-1,-1, + -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,0,1,-1,2,3,-1,-1,4,5,-1,6,7,8}; /* hash_compose_tab_18 */ +static CompEntry compose_tab_18[] = { +{65, 7840, 0, NULL, NULL}, +{66, 7684, 0, NULL, NULL}, +{68, 7692, 0, NULL, NULL}, +{69, 7864, 0, NULL, NULL}, +{72, 7716, 0, NULL, NULL}, +{73, 7882, 0, NULL, NULL}, +{75, 7730, 0, NULL, NULL}, +{76, 7734, 0, NULL, NULL}, +{77, 7746, 0, NULL, NULL}, +{78, 7750, 0, NULL, NULL}, +{79, 7884, 0, NULL, NULL}, +{82, 7770, 0, NULL, NULL}, +{83, 7778, 0, NULL, NULL}, +{84, 7788, 0, NULL, NULL}, +{85, 7908, 0, NULL, NULL}, +{86, 7806, 0, NULL, NULL}, +{87, 7816, 0, NULL, NULL}, +{89, 7924, 0, NULL, NULL}, +{90, 7826, 0, NULL, NULL}, +{97, 7841, 0, NULL, NULL}, +{98, 7685, 0, NULL, NULL}, +{100, 7693, 0, NULL, NULL}, +{101, 7865, 0, NULL, NULL}, +{104, 7717, 0, NULL, NULL}, +{105, 7883, 0, NULL, NULL}, +{107, 7731, 0, NULL, NULL}, +{108, 7735, 0, NULL, NULL}, +{109, 7747, 0, NULL, NULL}, +{110, 7751, 0, NULL, NULL}, +{111, 7885, 0, NULL, NULL}, +{114, 7771, 0, NULL, NULL}, +{115, 7779, 0, NULL, NULL}, +{116, 7789, 0, NULL, NULL}, +{117, 7909, 0, NULL, NULL}, +{118, 7807, 0, NULL, NULL}, +{119, 7817, 0, NULL, NULL}, +{121, 7925, 0, NULL, NULL}, +{122, 7827, 0, NULL, NULL}, +{795, 0, 4, compose_tab_18_38, hash_compose_tab_18_38} +}; /* compose_tab_18 */ +static int hash_compose_tab_19[4] = +{-1,0,1,-1}; /* hash_compose_tab_19 */ +static CompEntry compose_tab_19[] = { +{85, 7794, 0, NULL, NULL}, +{117, 7795, 0, NULL, NULL} +}; /* compose_tab_19 */ +static int hash_compose_tab_20[4] = +{-1,0,1,-1}; /* hash_compose_tab_20 */ +static CompEntry compose_tab_20[] = { +{65, 7680, 0, NULL, NULL}, +{97, 7681, 0, NULL, NULL} +}; /* compose_tab_20 */ +static int hash_compose_tab_21[40] = +{-1,-1,7,8,9,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,10,11,-1,-1,12,13,-1, + -1,0,1,14,15,2,3,16,17,4,5,18,6,19}; /* hash_compose_tab_21 */ +static CompEntry compose_tab_21[] = { +{67, 199, 0, NULL, NULL}, +{68, 7696, 0, NULL, NULL}, +{71, 290, 0, NULL, NULL}, +{72, 7720, 0, NULL, NULL}, +{75, 310, 0, NULL, NULL}, +{76, 315, 0, NULL, NULL}, +{78, 325, 0, NULL, NULL}, +{82, 342, 0, NULL, NULL}, +{83, 350, 0, NULL, NULL}, +{84, 354, 0, NULL, NULL}, +{99, 231, 0, NULL, NULL}, +{100, 7697, 0, NULL, NULL}, +{103, 291, 0, NULL, NULL}, +{104, 7721, 0, NULL, NULL}, +{107, 311, 0, NULL, NULL}, +{108, 316, 0, NULL, NULL}, +{110, 326, 0, NULL, NULL}, +{114, 343, 0, NULL, NULL}, +{115, 351, 0, NULL, NULL}, +{116, 355, 0, NULL, NULL} +}; /* compose_tab_21 */ +static int hash_compose_tab_22[20] = +{-1,6,-1,-1,-1,0,4,7,-1,1,-1,8,-1,2,-1,-1,-1,5,9,3}; /* hash_compose_tab_22 */ +static CompEntry compose_tab_22[] = { +{65, 260, 0, NULL, NULL}, +{69, 280, 0, NULL, NULL}, +{73, 302, 0, NULL, NULL}, +{79, 490, 0, NULL, NULL}, +{85, 370, 0, NULL, NULL}, +{97, 261, 0, NULL, NULL}, +{101, 281, 0, NULL, NULL}, +{105, 303, 0, NULL, NULL}, +{111, 491, 0, NULL, NULL}, +{117, 371, 0, NULL, NULL} +}; /* compose_tab_22 */ +static int hash_compose_tab_23[24] = +{-1,-1,-1,-1,2,6,3,7,-1,-1,-1,-1,4,5,8,9,-1,-1,-1,-1,0,1,10,11}; /* hash_compose_tab_23 */ +static CompEntry compose_tab_23[] = { +{68, 7698, 0, NULL, NULL}, +{69, 7704, 0, NULL, NULL}, +{76, 7740, 0, NULL, NULL}, +{78, 7754, 0, NULL, NULL}, +{84, 7792, 0, NULL, NULL}, +{85, 7798, 0, NULL, NULL}, +{100, 7699, 0, NULL, NULL}, +{101, 7705, 0, NULL, NULL}, +{108, 7741, 0, NULL, NULL}, +{110, 7755, 0, NULL, NULL}, +{116, 7793, 0, NULL, NULL}, +{117, 7799, 0, NULL, NULL} +}; /* compose_tab_23 */ +static int hash_compose_tab_24[4] = +{0,1,-1,-1}; /* hash_compose_tab_24 */ +static CompEntry compose_tab_24[] = { +{72, 7722, 0, NULL, NULL}, +{104, 7723, 0, NULL, NULL} +}; /* compose_tab_24 */ +static int hash_compose_tab_25[12] = +{-1,1,2,-1,-1,3,-1,-1,-1,0,4,5}; /* hash_compose_tab_25 */ +static CompEntry compose_tab_25[] = { +{69, 7706, 0, NULL, NULL}, +{73, 7724, 0, NULL, NULL}, +{85, 7796, 0, NULL, NULL}, +{101, 7707, 0, NULL, NULL}, +{105, 7725, 0, NULL, NULL}, +{117, 7797, 0, NULL, NULL} +}; /* compose_tab_25 */ +static int hash_compose_tab_26[34] = +{1,-1,10,-1,-1,11,12,2,3,13,4,-1,14,-1,5,15,6,-1,-1,-1,16,-1,7,-1,-1,-1,-1,-1, + -1,-1,8,-1,0,9}; /* hash_compose_tab_26 */ +static CompEntry compose_tab_26[] = { +{66, 7686, 0, NULL, NULL}, +{68, 7694, 0, NULL, NULL}, +{75, 7732, 0, NULL, NULL}, +{76, 7738, 0, NULL, NULL}, +{78, 7752, 0, NULL, NULL}, +{82, 7774, 0, NULL, NULL}, +{84, 7790, 0, NULL, NULL}, +{90, 7828, 0, NULL, NULL}, +{98, 7687, 0, NULL, NULL}, +{100, 7695, 0, NULL, NULL}, +{104, 7830, 0, NULL, NULL}, +{107, 7733, 0, NULL, NULL}, +{108, 7739, 0, NULL, NULL}, +{110, 7753, 0, NULL, NULL}, +{114, 7775, 0, NULL, NULL}, +{116, 7791, 0, NULL, NULL}, +{122, 7829, 0, NULL, NULL} +}; /* compose_tab_26 */ +static int hash_compose_tab_27_1[4] = +{-1,0,1,-1}; /* hash_compose_tab_27_1 */ +static CompEntry compose_tab_27_1[] = { +{953, 8151, 0, NULL, NULL}, +{965, 8167, 0, NULL, NULL} +}; /* compose_tab_27_1 */ +static int hash_compose_tab_27_2_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_27_2_0 */ +static CompEntry compose_tab_27_2_0[] = { +{913, 8078, 0, NULL, NULL}, +{919, 8094, 0, NULL, NULL}, +{937, 8110, 0, NULL, NULL}, +{945, 8070, 0, NULL, NULL}, +{951, 8086, 0, NULL, NULL}, +{969, 8102, 0, NULL, NULL} +}; /* compose_tab_27_2_0 */ +static int hash_compose_tab_27_2[20] = +{-1,3,-1,-1,-1,5,8,-1,-1,9,-1,6,-1,1,7,-1,-1,0,4,2}; /* hash_compose_tab_27_2 */ +static CompEntry compose_tab_27_2[] = { +{837, 0, 6, compose_tab_27_2_0, hash_compose_tab_27_2_0}, +{913, 7950, 0, NULL, NULL}, +{919, 7982, 0, NULL, NULL}, +{921, 7998, 0, NULL, NULL}, +{937, 8046, 0, NULL, NULL}, +{945, 7942, 0, NULL, NULL}, +{951, 7974, 0, NULL, NULL}, +{953, 7990, 0, NULL, NULL}, +{965, 8022, 0, NULL, NULL}, +{969, 8038, 0, NULL, NULL} +}; /* compose_tab_27_2 */ +static int hash_compose_tab_27_3_0[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_27_3_0 */ +static CompEntry compose_tab_27_3_0[] = { +{913, 8079, 0, NULL, NULL}, +{919, 8095, 0, NULL, NULL}, +{937, 8111, 0, NULL, NULL}, +{945, 8071, 0, NULL, NULL}, +{951, 8087, 0, NULL, NULL}, +{969, 8103, 0, NULL, NULL} +}; /* compose_tab_27_3_0 */ +static int hash_compose_tab_27_3[22] = +{-1,0,10,-1,-1,7,-1,8,-1,4,-1,1,-1,5,-1,-1,-1,2,-1,3,9,6}; /* hash_compose_tab_27_3 */ +static CompEntry compose_tab_27_3[] = { +{837, 0, 6, compose_tab_27_3_0, hash_compose_tab_27_3_0}, +{913, 7951, 0, NULL, NULL}, +{919, 7983, 0, NULL, NULL}, +{921, 7999, 0, NULL, NULL}, +{933, 8031, 0, NULL, NULL}, +{937, 8047, 0, NULL, NULL}, +{945, 7943, 0, NULL, NULL}, +{951, 7975, 0, NULL, NULL}, +{953, 7991, 0, NULL, NULL}, +{965, 8023, 0, NULL, NULL}, +{969, 8039, 0, NULL, NULL} +}; /* compose_tab_27_3 */ +static int hash_compose_tab_27_4[6] = +{-1,-1,-1,0,1,2}; /* hash_compose_tab_27_4 */ +static CompEntry compose_tab_27_4[] = { +{945, 8119, 0, NULL, NULL}, +{951, 8135, 0, NULL, NULL}, +{969, 8183, 0, NULL, NULL} +}; /* compose_tab_27_4 */ +static int hash_compose_tab_27[24] = +{0,-1,-1,-1,-1,8,11,-1,1,5,9,-1,-1,-1,-1,6,10,7,-1,2,3,4,-1,-1}; /* hash_compose_tab_27 */ +static CompEntry compose_tab_27[] = { +{168, 8129, 0, NULL, NULL}, +{776, 0, 2, compose_tab_27_1, hash_compose_tab_27_1}, +{787, 0, 10, compose_tab_27_2, hash_compose_tab_27_2}, +{788, 0, 11, compose_tab_27_3, hash_compose_tab_27_3}, +{837, 0, 3, compose_tab_27_4, hash_compose_tab_27_4}, +{945, 8118, 0, NULL, NULL}, +{951, 8134, 0, NULL, NULL}, +{953, 8150, 0, NULL, NULL}, +{965, 8166, 0, NULL, NULL}, +{969, 8182, 0, NULL, NULL}, +{8127, 8143, 0, NULL, NULL}, +{8190, 8159, 0, NULL, NULL} +}; /* compose_tab_27 */ +static int hash_compose_tab_28[12] = +{-1,0,2,4,-1,-1,-1,1,-1,3,5,-1}; /* hash_compose_tab_28 */ +static CompEntry compose_tab_28[] = { +{913, 8124, 0, NULL, NULL}, +{919, 8140, 0, NULL, NULL}, +{937, 8188, 0, NULL, NULL}, +{945, 8115, 0, NULL, NULL}, +{951, 8131, 0, NULL, NULL}, +{969, 8179, 0, NULL, NULL} +}; /* compose_tab_28 */ +static int hash_compose_tab_29[4] = +{0,-1,1,-1}; /* hash_compose_tab_29 */ +static CompEntry compose_tab_29[] = { +{1488, 64302, 0, NULL, NULL}, +{1522, 64287, 0, NULL, NULL} +}; /* compose_tab_29 */ +static int hash_compose_tab_30[2] = +{0,-1}; /* hash_compose_tab_30 */ +static CompEntry compose_tab_30[] = { +{1488, 64303, 0, NULL, NULL} +}; /* compose_tab_30 */ +static int hash_compose_tab_31[2] = +{-1,0}; /* hash_compose_tab_31 */ +static CompEntry compose_tab_31[] = { +{1493, 64331, 0, NULL, NULL} +}; /* compose_tab_31 */ +static int hash_compose_tab_32[44] = +{7,8,9,10,11,-1,12,-1,13,14,-1,15,16,-1,17,18,19,20,21,-1,-1,-1,-1,-1,-1,-1, + -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,0,1,2,3,4,5,6,-1}; /* hash_compose_tab_32 */ +static CompEntry compose_tab_32[] = { +{1488, 64304, 0, NULL, NULL}, +{1489, 64305, 0, NULL, NULL}, +{1490, 64306, 0, NULL, NULL}, +{1491, 64307, 0, NULL, NULL}, +{1492, 64308, 0, NULL, NULL}, +{1493, 64309, 0, NULL, NULL}, +{1494, 64310, 0, NULL, NULL}, +{1496, 64312, 0, NULL, NULL}, +{1497, 64313, 0, NULL, NULL}, +{1498, 64314, 0, NULL, NULL}, +{1499, 64315, 0, NULL, NULL}, +{1500, 64316, 0, NULL, NULL}, +{1502, 64318, 0, NULL, NULL}, +{1504, 64320, 0, NULL, NULL}, +{1505, 64321, 0, NULL, NULL}, +{1507, 64323, 0, NULL, NULL}, +{1508, 64324, 0, NULL, NULL}, +{1510, 64326, 0, NULL, NULL}, +{1511, 64327, 0, NULL, NULL}, +{1512, 64328, 0, NULL, NULL}, +{1513, 64329, 0, NULL, NULL}, +{1514, 64330, 0, NULL, NULL} +}; /* compose_tab_32 */ +static int hash_compose_tab_33[6] = +{-1,0,2,-1,-1,1}; /* hash_compose_tab_33 */ +static CompEntry compose_tab_33[] = { +{1489, 64332, 0, NULL, NULL}, +{1499, 64333, 0, NULL, NULL}, +{1508, 64334, 0, NULL, NULL} +}; /* compose_tab_33 */ +static int hash_compose_tab_34_0[2] = +{-1,0}; /* hash_compose_tab_34_0 */ +static CompEntry compose_tab_34_0[] = { +{1513, 64300, 0, NULL, NULL} +}; /* compose_tab_34_0 */ +static int hash_compose_tab_34[4] = +{0,1,-1,-1}; /* hash_compose_tab_34 */ +static CompEntry compose_tab_34[] = { +{1468, 0, 1, compose_tab_34_0, hash_compose_tab_34_0}, +{1513, 64298, 0, NULL, NULL} +}; /* compose_tab_34 */ +static int hash_compose_tab_35_0[2] = +{-1,0}; /* hash_compose_tab_35_0 */ +static CompEntry compose_tab_35_0[] = { +{1513, 64301, 0, NULL, NULL} +}; /* compose_tab_35_0 */ +static int hash_compose_tab_35[4] = +{0,1,-1,-1}; /* hash_compose_tab_35 */ +static CompEntry compose_tab_35[] = { +{1468, 0, 1, compose_tab_35_0, hash_compose_tab_35_0}, +{1513, 64299, 0, NULL, NULL} +}; /* compose_tab_35 */ +static int hash_compose_tab_36[22] = +{3,10,-1,-1,-1,4,5,-1,-1,-1,-1,-1,6,-1,-1,0,1,2,7,8,9,-1}; /* hash_compose_tab_36 */ +static CompEntry compose_tab_36[] = { +{2325, 2392, 0, NULL, NULL}, +{2326, 2393, 0, NULL, NULL}, +{2327, 2394, 0, NULL, NULL}, +{2332, 2395, 0, NULL, NULL}, +{2337, 2396, 0, NULL, NULL}, +{2338, 2397, 0, NULL, NULL}, +{2344, 2345, 0, NULL, NULL}, +{2347, 2398, 0, NULL, NULL}, +{2351, 2399, 0, NULL, NULL}, +{2352, 2353, 0, NULL, NULL}, +{2355, 2356, 0, NULL, NULL} +}; /* compose_tab_36 */ +static int hash_compose_tab_37[8] = +{-1,0,1,-1,2,-1,-1,3}; /* hash_compose_tab_37 */ +static CompEntry compose_tab_37[] = { +{2465, 2524, 0, NULL, NULL}, +{2466, 2525, 0, NULL, NULL}, +{2476, 2480, 0, NULL, NULL}, +{2479, 2527, 0, NULL, NULL} +}; /* compose_tab_37 */ +static int hash_compose_tab_38[2] = +{-1,0}; /* hash_compose_tab_38 */ +static CompEntry compose_tab_38[] = { +{2503, 2507, 0, NULL, NULL} +}; /* compose_tab_38 */ +static int hash_compose_tab_39[2] = +{-1,0}; /* hash_compose_tab_39 */ +static CompEntry compose_tab_39[] = { +{2503, 2508, 0, NULL, NULL} +}; /* compose_tab_39 */ +static int hash_compose_tab_40[10] = +{-1,-1,0,1,3,4,-1,-1,2,-1}; /* hash_compose_tab_40 */ +static CompEntry compose_tab_40[] = { +{2582, 2649, 0, NULL, NULL}, +{2583, 2650, 0, NULL, NULL}, +{2588, 2651, 0, NULL, NULL}, +{2593, 2652, 0, NULL, NULL}, +{2603, 2654, 0, NULL, NULL} +}; /* compose_tab_40 */ +static int hash_compose_tab_41[6] = +{1,2,-1,-1,-1,0}; /* hash_compose_tab_41 */ +static CompEntry compose_tab_41[] = { +{2849, 2908, 0, NULL, NULL}, +{2850, 2909, 0, NULL, NULL}, +{2863, 2911, 0, NULL, NULL} +}; /* compose_tab_41 */ +static int hash_compose_tab_42[2] = +{-1,0}; /* hash_compose_tab_42 */ +static CompEntry compose_tab_42[] = { +{2887, 2891, 0, NULL, NULL} +}; /* compose_tab_42 */ +static int hash_compose_tab_43[2] = +{-1,0}; /* hash_compose_tab_43 */ +static CompEntry compose_tab_43[] = { +{2887, 2888, 0, NULL, NULL} +}; /* compose_tab_43 */ +static int hash_compose_tab_44[2] = +{-1,0}; /* hash_compose_tab_44 */ +static CompEntry compose_tab_44[] = { +{2887, 2892, 0, NULL, NULL} +}; /* compose_tab_44 */ +static int hash_compose_tab_45[4] = +{-1,-1,0,1}; /* hash_compose_tab_45 */ +static CompEntry compose_tab_45[] = { +{3014, 3018, 0, NULL, NULL}, +{3015, 3019, 0, NULL, NULL} +}; /* compose_tab_45 */ +static int hash_compose_tab_46[4] = +{-1,-1,0,1}; /* hash_compose_tab_46 */ +static CompEntry compose_tab_46[] = { +{2962, 2964, 0, NULL, NULL}, +{3014, 3020, 0, NULL, NULL} +}; /* compose_tab_46 */ +static int hash_compose_tab_47[2] = +{0,-1}; /* hash_compose_tab_47 */ +static CompEntry compose_tab_47[] = { +{3142, 3144, 0, NULL, NULL} +}; /* compose_tab_47 */ +static int hash_compose_tab_48[2] = +{0,-1}; /* hash_compose_tab_48 */ +static CompEntry compose_tab_48[] = { +{3270, 3274, 0, NULL, NULL} +}; /* compose_tab_48 */ +static int hash_compose_tab_49_1[2] = +{0,-1}; /* hash_compose_tab_49_1 */ +static CompEntry compose_tab_49_1[] = { +{3270, 3275, 0, NULL, NULL} +}; /* compose_tab_49_1 */ +static int hash_compose_tab_49[6] = +{2,-1,1,-1,-1,0}; /* hash_compose_tab_49 */ +static CompEntry compose_tab_49[] = { +{3263, 3264, 0, NULL, NULL}, +{3266, 0, 1, compose_tab_49_1, hash_compose_tab_49_1}, +{3270, 3271, 0, NULL, NULL} +}; /* compose_tab_49 */ +static int hash_compose_tab_50[2] = +{0,-1}; /* hash_compose_tab_50 */ +static CompEntry compose_tab_50[] = { +{3270, 3272, 0, NULL, NULL} +}; /* compose_tab_50 */ +static int hash_compose_tab_51[4] = +{-1,-1,0,1}; /* hash_compose_tab_51 */ +static CompEntry compose_tab_51[] = { +{3398, 3402, 0, NULL, NULL}, +{3399, 3403, 0, NULL, NULL} +}; /* compose_tab_51 */ +static int hash_compose_tab_52[2] = +{0,-1}; /* hash_compose_tab_52 */ +static CompEntry compose_tab_52[] = { +{3398, 3404, 0, NULL, NULL} +}; /* compose_tab_52 */ +static int hash_compose_tab_53[2] = +{-1,0}; /* hash_compose_tab_53 */ +static CompEntry compose_tab_53[] = { +{3661, 3635, 0, NULL, NULL} +}; /* compose_tab_53 */ +static int hash_compose_tab_54[2] = +{-1,0}; /* hash_compose_tab_54 */ +static CompEntry compose_tab_54[] = { +{3789, 3763, 0, NULL, NULL} +}; /* compose_tab_54 */ +static int hash_compose_tab_55_2[4] = +{-1,-1,0,1}; /* hash_compose_tab_55_2 */ +static CompEntry compose_tab_55_2[] = { +{4018, 3959, 0, NULL, NULL}, +{4019, 3961, 0, NULL, NULL} +}; /* compose_tab_55_2 */ +static int hash_compose_tab_55[6] = +{0,-1,1,2,-1,-1}; /* hash_compose_tab_55 */ +static CompEntry compose_tab_55[] = { +{3954, 3955, 0, NULL, NULL}, +{3956, 3957, 0, NULL, NULL}, +{3968, 0, 2, compose_tab_55_2, hash_compose_tab_55_2} +}; /* compose_tab_55 */ +static int hash_compose_tab_56[4] = +{-1,-1,0,1}; /* hash_compose_tab_56 */ +static CompEntry compose_tab_56[] = { +{4018, 3958, 0, NULL, NULL}, +{4019, 3960, 0, NULL, NULL} +}; /* compose_tab_56 */ +static int hash_compose_tab_57[4] = +{0,1,-1,-1}; /* hash_compose_tab_57 */ +static CompEntry compose_tab_57[] = { +{3904, 3945, 0, NULL, NULL}, +{3984, 4025, 0, NULL, NULL} +}; /* compose_tab_57 */ +static int hash_compose_tab_58[20] = +{-1,2,7,-1,-1,-1,0,3,5,8,-1,4,9,-1,-1,-1,1,6,-1,-1}; /* hash_compose_tab_58 */ +static CompEntry compose_tab_58[] = { +{3906, 3907, 0, NULL, NULL}, +{3916, 3917, 0, NULL, NULL}, +{3921, 3922, 0, NULL, NULL}, +{3926, 3927, 0, NULL, NULL}, +{3931, 3932, 0, NULL, NULL}, +{3986, 3987, 0, NULL, NULL}, +{3996, 3997, 0, NULL, NULL}, +{4001, 4002, 0, NULL, NULL}, +{4006, 4007, 0, NULL, NULL}, +{4011, 4012, 0, NULL, NULL} +}; /* compose_tab_58 */ +static int hash_compose_tab_59[96] = +{33,12,34,-1,13,35,14,36,15,37,-1,-1,-1,-1,-1,16,38,-1,17,39,-1,18,40,-1,19, + 41,-1,20,42,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,43,44,45, + 46,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,21,47,-1,-1,-1,-1,-1,-1,-1,0,22,-1,-1,-1,1, + 23,2,24,3,25,4,26,5,27,6,28,7,29,8,30,9,31,10,32,11}; /* hash_compose_tab_59 */ +static CompEntry compose_tab_59[] = { +{12358, 12436, 0, NULL, NULL}, +{12363, 12364, 0, NULL, NULL}, +{12365, 12366, 0, NULL, NULL}, +{12367, 12368, 0, NULL, NULL}, +{12369, 12370, 0, NULL, NULL}, +{12371, 12372, 0, NULL, NULL}, +{12373, 12374, 0, NULL, NULL}, +{12375, 12376, 0, NULL, NULL}, +{12377, 12378, 0, NULL, NULL}, +{12379, 12380, 0, NULL, NULL}, +{12381, 12382, 0, NULL, NULL}, +{12383, 12384, 0, NULL, NULL}, +{12385, 12386, 0, NULL, NULL}, +{12388, 12389, 0, NULL, NULL}, +{12390, 12391, 0, NULL, NULL}, +{12392, 12393, 0, NULL, NULL}, +{12399, 12400, 0, NULL, NULL}, +{12402, 12403, 0, NULL, NULL}, +{12405, 12406, 0, NULL, NULL}, +{12408, 12409, 0, NULL, NULL}, +{12411, 12412, 0, NULL, NULL}, +{12445, 12446, 0, NULL, NULL}, +{12454, 12532, 0, NULL, NULL}, +{12459, 12460, 0, NULL, NULL}, +{12461, 12462, 0, NULL, NULL}, +{12463, 12464, 0, NULL, NULL}, +{12465, 12466, 0, NULL, NULL}, +{12467, 12468, 0, NULL, NULL}, +{12469, 12470, 0, NULL, NULL}, +{12471, 12472, 0, NULL, NULL}, +{12473, 12474, 0, NULL, NULL}, +{12475, 12476, 0, NULL, NULL}, +{12477, 12478, 0, NULL, NULL}, +{12479, 12480, 0, NULL, NULL}, +{12481, 12482, 0, NULL, NULL}, +{12484, 12485, 0, NULL, NULL}, +{12486, 12487, 0, NULL, NULL}, +{12488, 12489, 0, NULL, NULL}, +{12495, 12496, 0, NULL, NULL}, +{12498, 12499, 0, NULL, NULL}, +{12501, 12502, 0, NULL, NULL}, +{12504, 12505, 0, NULL, NULL}, +{12507, 12508, 0, NULL, NULL}, +{12527, 12535, 0, NULL, NULL}, +{12528, 12536, 0, NULL, NULL}, +{12529, 12537, 0, NULL, NULL}, +{12530, 12538, 0, NULL, NULL}, +{12541, 12542, 0, NULL, NULL} +}; /* compose_tab_59 */ +static int hash_compose_tab_60[20] = +{-1,7,1,-1,8,2,-1,9,3,-1,-1,4,-1,-1,-1,5,-1,-1,6,0}; /* hash_compose_tab_60 */ +static CompEntry compose_tab_60[] = { +{12399, 12401, 0, NULL, NULL}, +{12402, 12404, 0, NULL, NULL}, +{12405, 12407, 0, NULL, NULL}, +{12408, 12410, 0, NULL, NULL}, +{12411, 12413, 0, NULL, NULL}, +{12495, 12497, 0, NULL, NULL}, +{12498, 12500, 0, NULL, NULL}, +{12501, 12503, 0, NULL, NULL}, +{12504, 12506, 0, NULL, NULL}, +{12507, 12509, 0, NULL, NULL} +}; /* compose_tab_60 */ +static int hash_compose_tab[122] = +{30,31,52,60,32,-1,-1,33,-1,34,35,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, + -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,0,1,2,3,4,-1,5,6,7,8,9,10,11,12,36,13,37,14, + 38,15,16,55,40,-1,-1,-1,-1,17,56,-1,-1,-1,-1,-1,41,18,19,20,42,21,22,-1,45, + 39,-1,23,24,-1,25,26,-1,-1,-1,-1,-1,-1,-1,-1,48,-1,43,44,51,53,-1,-1,27,46, + 54,28,-1,-1,47,-1,-1,-1,-1,49,50,-1,-1,57,-1,58,59,29}; /* hash_compose_tab */ +static CompEntry compose_tab[] = { +{768, 0, 39, compose_tab_0, hash_compose_tab_0}, +{769, 0, 70, compose_tab_1, hash_compose_tab_1}, +{770, 0, 27, compose_tab_2, hash_compose_tab_2}, +{771, 0, 19, compose_tab_3, hash_compose_tab_3}, +{772, 0, 28, compose_tab_4, hash_compose_tab_4}, +{774, 0, 30, compose_tab_5, hash_compose_tab_5}, +{775, 0, 40, compose_tab_6, hash_compose_tab_6}, +{776, 0, 50, compose_tab_7, hash_compose_tab_7}, +{777, 0, 15, compose_tab_8, hash_compose_tab_8}, +{778, 0, 6, compose_tab_9, hash_compose_tab_9}, +{779, 0, 6, compose_tab_10, hash_compose_tab_10}, +{780, 0, 34, compose_tab_11, hash_compose_tab_11}, +{781, 0, 17, compose_tab_12, hash_compose_tab_12}, +{783, 0, 14, compose_tab_13, hash_compose_tab_13}, +{785, 0, 12, compose_tab_14, hash_compose_tab_14}, +{787, 0, 15, compose_tab_15, hash_compose_tab_15}, +{788, 0, 17, compose_tab_16, hash_compose_tab_16}, +{795, 0, 4, compose_tab_17, hash_compose_tab_17}, +{803, 0, 39, compose_tab_18, hash_compose_tab_18}, +{804, 0, 2, compose_tab_19, hash_compose_tab_19}, +{805, 0, 2, compose_tab_20, hash_compose_tab_20}, +{807, 0, 20, compose_tab_21, hash_compose_tab_21}, +{808, 0, 10, compose_tab_22, hash_compose_tab_22}, +{813, 0, 12, compose_tab_23, hash_compose_tab_23}, +{814, 0, 2, compose_tab_24, hash_compose_tab_24}, +{816, 0, 6, compose_tab_25, hash_compose_tab_25}, +{817, 0, 17, compose_tab_26, hash_compose_tab_26}, +{834, 0, 12, compose_tab_27, hash_compose_tab_27}, +{837, 0, 6, compose_tab_28, hash_compose_tab_28}, +{1463, 0, 2, compose_tab_29, hash_compose_tab_29}, +{1464, 0, 1, compose_tab_30, hash_compose_tab_30}, +{1465, 0, 1, compose_tab_31, hash_compose_tab_31}, +{1468, 0, 22, compose_tab_32, hash_compose_tab_32}, +{1471, 0, 3, compose_tab_33, hash_compose_tab_33}, +{1473, 0, 2, compose_tab_34, hash_compose_tab_34}, +{1474, 0, 2, compose_tab_35, hash_compose_tab_35}, +{2364, 0, 11, compose_tab_36, hash_compose_tab_36}, +{2492, 0, 4, compose_tab_37, hash_compose_tab_37}, +{2494, 0, 1, compose_tab_38, hash_compose_tab_38}, +{2519, 0, 1, compose_tab_39, hash_compose_tab_39}, +{2620, 0, 5, compose_tab_40, hash_compose_tab_40}, +{2876, 0, 3, compose_tab_41, hash_compose_tab_41}, +{2878, 0, 1, compose_tab_42, hash_compose_tab_42}, +{2902, 0, 1, compose_tab_43, hash_compose_tab_43}, +{2903, 0, 1, compose_tab_44, hash_compose_tab_44}, +{3006, 0, 2, compose_tab_45, hash_compose_tab_45}, +{3031, 0, 2, compose_tab_46, hash_compose_tab_46}, +{3158, 0, 1, compose_tab_47, hash_compose_tab_47}, +{3266, 0, 1, compose_tab_48, hash_compose_tab_48}, +{3285, 0, 3, compose_tab_49, hash_compose_tab_49}, +{3286, 0, 1, compose_tab_50, hash_compose_tab_50}, +{3390, 0, 2, compose_tab_51, hash_compose_tab_51}, +{3415, 0, 1, compose_tab_52, hash_compose_tab_52}, +{3634, 0, 1, compose_tab_53, hash_compose_tab_53}, +{3762, 0, 1, compose_tab_54, hash_compose_tab_54}, +{3953, 0, 3, compose_tab_55, hash_compose_tab_55}, +{3968, 0, 2, compose_tab_56, hash_compose_tab_56}, +{4021, 0, 2, compose_tab_57, hash_compose_tab_57}, +{4023, 0, 10, compose_tab_58, hash_compose_tab_58}, +{12441, 0, 48, compose_tab_59, hash_compose_tab_59}, +{12442, 0, 10, compose_tab_60, hash_compose_tab_60} +}; /* compose_tab */ +#define COMP_CANDIDATE_MAP_OFFSET 24 +static Uint32 comp_candidate_map[] = { + 0x081ABFDFU, + 0x000361B8U, + 0x00000024U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x93800000U, + 0x00000006U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x10000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x50000000U, + 0x00800000U, + 0x00000000U, + 0x00000000U, + 0x10000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x50000000U, + 0x00C00000U, + 0x00000000U, + 0x00000000U, + 0x40000000U, + 0x00800000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00400000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00600004U, + 0x00000000U, + 0x00000000U, + 0x40000000U, + 0x00800000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00040000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00040000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00000000U, + 0x00020000U, + 0x00000001U, + 0x00A00000U +}; diff --git a/erts/emulator/beam/global.h b/erts/emulator/beam/global.h index 89c6625550..524db2a2eb 100644 --- a/erts/emulator/beam/global.h +++ b/erts/emulator/beam/global.h @@ -1596,6 +1596,19 @@ Sint erts_binary_set_loop_limit(Sint limit); /* erl_unicode.c */ void erts_init_unicode(void); Sint erts_unicode_set_loop_limit(Sint limit); + +void erts_native_filename_put(Eterm ioterm, int encoding, byte *p) ; +Sint erts_native_filename_need(Eterm ioterm, int encoding); +void erts_copy_utf8_to_utf16_little(byte *target, byte *bytes, int num_chars); +int erts_analyze_utf8(byte *source, Uint size, + byte **err_pos, Uint *num_chars, int *left); +char *erts_convert_filename_to_native(Eterm name, ErtsAlcType_t alloc_type, int allow_empty); + +#define ERTS_UTF8_OK 0 +#define ERTS_UTF8_INCOMPLETE 1 +#define ERTS_UTF8_ERROR 2 +#define ERTS_UTF8_ANALYZE_MORE 3 + /* erl_trace.c */ void erts_init_trace(void); void erts_trace_check_exiting(Eterm exiting); diff --git a/erts/emulator/beam/sys.h b/erts/emulator/beam/sys.h index 0d15272aa8..27c5f99320 100644 --- a/erts/emulator/beam/sys.h +++ b/erts/emulator/beam/sys.h @@ -1253,6 +1253,22 @@ char* win32_errorstr(int); #endif +/************************************************************************ + * Find out the native filename encoding of the process (look at locale of + * Unix processes and just do UTF16 on windows + ************************************************************************/ +#define ERL_FILENAME_UNKNOWN 0 +#define ERL_FILENAME_LATIN1 1 +#define ERL_FILENAME_UTF8 2 +#define ERL_FILENAME_UTF8_MAC 3 +#define ERL_FILENAME_WIN_WCHAR 4 + +int erts_get_native_filename_encoding(void); +/* The set function is only to be used by erl_init! */ +void erts_set_user_requested_filename_encoding(int encoding); +int erts_get_user_requested_filename_encoding(void); + +void erts_init_sys_common_misc(void); #endif diff --git a/erts/emulator/drivers/common/efile_drv.c b/erts/emulator/drivers/common/efile_drv.c index c450f10f48..786fa7da77 100644 --- a/erts/emulator/drivers/common/efile_drv.c +++ b/erts/emulator/drivers/common/efile_drv.c @@ -67,6 +67,8 @@ #define FILE_RESP_LDATA 6 #define FILE_RESP_N2DATA 7 #define FILE_RESP_EOF 8 +#define FILE_RESP_FNAME 9 +#define FILE_RESP_ALL_DATA 10 /* Options */ @@ -109,11 +111,11 @@ void erl_exit(int n, char *fmt, ...); static ErlDrvSysInfo sys_info; -/*#define TRACE 1*/ +/* #define TRACE 1 */ #ifdef TRACE -# define TRACE_C(c) (putchar(c)) -# define TRACE_S(s) (fputs((s), stdout)) -# define TRACE_F(args) (printf args) +# define TRACE_C(c) do { putchar(c); fflush(stdout); } while (0) +# define TRACE_S(s) do { fputs((s), stdout); fflush(stdout); } while (0) +# define TRACE_F(args) do { printf args ;fflush(stdout); } while (0) #else # define TRACE_C(c) ((void)(0)) # define TRACE_S(s) ((void)(0)) @@ -137,24 +139,54 @@ static ErlDrvSysInfo sys_info; #define MUTEX_UNLOCK(m) #endif - - #if 0 /* Experimental, for forcing all file operations to use the same thread. */ -static unsigned file_fixed_key = 1; -#define KEY(desc) (&file_fixed_key) + static unsigned file_fixed_key = 1; +# define KEY(desc) (&file_fixed_key) #else -#define KEY(desc) (&(desc)->key) +# define KEY(desc) (&(desc)->key) #endif +#ifdef FILENAMES_16BIT +# define FILENAME_BYTELEN(Str) filename_len_16bit(Str) +# define FILENAME_COPY(To,From) filename_cpy_16bit((To),(From)) +# define FILENAME_CHARSIZE 2 + + static int filename_len_16bit(char *str) + { + char *p = str; + while(*p != '\0' || p[1] != '\0') { + p += 2; + } + return (p - str); + } + + static void filename_cpy_16bit(char *to, char *from) + { + while(*from != '\0' || from[1] != '\0') { + *to++ = *from++; + *to++ = *from++; + } + *to++ = *from++; + *to++ = *from++; + } + +#else +# define FILENAME_BYTELEN(Str) strlen(Str) +# define FILENAME_COPY(To,From) strcpy(To,From) +# define FILENAME_CHARSIZE 1 +#endif -#if MAXPATHLEN >= BUFSIZ -#define RESBUFSIZE MAXPATHLEN+1 +#if (MAXPATHLEN+1)*FILENAME_CHARSIZE+1 > BUFSIZ +# define RESBUFSIZE ((MAXPATHLEN+1)*FILENAME_CHARSIZE+1) #else -#define RESBUFSIZE BUFSIZ +# define RESBUFSIZE BUFSIZ #endif + + + #define GET_TIME(i, b) \ (i).year = get_int32((b) + 0 * 4); \ (i).month = get_int32((b) + 1 * 4); \ @@ -286,9 +318,9 @@ struct t_preadv { }; #define READDIR_BUFSIZE (8*1024) -#if READDIR_BUFSIZE < (2*MAXPATHLEN) -#undef READDIR_BUFSIZE -#define READDIR_BUFSIZE (2*MAXPATHLEN) +#if READDIR_BUFSIZE < (FILENAME_CHARSIZE*2*(MAXPATHLEN+1)) +# undef READDIR_BUFSIZE +# define READDIR_BUFSIZE (FILENAME_CHARSIZE*2*(MAXPATHLEN+1)) #endif struct t_readdir_buf { @@ -369,6 +401,7 @@ struct t_data }; + #define EF_ALLOC(S) driver_alloc((S)) #define EF_REALLOC(P, S) driver_realloc((P), (S)) #define EF_SAFE_ALLOC(S) ef_safe_alloc((S)) @@ -1288,7 +1321,7 @@ static void invoke_writev(void *data) { p < size && iovcnt < iovlen; p += iov0[iovcnt++].iov_len) ; - iov = EF_ALLOC(sizeof(SysIOVec)*iovcnt); + iov = EF_SAFE_ALLOC(sizeof(SysIOVec)*iovcnt); memcpy(iov,iov0,iovcnt*sizeof(SysIOVec)); MUTEX_UNLOCK(d->c.writev.q_mtx); /* Let go of lock until we deque from original vector */ @@ -1368,7 +1401,7 @@ static void invoke_readlink(void *data) d->result_ok = efile_readlink(&d->errInfo, d->b, resbuf+1, RESBUFSIZE-1); if (d->result_ok != 0) - strcpy((char *) d->b + 1, resbuf+1); + FILENAME_COPY((char *) d->b + 1, resbuf+1); } static void invoke_altname(void *data) @@ -1380,7 +1413,7 @@ static void invoke_altname(void *data) d->result_ok = efile_altname(&d->errInfo, d->b, resbuf+1, RESBUFSIZE-1); if (d->result_ok != 0) - strcpy((char *) d->b + 1, resbuf+1); + FILENAME_COPY((char *) d->b + 1, resbuf+1); } static void invoke_pwritev(void *data) { @@ -1405,7 +1438,7 @@ static void invoke_pwritev(void *data) { /* Lock the queue just for a while, we don't want it locked during write */ MUTEX_LOCK(c->q_mtx); iov0 = driver_peekq(c->port, &iovlen); - iov = EF_ALLOC(sizeof(SysIOVec)*iovlen); + iov = EF_SAFE_ALLOC(sizeof(SysIOVec)*iovlen); memcpy(iov,iov0,sizeof(SysIOVec)*iovlen); MUTEX_UNLOCK(c->q_mtx); @@ -1499,7 +1532,7 @@ static void invoke_link(void *data) char *new_name; d->again = 0; - new_name = name+strlen(name)+1; + new_name = name+FILENAME_BYTELEN(name)+FILENAME_CHARSIZE; d->result_ok = efile_link(&d->errInfo, name, new_name); } @@ -1510,7 +1543,7 @@ static void invoke_symlink(void *data) char *new_name; d->again = 0; - new_name = name+strlen(name)+1; + new_name = name+FILENAME_BYTELEN(name)+FILENAME_CHARSIZE; d->result_ok = efile_symlink(&d->errInfo, name, new_name); } @@ -1521,7 +1554,7 @@ static void invoke_rename(void *data) char *new_name; d->again = 0; - new_name = name+strlen(name)+1; + new_name = name+FILENAME_BYTELEN(name)+FILENAME_CHARSIZE; d->result_ok = efile_rename(&d->errInfo, name, new_name); } @@ -1569,13 +1602,15 @@ static void invoke_readdir(void *data) int s; char *p = NULL; int buf_sz = 0; + size_t tmp_bs; d->again = 0; d->errInfo.posix_errno = 0; while (1) { char *str; - if (buf_sz < (4 /* sz */ + 1 /* cmd */ + MAXPATHLEN + 1 /* '\0' */)) { + if (buf_sz < (4 /* sz */ + 1 /* cmd */ + + FILENAME_CHARSIZE*(MAXPATHLEN + 1))) { struct t_readdir_buf *b; if (p) { put_int32(0, p); /* EOB */ @@ -1591,18 +1626,18 @@ static void invoke_readdir(void *data) buf_sz = READDIR_BUFSIZE - 4/* EOB */; } - p[4] = FILE_RESP_OK; + p[4] = FILE_RESP_FNAME; buf_sz -= 4 + 1; str = p + 4 + 1; ASSERT(buf_sz >= MAXPATHLEN + 1); - s = efile_readdir(&d->errInfo, d->b, &d->dir_handle, str, buf_sz); + tmp_bs = buf_sz; + s = efile_readdir(&d->errInfo, d->b, &d->dir_handle, str, &tmp_bs); if (s) { - int str_sz = strlen(str); - int sz = str_sz + 1; - put_int32(sz, p); - p += 4 + sz; - buf_sz -= str_sz; + put_int32(tmp_bs + 1 /* 1 byte for opcode */, p); + p += 4 + tmp_bs + 1; + ASSERT(p == (str + tmp_bs)); + buf_sz -= tmp_bs; } else { put_int32(1, p); @@ -1911,7 +1946,7 @@ file_async_ready(ErlDrvData e, ErlDrvThreadData data) if (!d->result_ok) reply_error(desc, &d->errInfo); else { - header[0] = FILE_RESP_OK; + header[0] = FILE_RESP_ALL_DATA; TRACE_C('R'); driver_output_binary(desc->port, header, 1, d->c.read_file.binp, @@ -1968,10 +2003,10 @@ file_async_ready(ErlDrvData e, ErlDrvThreadData data) if (!d->result_ok) reply_error(desc, &d->errInfo); else { - resbuf[0] = FILE_RESP_OK; - length = 1+strlen((char*) resbuf+1); + resbuf[0] = FILE_RESP_FNAME; + length = 1+FILENAME_BYTELEN((char*) resbuf+1); TRACE_C('R'); - driver_output2(desc->port, resbuf, length, NULL, 0); + driver_output2(desc->port, resbuf, 1, resbuf+1, length-1); } free_data(data); break; @@ -2031,13 +2066,18 @@ file_async_ready(ErlDrvData e, ErlDrvThreadData data) int sz = get_int32(p); while (sz) { /* 0 == EOB */ p += 4; - driver_output2(desc->port, p, sz, NULL, 0); + if (sz - 1 > 0) { + driver_output2(desc->port, p, 1, p+1, sz-1); + } else { + driver_output2(desc->port, p, 1, NULL, 0); + } p += sz; sz = get_int32(p); } b1 = b1->next; EF_FREE(b2); } + d->c.read_dir.first_buf = NULL; d->c.read_dir.last_buf = NULL; } @@ -2113,9 +2153,9 @@ file_output(ErlDrvData e, char* buf, int count) case FILE_MKDIR: { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(name) + 1); + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE); - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->command = command; d->invoke = invoke_mkdir; d->free = free_data; @@ -2124,9 +2164,9 @@ file_output(ErlDrvData e, char* buf, int count) } case FILE_RMDIR: { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(name) + 1); + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE); - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->command = command; d->invoke = invoke_rmdir; d->free = free_data; @@ -2135,9 +2175,9 @@ file_output(ErlDrvData e, char* buf, int count) } case FILE_DELETE: { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(name) + 1); + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE); - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->command = command; d->invoke = invoke_delete_file; d->free = free_data; @@ -2147,14 +2187,14 @@ file_output(ErlDrvData e, char* buf, int count) case FILE_RENAME: { char* new_name; - - new_name = name+strlen(name)+1; + int namelen = FILENAME_BYTELEN(name)+FILENAME_CHARSIZE; + new_name = name+namelen; d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 - + strlen(name) + 1 - + strlen(new_name) + 1); + + namelen + + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE); - strcpy(d->b, name); - strcpy(d->b + strlen(name) + 1, new_name); + FILENAME_COPY(d->b, name); + FILENAME_COPY(d->b + namelen, new_name); d->flags = desc->flags; d->fd = fd; d->command = command; @@ -2165,9 +2205,9 @@ file_output(ErlDrvData e, char* buf, int count) } case FILE_CHDIR: { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(name) + 1); + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE); - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->command = command; d->invoke = invoke_chdir; d->free = free_data; @@ -2190,9 +2230,10 @@ file_output(ErlDrvData e, char* buf, int count) #ifdef USE_THREADS if (sys_info.async_threads > 0) { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(name) + 1); + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + + FILENAME_CHARSIZE); - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->dir_handle = NULL; d->command = command; d->invoke = invoke_readdir; @@ -2205,17 +2246,19 @@ file_output(ErlDrvData e, char* buf, int count) else #endif { + size_t resbufsize; char resbuf[RESBUFSIZE+1]; EFILE_DIR_HANDLE dir_handle; /* Handle to open directory. */ errInfo.posix_errno = 0; dir_handle = NULL; - resbuf[0] = FILE_RESP_OK; + resbuf[0] = FILE_RESP_FNAME; + resbufsize = RESBUFSIZE; while (efile_readdir(&errInfo, name, &dir_handle, - resbuf+1, RESBUFSIZE)) { - int length = 1 + strlen(resbuf+1); - driver_output2(desc->port, resbuf, length, NULL, 0); + resbuf+1, &resbufsize)) { + driver_output2(desc->port, resbuf, 1, resbuf+1, resbufsize); + resbufsize = RESBUFSIZE; } if (errInfo.posix_errno != 0) { reply_error(desc, &errInfo); @@ -2227,11 +2270,12 @@ file_output(ErlDrvData e, char* buf, int count) } case FILE_OPEN: { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(buf+4) + 1); + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(buf+4) + + FILENAME_CHARSIZE); d->flags = get_int32((uchar*)buf); name = buf+4; - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->command = command; d->invoke = invoke_open; d->free = free_data; @@ -2240,44 +2284,45 @@ file_output(ErlDrvData e, char* buf, int count) } case FILE_FDATASYNC: - { + { d = EF_SAFE_ALLOC(sizeof(struct t_data)); - + d->fd = fd; d->command = command; d->invoke = invoke_fdatasync; d->free = free_data; d->level = 2; goto done; - } + } case FILE_FSYNC: - { - d = EF_SAFE_ALLOC(sizeof(struct t_data)); - - d->fd = fd; - d->command = command; - d->invoke = invoke_fsync; - d->free = free_data; - d->level = 2; - goto done; - } + { + d = EF_SAFE_ALLOC(sizeof(struct t_data)); + + d->fd = fd; + d->command = command; + d->invoke = invoke_fsync; + d->free = free_data; + d->level = 2; + goto done; + } case FILE_FSTAT: case FILE_LSTAT: - { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + strlen(name) + 1); + { + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + + FILENAME_CHARSIZE); + + FILENAME_COPY(d->b, name); + d->fd = fd; + d->command = command; + d->invoke = invoke_flstat; + d->free = free_data; + d->level = 2; + goto done; + } - strcpy(d->b, name); - d->fd = fd; - d->command = command; - d->invoke = invoke_flstat; - d->free = free_data; - d->level = 2; - goto done; - } - case FILE_TRUNCATE: { d = EF_SAFE_ALLOC(sizeof(struct t_data)); @@ -2294,7 +2339,7 @@ file_output(ErlDrvData e, char* buf, int count) case FILE_WRITE_INFO: { d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 - + strlen(buf+21*4) + 1); + + FILENAME_BYTELEN(buf+21*4) + FILENAME_CHARSIZE); d->info.mode = get_int32(buf + 0 * 4); d->info.uid = get_int32(buf + 1 * 4); @@ -2302,7 +2347,7 @@ file_output(ErlDrvData e, char* buf, int count) GET_TIME(d->info.accessTime, buf + 3 * 4); GET_TIME(d->info.modifyTime, buf + 9 * 4); GET_TIME(d->info.cTime, buf + 15 * 4); - strcpy(d->b, buf+21*4); + FILENAME_COPY(d->b, buf+21*4); d->command = command; d->invoke = invoke_write_info; d->free = free_data; @@ -2314,7 +2359,7 @@ file_output(ErlDrvData e, char* buf, int count) { d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + RESBUFSIZE + 1); - strcpy(d->b, name); + FILENAME_COPY(d->b, name); d->command = command; d->invoke = invoke_readlink; d->free = free_data; @@ -2323,28 +2368,29 @@ file_output(ErlDrvData e, char* buf, int count) } case FILE_ALTNAME: - { - d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + RESBUFSIZE + 1); - strcpy(d->b, name); - d->command = command; - d->invoke = invoke_altname; - d->free = free_data; - d->level = 2; - goto done; - } + { + d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + RESBUFSIZE + 1); + FILENAME_COPY(d->b, name); + d->command = command; + d->invoke = invoke_altname; + d->free = free_data; + d->level = 2; + goto done; + } case FILE_LINK: { char* new_name; + int namelen = FILENAME_BYTELEN(name) + FILENAME_CHARSIZE; - new_name = name+strlen(name)+1; + new_name = name+namelen; d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 - + strlen(name) + 1 - + strlen(new_name) + 1); + + namelen + + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE); - strcpy(d->b, name); - strcpy(d->b + strlen(name) + 1, new_name); + FILENAME_COPY(d->b, name); + FILENAME_COPY(d->b + namelen, new_name); d->flags = desc->flags; d->fd = fd; d->command = command; @@ -2357,14 +2403,15 @@ file_output(ErlDrvData e, char* buf, int count) case FILE_SYMLINK: { char* new_name; + int namelen = FILENAME_BYTELEN(name) + FILENAME_CHARSIZE; - new_name = name+strlen(name)+1; + new_name = name+namelen; d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 - + strlen(name) + 1 - + strlen(new_name) + 1); + + namelen + + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE); - strcpy(d->b, name); - strcpy(d->b + strlen(name) + 1, new_name); + FILENAME_COPY(d->b, name); + FILENAME_COPY(d->b + namelen, new_name); d->flags = desc->flags; d->fd = fd; d->command = command; @@ -3004,6 +3051,7 @@ file_outputv(ErlDrvData e, ErlIOVec *ev) { case FILE_READ_FILE: { struct t_data *d; + char *filename; if (ev->size < 1+1) { /* Buffer contains empty name */ reply_posix_error(desc, ENOENT); @@ -3014,7 +3062,8 @@ file_outputv(ErlDrvData e, ErlIOVec *ev) { reply_posix_error(desc, EINVAL); goto done; } - d = EF_ALLOC(sizeof(struct t_data) + ev->size); + filename = EV_CHAR_P(ev, p, q); + d = EF_ALLOC(sizeof(struct t_data) -1 + FILENAME_BYTELEN(filename) + FILENAME_CHARSIZE); if (! d) { reply_posix_error(desc, ENOMEM); goto done; @@ -3022,8 +3071,7 @@ file_outputv(ErlDrvData e, ErlIOVec *ev) { d->command = command; d->reply = !0; /* Copy name */ - memcpy(d->c.read_file.name, EV_CHAR_P(ev, p, q), ev->size-1); - d->c.read_file.name[ev->size-1] = '\0'; + FILENAME_COPY(d->c.read_file.name, filename); d->c.read_file.binp = NULL; d->invoke = invoke_read_file; d->free = free_read_file; diff --git a/erts/emulator/drivers/common/erl_efile.h b/erts/emulator/drivers/common/erl_efile.h index ac95c1f949..3097ded3f1 100644 --- a/erts/emulator/drivers/common/erl_efile.h +++ b/erts/emulator/drivers/common/erl_efile.h @@ -59,6 +59,14 @@ #define FA_WRITE 1 #define FA_READ 2 +/* Some OS'es (i.e. Windows) has filenames in wide charaqcters. That requires special handling */ +/* Note that we do *not* honor alignment in the communication to the OS specific driver, */ +/* which is not a problem on x86, but might be on other platforms. The OS specific efile */ +/* implementation is expected to align if needed */ +#ifdef __WIN32__ +#define FILENAMES_16BIT 1 +#endif + /* * An handle to an open directory. To be cast to the correct type * in the system-dependent directory functions. @@ -123,7 +131,7 @@ int efile_getdcwd(Efile_error* errInfo, int drive, char* buffer, size_t size); int efile_readdir(Efile_error* errInfo, char* name, EFILE_DIR_HANDLE* dir_handle, - char* buffer, size_t size); + char* buffer, size_t *size); int efile_openfile(Efile_error* errInfo, char* name, int flags, int* pfd, Sint64* pSize); void efile_closefile(int fd); diff --git a/erts/emulator/drivers/common/gzio.c b/erts/emulator/drivers/common/gzio.c index 801bc61d4d..5531a275ea 100644 --- a/erts/emulator/drivers/common/gzio.c +++ b/erts/emulator/drivers/common/gzio.c @@ -28,6 +28,7 @@ #ifdef __WIN32__ #define HAVE_CONFLICTING_FREAD_DECLARATION +#define FILENAMES_16BIT 1 #endif #ifdef STDC @@ -102,6 +103,40 @@ local uLong getLong OF((gz_stream *s)); # define ERTS_GZREAD(File, Buf, Count) fread((Buf), 1, (Count), (File)) #endif +/* + * Ripped from efile_drv.c + */ + +#ifdef FILENAMES_16BIT +# define FILENAME_BYTELEN(Str) filename_len_16bit(Str) +# define FILENAME_COPY(To,From) filename_cpy_16bit((To),(From)) +# define FILENAME_CHARSIZE 2 + + static int filename_len_16bit(const char *str) + { + const char *p = str; + while(*p != '\0' || p[1] != '\0') { + p += 2; + } + return (p - str); + } + + static void filename_cpy_16bit(char *to, const char *from) + { + while(*from != '\0' || from[1] != '\0') { + *to++ = *from++; + *to++ = *from++; + } + *to++ = *from++; + *to++ = *from++; + } + +#else +# define FILENAME_BYTELEN(Str) strlen(Str) +# define FILENAME_COPY(To,From) strcpy(To,From) +# define FILENAME_CHARSIZE 1 +#endif + /* =========================================================================== Opens a gzip (.gz) file for reading or writing. The mode parameter is as in fopen ("rb" or "wb"). The file is given either by file descriptor @@ -144,11 +179,11 @@ local gzFile gz_open (path, mode) s->position = 0; s->destroy = destroy; - s->path = (char*)ALLOC(strlen(path)+1); + s->path = (char*)ALLOC(FILENAME_BYTELEN(path)+FILENAME_CHARSIZE); if (s->path == NULL) { return s->destroy(s), (gzFile)Z_NULL; } - strcpy(s->path, path); /* do this early for debugging */ + FILENAME_COPY(s->path, path); /* do this early for debugging */ s->mode = '\0'; do { @@ -194,7 +229,22 @@ local gzFile gz_open (path, mode) s->stream.avail_out = Z_BUFSIZE; errno = 0; -#ifdef UNIX +#if defined(FILENAMES_16BIT) + { + char wfmode[160]; + int i=0,j; + for(j=0;fmode[j] != '\0';++j) { + wfmode[i++]=fmode[j]; + wfmode[i++]='\0'; + } + wfmode[i++] = '\0'; + wfmode[i++] = '\0'; + s->file = F_OPEN(path, wfmode); + if (s->file == NULL) { + return s->destroy(s), (gzFile)Z_NULL; + } + } +#elif defined(UNIX) if (s->mode == 'r') { s->file = open(path, O_RDONLY); } else { diff --git a/erts/emulator/drivers/common/inet_drv.c b/erts/emulator/drivers/common/inet_drv.c index 18f7cdd15a..6f56cab575 100644 --- a/erts/emulator/drivers/common/inet_drv.c +++ b/erts/emulator/drivers/common/inet_drv.c @@ -88,9 +88,21 @@ #include <winsock2.h> #endif #include <windows.h> +#include <Ws2tcpip.h> /* NEED VC 6.0 or higher */ + +/* Visual studio 2008+: NTDDI_VERSION needs to be set for iphlpapi.h + to define the right structures. It needs to be set to WINXP (or LONGHORN) + for IPV6 to work and it's set lower by default, so we need to change it. */ +#ifdef HAVE_SDKDDKVER_H +# include <sdkddkver.h> +# ifdef NTDDI_VERSION +# undef NTDDI_VERSION +# endif +# define NTDDI_VERSION NTDDI_WINXP +#endif + #include <iphlpapi.h> -#include <Ws2tcpip.h> /* NEED VC 6.0 !!! */ #undef WANT_NONBLOCKING #include "sys.h" diff --git a/erts/emulator/drivers/unix/unix_efile.c b/erts/emulator/drivers/unix/unix_efile.c index b19f632f52..6297ccb8bc 100644 --- a/erts/emulator/drivers/unix/unix_efile.c +++ b/erts/emulator/drivers/unix/unix_efile.c @@ -587,7 +587,8 @@ efile_readdir(Efile_error* errInfo, /* Where to return error codes. */ open directory.*/ char* buffer, /* Pointer to buffer for one filename. */ - size_t size) /* Size of buffer. */ + size_t *size) /* in-out Size of buffer, length + of name. */ { DIR *dp; /* Pointer to directory structure. */ struct dirent* dirp; /* Pointer to directory entry. */ @@ -620,6 +621,7 @@ efile_readdir(Efile_error* errInfo, /* Where to return error codes. */ continue; buffer[0] = '\0'; strncat(buffer, dirp->d_name, size-1); + *size = strlen(dirp->d_name); return 1; } } diff --git a/erts/emulator/drivers/win32/win_efile.c b/erts/emulator/drivers/win32/win_efile.c index 6de08e2fa6..4ec9579529 100755 --- a/erts/emulator/drivers/win32/win_efile.c +++ b/erts/emulator/drivers/win32/win_efile.c @@ -23,20 +23,20 @@ #include <windows.h> #include "sys.h" #include <ctype.h> - +#include <wchar.h> #include "erl_efile.h" /* * Microsoft-specific function to map a WIN32 error code to a Posix errno. */ -#define ISSLASH(a) ((a) == '\\' || (a) == '/') +#define ISSLASH(a) ((a) == L'\\' || (a) == L'/') #define ISDIR(st) (((st).st_mode&S_IFMT) == S_IFDIR) #define ISREG(st) (((st).st_mode&S_IFMT) == S_IFREG) #define IS_DOT_OR_DOTDOT(s) \ - (s[0] == '.' && (s[1] == '\0' || (s[1] == '.' && s[2] == '\0'))) + ((s)[0] == L'.' && ((s)[1] == L'\0' || ((s)[1] == L'.' && (s)[2] == L'\0'))) #ifndef INVALID_FILE_ATTRIBUTES #define INVALID_FILE_ATTRIBUTES ((DWORD) 0xFFFFFFFF) @@ -44,9 +44,9 @@ static int check_error(int result, Efile_error* errInfo); static int set_error(Efile_error* errInfo); -static int IsRootUNCName(const char* path); -static int extract_root(char* name); -static unsigned short dos_to_posix_mode(int attr, const char *name); +static int is_root_unc_name(const WCHAR *path); +static int extract_root(WCHAR *name); +static unsigned short dos_to_posix_mode(int attr, const WCHAR *name); static int errno_map(DWORD last_error) { @@ -196,27 +196,26 @@ win_writev(Efile_error* errInfo, int -efile_mkdir(errInfo, name) -Efile_error* errInfo; /* Where to return error codes. */ -char* name; /* Name of directory to create. */ +efile_mkdir(Efile_error* errInfo, /* Where to return error codes. */ + char* name) /* Name of directory to create. */ { - return check_error(mkdir(name), errInfo); + return check_error(_wmkdir((WCHAR *) name), errInfo); } int -efile_rmdir(errInfo, name) -Efile_error* errInfo; /* Where to return error codes. */ -char* name; /* Name of directory to delete. */ +efile_rmdir(Efile_error* errInfo, /* Where to return error codes. */ + char* name) /* Name of directory to delete. */ { OSVERSIONINFO os; DWORD attr; + WCHAR *wname = (WCHAR *) name; - if (RemoveDirectory(name) != FALSE) { + if (RemoveDirectoryW(wname) != FALSE) { return 1; } errno = errno_map(GetLastError()); if (errno == EACCES) { - attr = GetFileAttributes(name); + attr = GetFileAttributesW(wname); if (attr != (DWORD) -1) { if ((attr & FILE_ATTRIBUTE_DIRECTORY) == 0) { /* @@ -238,21 +237,21 @@ char* name; /* Name of directory to delete. */ GetVersionEx(&os); if (os.dwPlatformId == VER_PLATFORM_WIN32_WINDOWS) { HANDLE handle; - WIN32_FIND_DATA data; - char buffer[2*MAX_PATH]; + WIN32_FIND_DATAW data; + WCHAR buffer[2*MAX_PATH]; int len; - len = strlen(name); - strcpy(buffer, name); - if (buffer[0] && buffer[len-1] != '\\' && buffer[len-1] != '/') { - strcat(buffer, "\\"); + len = wcslen(wname); + wcscpy(buffer, wname); + if (buffer[0] && buffer[len-1] != L'\\' && buffer[len-1] != L'/') { + wcscat(buffer, L"\\"); } - strcat(buffer, "*.*"); - handle = FindFirstFile(buffer, &data); + wcscat(buffer, L"*.*"); + handle = FindFirstFileW(buffer, &data); if (handle != INVALID_HANDLE_VALUE) { while (1) { - if ((strcmp(data.cFileName, ".") != 0) - && (strcmp(data.cFileName, "..") != 0)) { + if ((wcscmp(data.cFileName, L".") != 0) + && (wcscmp(data.cFileName, L"..") != 0)) { /* * Found something in this directory. */ @@ -260,7 +259,7 @@ char* name; /* Name of directory to delete. */ errno = EEXIST; break; } - if (FindNextFile(handle, &data) == FALSE) { + if (FindNextFileW(handle, &data) == FALSE) { break; } } @@ -284,19 +283,19 @@ char* name; /* Name of directory to delete. */ } int -efile_delete_file(errInfo, name) -Efile_error* errInfo; /* Where to return error codes. */ -char* name; /* Name of file to delete. */ +efile_delete_file(Efile_error* errInfo, /* Where to return error codes. */ + char* name) /* Name of file to delete. */ { DWORD attr; + WCHAR *wname = (WCHAR *) name; - if (DeleteFile(name) != FALSE) { + if (DeleteFileW(wname) != FALSE) { return 1; } errno = errno_map(GetLastError()); if (errno == EACCES) { - attr = GetFileAttributes(name); + attr = GetFileAttributesW(wname); if (attr != (DWORD) -1) { if (attr & FILE_ATTRIBUTE_DIRECTORY) { /* @@ -308,7 +307,7 @@ char* name; /* Name of file to delete. */ } } } else if (errno == ENOENT) { - attr = GetFileAttributes(name); + attr = GetFileAttributesW(wname); if (attr != (DWORD) -1) { if (attr & FILE_ATTRIBUTE_DIRECTORY) { /* @@ -362,20 +361,21 @@ char* name; /* Name of file to delete. */ */ int -efile_rename(errInfo, src, dst) -Efile_error* errInfo; /* Where to return error codes. */ -char* src; /* Original name. */ -char* dst; /* New name. */ +efile_rename(Efile_error* errInfo, /* Where to return error codes. */ + char* src, /* Original name. */ + char* dst) /* New name. */ { DWORD srcAttr, dstAttr; + WCHAR *wsrc = (WCHAR *) src; + WCHAR *wdst = (WCHAR *) dst; - if (MoveFile(src, dst) != FALSE) { + if (MoveFileW(wsrc, wdst) != FALSE) { return 1; } errno = errno_map(GetLastError()); - srcAttr = GetFileAttributes(src); - dstAttr = GetFileAttributes(dst); + srcAttr = GetFileAttributesW(wsrc); + dstAttr = GetFileAttributesW(wdst); if (srcAttr == (DWORD) -1) { srcAttr = 0; } @@ -390,22 +390,22 @@ char* dst; /* New name. */ if (errno == EACCES) { decode: if (srcAttr & FILE_ATTRIBUTE_DIRECTORY) { - char srcPath[MAX_PATH], dstPath[MAX_PATH]; - char *srcRest, *dstRest; + WCHAR srcPath[MAX_PATH], dstPath[MAX_PATH]; + WCHAR *srcRest, *dstRest; int size; - size = GetFullPathName(src, sizeof(srcPath), srcPath, &srcRest); - if ((size == 0) || (size > sizeof(srcPath))) { + size = GetFullPathNameW(wsrc, MAX_PATH, srcPath, &srcRest); + if ((size == 0) || (size > MAX_PATH)) { return check_error(-1, errInfo); } - size = GetFullPathName(dst, sizeof(dstPath), dstPath, &dstRest); - if ((size == 0) || (size > sizeof(dstPath))) { + size = GetFullPathNameW(wdst, MAX_PATH, dstPath, &dstRest); + if ((size == 0) || (size > MAX_PATH)) { return check_error(-1, errInfo); } if (srcRest == NULL) { - srcRest = srcPath + strlen(srcPath); + srcRest = srcPath + wcslen(srcPath); } - if (strnicmp(srcPath, dstPath, srcRest - srcPath) == 0) { + if (_wcsnicmp(srcPath, dstPath, srcRest - srcPath) == 0) { /* * Trying to move a directory into itself. */ @@ -420,14 +420,14 @@ char* dst; /* New name. */ } (void) extract_root(dstPath); - if (dstPath[0] == '\0') { + if (dstPath[0] == L'\0') { /* * The filename was invalid. (Don't know why, * but play it safe.) */ errno = EINVAL; } - if (stricmp(srcPath, dstPath) != 0) { + if (_wcsicmp(srcPath, dstPath) != 0) { /* * If src is a directory and dst filesystem != src * filesystem, errno should be EXDEV. It is very @@ -463,14 +463,14 @@ char* dst; /* New name. */ * fails, it's because it wasn't empty. */ - if (RemoveDirectory(dst)) { + if (RemoveDirectoryW(wdst)) { /* * Now that that empty directory is gone, we can try * renaming again. If that fails, we'll put this empty * directory back, for completeness. */ - if (MoveFile(src, dst) != FALSE) { + if (MoveFileW(wsrc, wdst) != FALSE) { return 1; } @@ -480,8 +480,8 @@ char* dst; /* New name. */ */ errno = errno_map(GetLastError()); - CreateDirectory(dst, NULL); - SetFileAttributes(dst, dstAttr); + CreateDirectoryW(wdst, NULL); + SetFileAttributesW(wdst, dstAttr); if (errno == EACCES) { /* * Decode the EACCES to a more meaningful error. @@ -506,17 +506,17 @@ char* dst; /* New name. */ * put temp file back to old name. */ - char tempName[MAX_PATH]; + WCHAR tempName[MAX_PATH]; int result, size; - char *rest; + WCHAR *rest; - size = GetFullPathName(dst, sizeof(tempName), tempName, &rest); - if ((size == 0) || (size > sizeof(tempName)) || (rest == NULL)) { + size = GetFullPathNameW(wdst, MAX_PATH, tempName, &rest); + if ((size == 0) || (size > MAX_PATH) || (rest == NULL)) { return check_error(-1, errInfo); } - *rest = '\0'; + *rest = L'\0'; result = -1; - if (GetTempFileName(tempName, "erlr", 0, tempName) != 0) { + if (GetTempFileNameW(tempName, L"erlr", 0, tempName) != 0) { /* * Strictly speaking, need the following DeleteFile and * MoveFile to be joined as an atomic operation so no @@ -524,15 +524,15 @@ char* dst; /* New name. */ * same temp file. */ - DeleteFile(tempName); - if (MoveFile(dst, tempName) != FALSE) { - if (MoveFile(src, dst) != FALSE) { - SetFileAttributes(tempName, FILE_ATTRIBUTE_NORMAL); - DeleteFile(tempName); + DeleteFileW(tempName); + if (MoveFileW(wdst, tempName) != FALSE) { + if (MoveFileW(wsrc, wdst) != FALSE) { + SetFileAttributesW(tempName, FILE_ATTRIBUTE_NORMAL); + DeleteFileW(tempName); return 1; } else { - DeleteFile(dst); - MoveFile(tempName, dst); + DeleteFileW(wdst); + MoveFileW(tempName, wdst); } } @@ -558,11 +558,10 @@ char* dst; /* New name. */ } int -efile_chdir(errInfo, name) -Efile_error* errInfo; /* Where to return error codes. */ -char* name; /* Name of directory to make current. */ +efile_chdir(Efile_error* errInfo, /* Where to return error codes. */ + char* name) /* Name of directory to make current. */ { - int success = check_error(chdir(name), errInfo); + int success = check_error(_wchdir((WCHAR *) name), errInfo); if (!success && errInfo->posix_errno == EINVAL) /* POSIXification of errno */ errInfo->posix_errno = ENOENT; @@ -570,59 +569,65 @@ char* name; /* Name of directory to make current. */ } int -efile_getdcwd(errInfo, drive, buffer, size) -Efile_error* errInfo; /* Where to return error codes. */ -int drive; /* 0 - current, 1 - A, 2 - B etc. */ -char* buffer; /* Where to return the current directory. */ -size_t size; /* Size of buffer. */ +efile_getdcwd(Efile_error* errInfo, /* Where to return error codes. */ + int drive, /* 0 - current, 1 - A, 2 - B etc. */ + char* buffer, /* Where to return the current directory. */ + size_t size) /* Size of buffer. */ { - if (_getdcwd(drive, buffer, size) == NULL) + WCHAR *wbuffer = (WCHAR *) buffer; + size_t wbuffer_size = size / 2; + if (_wgetdcwd(drive, wbuffer, wbuffer_size) == NULL) return check_error(-1, errInfo); - for ( ; *buffer; buffer++) - if (*buffer == '\\') - *buffer = '/'; + for ( ; *wbuffer; wbuffer++) + if (*wbuffer == L'\\') + *wbuffer = L'/'; return 1; } int -efile_readdir(errInfo, name, dir_handle, buffer, size) -Efile_error* errInfo; /* Where to return error codes. */ -char* name; /* Name of directory to open. */ -EFILE_DIR_HANDLE* dir_handle; /* Directory handle of open directory. */ -char* buffer; /* Pointer to buffer for one filename. */ -size_t size; /* Size of buffer. */ +efile_readdir(Efile_error* errInfo, /* Where to return error codes. */ + char* name, /* Name of directory to list */ + EFILE_DIR_HANDLE* dir_handle, /* Handle of opened directory or NULL */ + char* buffer, /* Buffer to put one filename in */ + size_t *size) /* in-out size of buffer/size of filename excluding zero + termination in bytes*/ { HANDLE dir; /* Handle to directory. */ - char wildcard[MAX_PATH]; /* Wildcard to search for. */ - WIN32_FIND_DATA findData; /* Data found by FindFirstFile() or FindNext(). */ + WCHAR wildcard[MAX_PATH]; /* Wildcard to search for. */ + WIN32_FIND_DATAW findData; /* Data found by FindFirstFile() or FindNext(). */ + /* Alignment is not honored, this works on x86 because of alignment fixup by processor. + Not perfect, but faster than alinging by hand (really) */ + WCHAR *wname = (WCHAR *) name; + WCHAR *wbuffer = (WCHAR *) buffer; /* * First time we must setup everything. */ if (*dir_handle == NULL) { - int length = strlen(name); - char* s; + int length = wcslen(wname); + WCHAR* s; if (length+3 >= MAX_PATH) { errno = ENAMETOOLONG; return check_error(-1, errInfo); } - strcpy(wildcard, name); + wcscpy(wildcard, wname); s = wildcard+length-1; - if (*s != '/' && *s != '\\') - *++s = '\\'; - *++s = '*'; - *++s = '\0'; - DEBUGF(("Reading %s\n", wildcard)); - dir = FindFirstFile(wildcard, &findData); + if (*s != L'/' && *s != L'\\') + *++s = L'\\'; + *++s = L'*'; + *++s = L'\0'; + DEBUGF(("Reading %ws\n", wildcard)); + dir = FindFirstFileW(wildcard, &findData); if (dir == INVALID_HANDLE_VALUE) return set_error(errInfo); *dir_handle = (EFILE_DIR_HANDLE) dir; if (!IS_DOT_OR_DOTDOT(findData.cFileName)) { - strcpy(buffer, findData.cFileName); + wcscpy(wbuffer, findData.cFileName); + *size = wcslen(wbuffer)*2; return 1; } } @@ -635,10 +640,11 @@ size_t size; /* Size of buffer. */ dir = (HANDLE) *dir_handle; for (;;) { - if (FindNextFile(dir, &findData)) { + if (FindNextFileW(dir, &findData)) { if (IS_DOT_OR_DOTDOT(findData.cFileName)) continue; - strcpy(buffer, findData.cFileName); + wcscpy(wbuffer, findData.cFileName); + *size = wcslen(wbuffer)*2; return 1; } @@ -655,17 +661,17 @@ size_t size; /* Size of buffer. */ } int -efile_openfile(errInfo, name, flags, pfd, pSize) -Efile_error* errInfo; /* Where to return error codes. */ -char* name; /* Name of directory to open. */ -int flags; /* Flags to use for opening. */ -int* pfd; /* Where to store the file descriptor. */ -Sint64* pSize; /* Where to store the size of the file. */ +efile_openfile(Efile_error* errInfo, /* Where to return error codes. */ + char* name, /* Name of directory to open. */ + int flags, /* Flags to use for opening. */ + int* pfd, /* Where to store the file descriptor. */ + Sint64* pSize) /* Where to store the size of the file. */ { BY_HANDLE_FILE_INFORMATION fileInfo; /* File information from a handle. */ HANDLE fd; /* Handle to open file. */ DWORD access; /* Access mode: GENERIC_READ, GENERIC_WRITE. */ DWORD crFlags; + WCHAR *wname = (WCHAR *) name; switch (flags & (EFILE_MODE_READ|EFILE_MODE_WRITE)) { case EFILE_MODE_READ: @@ -692,7 +698,7 @@ Sint64* pSize; /* Where to store the size of the file. */ if (flags & EFILE_MODE_EXCL) { crFlags = CREATE_NEW; } - fd = CreateFile(name, access, + fd = CreateFileW(wname, access, FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE, NULL, crFlags, FILE_ATTRIBUTE_NORMAL, NULL); @@ -711,7 +717,7 @@ Sint64* pSize; /* Where to store the size of the file. */ * to EISDIR. */ if (errInfo->posix_errno && - (attr = GetFileAttributes(name)) != INVALID_FILE_ATTRIBUTES && + (attr = GetFileAttributesW(wname)) != INVALID_FILE_ATTRIBUTES && (attr & FILE_ATTRIBUTE_DIRECTORY)) { errInfo->posix_errno = EISDIR; } @@ -735,9 +741,10 @@ Sint64* pSize; /* Where to store the size of the file. */ int efile_may_openfile(Efile_error* errInfo, char *name) { + WCHAR *wname = (WCHAR *) name; DWORD attr; - if ((attr = GetFileAttributes(name)) == INVALID_FILE_ATTRIBUTES) { + if ((attr = GetFileAttributesW(wname)) == INVALID_FILE_ATTRIBUTES) { return check_error(-1, errInfo); } @@ -746,18 +753,6 @@ efile_may_openfile(Efile_error* errInfo, char *name) { return check_error(-1, errInfo); } return 1; -#if 0 - struct stat statbuf; - - if (stat(name, &statbuf)) { - return check_error(-1, errInfo); - } - if (ISDIR(statbuf)) { - errno = EISDIR; - return check_error(-1, errInfo); - } - return 1; -#endif } void @@ -792,16 +787,17 @@ efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo, char* orig_name, int info_for_link) { HANDLE findhandle; /* Handle returned by FindFirstFile(). */ - WIN32_FIND_DATA findbuf; /* Data return by FindFirstFile(). */ - char name[_MAX_PATH]; + WIN32_FIND_DATAW findbuf; /* Data return by FindFirstFile(). */ + WCHAR name[_MAX_PATH]; int name_len; - char* path; - char pathbuf[_MAX_PATH]; + WCHAR *path; + WCHAR pathbuf[_MAX_PATH]; int drive; /* Drive for filename (1 = A:, 2 = B: etc). */ + WCHAR *worig_name = (WCHAR *) orig_name; /* Don't allow wildcards to be interpreted by system */ - if (strpbrk(orig_name, "?*")) { + if (wcspbrk(worig_name, L"?*")) { enoent: errInfo->posix_errno = ENOENT; errInfo->os_errno = ERROR_FILE_NOT_FOUND; @@ -813,25 +809,25 @@ efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo, * slash, because it causes FindFirstFile() to fail on Win95. */ - if ((name_len = strlen(orig_name)) >= _MAX_PATH) { + if ((name_len = wcslen(worig_name)) >= _MAX_PATH) { goto enoent; } else { - strcpy(name, orig_name); + wcscpy(name, worig_name); if (name_len > 2 && ISSLASH(name[name_len-1]) && - name[name_len-2] != ':') { - name[name_len-1] = '\0'; + name[name_len-2] != L':') { + name[name_len-1] = L'\0'; } } /* Try to get disk from name. If none, get current disk. */ - if (name[1] != ':') { + if (name[1] != L':') { drive = 0; - if (GetCurrentDirectory(sizeof(pathbuf), pathbuf) && - pathbuf[1] == ':') { - drive = tolower(pathbuf[0]) - 'a' + 1; + if (GetCurrentDirectoryW(_MAX_PATH, pathbuf) && + pathbuf[1] == L':') { + drive = towlower(pathbuf[0]) - L'a' + 1; } - } else if (*name && name[2] == '\0') { + } else if (*name && name[2] == L'\0') { /* * X: and nothing more is an error. */ @@ -839,15 +835,15 @@ efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo, errInfo->os_errno = ERROR_FILE_NOT_FOUND; return 0; } else - drive = tolower(*name) - 'a' + 1; + drive = towlower(*name) - L'a' + 1; - findhandle = FindFirstFile(name, &findbuf); + findhandle = FindFirstFileW(name, &findbuf); if (findhandle == INVALID_HANDLE_VALUE) { - if (!(strpbrk(name, "./\\") && - (path = _fullpath(pathbuf, name, _MAX_PATH)) && + if (!(wcspbrk(name, L"./\\") && + (path = _wfullpath(pathbuf, name, _MAX_PATH)) && /* root dir. ('C:\') or UNC root dir. ('\\server\share\') */ - ((strlen(path) == 3) || IsRootUNCName(path)) && - (GetDriveType(path) > 1) ) ) { + ((wcslen(path) == 3) || is_root_unc_name(path)) && + (GetDriveTypeW(path) > 1) ) ) { errInfo->posix_errno = ENOENT; errInfo->os_errno = ERROR_FILE_NOT_FOUND; return 0; @@ -860,7 +856,7 @@ efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo, findbuf.dwFileAttributes = FILE_ATTRIBUTE_DIRECTORY; findbuf.nFileSizeHigh = 0; findbuf.nFileSizeLow = 0; - findbuf.cFileName[0] = '\0'; + findbuf.cFileName[0] = L'\0'; pInfo->links = 1; pInfo->modifyTime.year = 1980; @@ -881,26 +877,28 @@ efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo, /* * given that we know this is a symlink, we should be able to find its target */ - char target_name[256]; - if (efile_readlink(errInfo, name, target_name,256) == 1) { + WCHAR target_name[_MAX_PATH]; + if (efile_readlink(errInfo, (char *) name, + (char *) target_name,256) == 1) { + FindClose(findhandle); return efile_fileinfo(errInfo, pInfo, - target_name, info_for_link); + (char *) target_name, info_for_link); } } -#if 0 /* number of links: */ { HANDLE handle; /* Handle returned by CreateFile() */ BY_HANDLE_FILE_INFORMATION fileInfo; /* from CreateFile() */ - if (handle = CreateFile(name, GENERIC_READ, 0,NULL, + if (handle = CreateFileW(name, GENERIC_READ, 0,NULL, OPEN_EXISTING, 0, NULL)) { GetFileInformationByHandle(handle, &fileInfo); pInfo->links = fileInfo.nNumberOfLinks; CloseHandle(handle); - } + } else { + pInfo->links = 1; + } } -#endif #define GET_TIME(dst, src) \ if (!FileTimeToLocalFileTime(&findbuf.src, &LocalFTime) || \ @@ -960,10 +958,9 @@ if (!FileTimeToLocalFileTime(&findbuf.src, &LocalFTime) || \ } int -efile_write_info(errInfo, pInfo, name) -Efile_error* errInfo; -Efile_info* pInfo; -char* name; +efile_write_info(Efile_error* errInfo, + Efile_info* pInfo, + char* name) { SYSTEMTIME timebuf; FILETIME LocalFileTime; @@ -977,12 +974,13 @@ char* name; DWORD attr; DWORD tempAttr; BOOL modifyTime = FALSE; + WCHAR *wname = (WCHAR *) name; /* * Get the attributes for the file. */ - tempAttr = attr = GetFileAttributes((LPTSTR)name); + tempAttr = attr = GetFileAttributesW(wname); if (attr == 0xffffffff) { return set_error(errInfo); } @@ -1036,12 +1034,12 @@ char* name; if (tempAttr & FILE_ATTRIBUTE_READONLY) { tempAttr &= ~FILE_ATTRIBUTE_READONLY; - if (!SetFileAttributes((LPTSTR) name, tempAttr)) { + if (!SetFileAttributesW(wname, tempAttr)) { return set_error(errInfo); } } - fd = CreateFile(name, GENERIC_READ|GENERIC_WRITE, + fd = CreateFileW(wname, GENERIC_READ|GENERIC_WRITE, FILE_SHARE_READ | FILE_SHARE_WRITE, NULL, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL); if (fd != INVALID_HANDLE_VALUE) { @@ -1059,7 +1057,7 @@ char* name; */ if (tempAttr != attr) { - if (!SetFileAttributes((LPTSTR) name, attr)) { + if (!SetFileAttributesW(wname, attr)) { return set_error(errInfo); } } @@ -1235,7 +1233,7 @@ int flags; /* - * IsRootUNCName - returns TRUE if the argument is a UNC name specifying + * is_root_unc_name - returns TRUE if the argument is a UNC name specifying * a root share. That is, if it is of the form \\server\share\. * This routine will also return true if the argument is of the * form \\server\share (no trailing slash) but Win32 currently @@ -1245,16 +1243,16 @@ int flags; */ static int -IsRootUNCName(const char* path) +is_root_unc_name(const WCHAR *path) { /* * If a root UNC name, path will start with 2 (but not 3) slashes */ - if ((strlen(path) >= 5) /* minimum string is "//x/y" */ + if ((wcslen(path) >= 5) /* minimum string is "//x/y" */ && ISSLASH(path[0]) && ISSLASH(path[1])) { - const char * p = path + 2 ; + const WCHAR *p = path + 2; /* * find the slash between the server name and share name @@ -1297,19 +1295,19 @@ IsRootUNCName(const char* path) */ static int -extract_root(char* name) +extract_root(WCHAR* name) { - int len = strlen(name); + int len = wcslen(name); - if (isalpha(name[0]) && name[1] == ':' && ISSLASH(name[2])) { - int c = name[3]; - name[3] = '\0'; - return c == '\0'; + if (iswalpha(name[0]) && name[1] == L':' && ISSLASH(name[2])) { + WCHAR c = name[3]; + name[3] = L'\0'; + return c == L'\0'; } else if (len < 5 || !ISSLASH(name[0]) || !ISSLASH(name[1])) { goto error; } else { /* Try to find the end of the UNC name. */ - char* p; - int c; + WCHAR* p; + WCHAR c; /* * Find the slash between the server name and share name. @@ -1318,7 +1316,7 @@ extract_root(char* name) for (p = name + 2; *p; p++) if (ISSLASH(*p)) break; - if (*p == '\0') + if (*p == L'\0') goto error; /* @@ -1329,24 +1327,24 @@ extract_root(char* name) if (ISSLASH(*p)) break; c = *p; - *p = '\0'; - return c == '\0' || p[1] == '\0'; + *p = L'\0'; + return c == L'\0' || p[1] == L'\0'; } error: - *name = '\0'; + *name = L'\0'; return 1; } static unsigned short -dos_to_posix_mode(int attr, const char *name) +dos_to_posix_mode(int attr, const WCHAR *name) { register unsigned short uxmode; unsigned dosmode; - register const char *p; + register const WCHAR *p; dosmode = attr & 0xff; - if ((p = name)[1] == ':') + if ((p = name)[1] == L':') p += 2; /* check to see if this is a directory - note we must make a special @@ -1355,7 +1353,7 @@ dos_to_posix_mode(int attr, const char *name) uxmode = (unsigned short) (((ISSLASH(*p) && !p[1]) || (dosmode & FILE_ATTRIBUTE_DIRECTORY) || - *p == '\0') ? _S_IFDIR|_S_IEXEC : _S_IFREG); + *p == L'\0') ? _S_IFDIR|_S_IEXEC : _S_IFREG); /* If attribute byte does not have read-only bit, it is read-write */ @@ -1364,11 +1362,11 @@ dos_to_posix_mode(int attr, const char *name) /* see if file appears to be executable - check extension of name */ - if (p = strrchr(name, '.')) { - if (!stricmp(p, ".exe") || - !stricmp(p, ".cmd") || - !stricmp(p, ".bat") || - !stricmp(p, ".com")) + if (p = wcsrchr(name, L'.')) { + if (!_wcsicmp(p, L".exe") || + !_wcsicmp(p, L".cmd") || + !_wcsicmp(p, L".bat") || + !_wcsicmp(p, L".com")) uxmode |= _S_IEXEC; } @@ -1388,28 +1386,40 @@ efile_readlink(Efile_error* errInfo, char* name, char* buffer, size_t size) * (Vista only) */ HINSTANCE hModule = NULL; + WCHAR *wname = (WCHAR *) name; + WCHAR *wbuffer = (WCHAR *) buffer; if ((hModule = LoadLibrary("kernel32.dll")) != NULL) { typedef DWORD (WINAPI * GETFINALPATHNAMEBYHANDLEPTR)( HANDLE hFile, - LPCSTR lpFilePath, + LPCWSTR lpFilePath, DWORD cchFilePath, DWORD dwFlags); GETFINALPATHNAMEBYHANDLEPTR pGetFinalPathNameByHandle = - (GETFINALPATHNAMEBYHANDLEPTR)GetProcAddress(hModule, "GetFinalPathNameByHandleA"); + (GETFINALPATHNAMEBYHANDLEPTR)GetProcAddress(hModule, "GetFinalPathNameByHandleW"); if (pGetFinalPathNameByHandle == NULL) { FreeLibrary(hModule); } else { /* first check if file is a symlink; {error, einval} otherwise */ - DWORD fileAttributes = GetFileAttributes(name); + DWORD fileAttributes = GetFileAttributesW(wname); if ((fileAttributes & FILE_ATTRIBUTE_REPARSE_POINT)) { BOOLEAN success = 0; - HANDLE h = CreateFile(name, GENERIC_READ, 0,NULL, OPEN_EXISTING, 0, NULL); + HANDLE h = CreateFileW(wname, GENERIC_READ, 0,NULL, OPEN_EXISTING, 0, NULL); + int len; if(h != INVALID_HANDLE_VALUE) { - success = pGetFinalPathNameByHandle(h, buffer, size,0); + success = pGetFinalPathNameByHandle(h, wbuffer, size,0); /* GetFinalPathNameByHandle prepends path with "\\?\": */ - sprintf(buffer, buffer+4); + len = wcslen(wbuffer); + wmemmove(wbuffer,wbuffer+4,len-3); + if (len - 4 >= 2 && wbuffer[1] == L':' && wbuffer[0] >= L'A' && + wbuffer[0] <= L'Z') { + wbuffer[0] = wbuffer[0] + L'a' - L'A'; + } + + for ( ; *wbuffer; wbuffer++) + if (*wbuffer == L'\\') + *wbuffer = L'/'; CloseHandle(h); } FreeLibrary(hModule); @@ -1433,17 +1443,20 @@ efile_readlink(Efile_error* errInfo, char* name, char* buffer, size_t size) int efile_altname(Efile_error* errInfo, char* orig_name, char* buffer, size_t size) { - WIN32_FIND_DATA wfd; + WIN32_FIND_DATAW wfd; HANDLE fh; - char name[_MAX_PATH]; + WCHAR name[_MAX_PATH+1]; int name_len; - char* path; - char pathbuf[_MAX_PATH]; + WCHAR* path; + WCHAR pathbuf[_MAX_PATH+1]; /* Unclear weather GetCurrentDirectory will access one char after + _MAX_PATH */ + WCHAR *worig_name = (WCHAR *) orig_name; + WCHAR *wbuffer = (WCHAR *) buffer; int drive; /* Drive for filename (1 = A:, 2 = B: etc). */ /* Don't allow wildcards to be interpreted by system */ - if (strpbrk(orig_name, "?*")) { + if (wcspbrk(worig_name, L"?*")) { enoent: errInfo->posix_errno = ENOENT; errInfo->os_errno = ERROR_FILE_NOT_FOUND; @@ -1455,61 +1468,64 @@ efile_altname(Efile_error* errInfo, char* orig_name, char* buffer, size_t size) * slash, because it causes FindFirstFile() to fail on Win95. */ - if ((name_len = strlen(orig_name)) >= _MAX_PATH) { + if ((name_len = wcslen(worig_name)) >= _MAX_PATH) { goto enoent; } else { - strcpy(name, orig_name); + wcscpy(name, worig_name); if (name_len > 2 && ISSLASH(name[name_len-1]) && - name[name_len-2] != ':') { - name[name_len-1] = '\0'; + name[name_len-2] != L':') { + name[name_len-1] = L'\0'; } } /* Try to get disk from name. If none, get current disk. */ - if (name[1] != ':') { + if (name[1] != L':') { drive = 0; - if (GetCurrentDirectory(sizeof(pathbuf), pathbuf) && - pathbuf[1] == ':') { - drive = tolower(pathbuf[0]) - 'a' + 1; + if (GetCurrentDirectoryW(_MAX_PATH, pathbuf) && + pathbuf[1] == L':') { + drive = towlower(pathbuf[0]) - L'a' + 1; } - } else if (*name && name[2] == '\0') { + } else if (*name && name[2] == L'\0') { /* * X: and nothing more is an error. */ goto enoent; } else { - drive = tolower(*name) - 'a' + 1; + drive = towlower(*name) - L'a' + 1; } - fh = FindFirstFile(name,&wfd); + fh = FindFirstFileW(name,&wfd); if (fh == INVALID_HANDLE_VALUE) { - if (!(strpbrk(name, "./\\") && - (path = _fullpath(pathbuf, name, _MAX_PATH)) && + if (!(wcspbrk(name, L"./\\") && + (path = _wfullpath(pathbuf, name, _MAX_PATH)) && /* root dir. ('C:\') or UNC root dir. ('\\server\share\') */ - ((strlen(path) == 3) || IsRootUNCName(path)) && - (GetDriveType(path) > 1) ) ) { + ((wcslen(path) == 3) || is_root_unc_name(path)) && + (GetDriveTypeW(path) > 1) ) ) { errno = errno_map(GetLastError()); return check_error(-1, errInfo); } /* * Root directories (such as C:\ or \\server\share\ are fabricated. */ - strcpy(buffer,name); + wcscpy(wbuffer,name); return 1; } - strcpy(buffer,wfd.cAlternateFileName); - if (!*buffer) { - strcpy(buffer,wfd.cFileName); + wcscpy(wbuffer,wfd.cAlternateFileName); + if (!*wbuffer) { + wcscpy(wbuffer,wfd.cFileName); } - + FindClose(fh); return 1; } + int efile_link(Efile_error* errInfo, char* old, char* new) { - if(!CreateHardLink(new, old, NULL)) { + WCHAR *wold = (WCHAR *) old; + WCHAR *wnew = (WCHAR *) new; + if(!CreateHardLinkW(wnew, wold, NULL)) { return set_error(errInfo); } return 1; @@ -1523,22 +1539,24 @@ efile_symlink(Efile_error* errInfo, char* old, char* new) * (Vista only) */ HINSTANCE hModule = NULL; + WCHAR *wold = (WCHAR *) old; + WCHAR *wnew = (WCHAR *) new; if ((hModule = LoadLibrary("kernel32.dll")) != NULL) { typedef BOOLEAN (WINAPI * CREATESYMBOLICLINKFUNCPTR) ( - LPCSTR lpSymlinkFileName, - LPCSTR lpTargetFileName, + LPCWSTR lpSymlinkFileName, + LPCWSTR lpTargetFileName, DWORD dwFlags); CREATESYMBOLICLINKFUNCPTR pCreateSymbolicLink = (CREATESYMBOLICLINKFUNCPTR) GetProcAddress(hModule, - "CreateSymbolicLinkA"); - /* A for MBCS, W for UNICODE... char* above implies 'A'! */ + "CreateSymbolicLinkW"); + /* A for MBCS, W for UNICODE... char* above implies 'W'! */ if (pCreateSymbolicLink != NULL) { - DWORD attr = GetFileAttributes(old); + DWORD attr = GetFileAttributesW(wold); int flag = (attr != INVALID_FILE_ATTRIBUTES && attr & FILE_ATTRIBUTE_DIRECTORY) ? 1 : 0; /* SYMBOLIC_LINK_FLAG_DIRECTORY = 1 */ - BOOLEAN success = pCreateSymbolicLink(new, old, flag); + BOOLEAN success = pCreateSymbolicLink(wnew, wold, flag); FreeLibrary(hModule); if (success) { diff --git a/erts/emulator/internal_doc/dec.dat b/erts/emulator/internal_doc/dec.dat new file mode 100644 index 0000000000..771ef51baa --- /dev/null +++ b/erts/emulator/internal_doc/dec.dat @@ -0,0 +1,942 @@ +{[59],894}. +{[96],8175}. +{[180],8189}. +{[183],903}. +{[198],1236}. +{[230],1237}. +{[399],1240}. +{[415],1256}. +{[439],1248}. +{[601],1241}. +{[629],1257}. +{[658],1249}. +{[697],884}. +{[768],832}. +{[768,65],192}. +{[768,69],200}. +{[768,73],204}. +{[768,79],210}. +{[768,85],217}. +{[768,87],7808}. +{[768,89],7922}. +{[768,97],224}. +{[768,101],232}. +{[768,105],236}. +{[768,111],242}. +{[768,117],249}. +{[768,119],7809}. +{[768,121],7923}. +{[768,168],8173}. +{[768,770,65],7846}. +{[768,770,69],7872}. +{[768,770,79],7890}. +{[768,770,97],7847}. +{[768,770,101],7873}. +{[768,770,111],7891}. +{[768,772,69],7700}. +{[768,772,79],7760}. +{[768,772,101],7701}. +{[768,772,111],7761}. +{[768,774,65],7856}. +{[768,774,97],7857}. +{[768,776,85],475}. +{[768,776,117],476}. +{[768,776,953],8146}. +{[768,776,965],8162}. +{[768,787,837,913],8074}. +{[768,787,837,919],8090}. +{[768,787,837,937],8106}. +{[768,787,837,945],8066}. +{[768,787,837,951],8082}. +{[768,787,837,969],8098}. +{[768,787,913],7946}. +{[768,787,917],7962}. +{[768,787,919],7978}. +{[768,787,921],7994}. +{[768,787,927],8010}. +{[768,787,937],8042}. +{[768,787,945],7938}. +{[768,787,949],7954}. +{[768,787,951],7970}. +{[768,787,953],7986}. +{[768,787,959],8002}. +{[768,787,965],8018}. +{[768,787,969],8034}. +{[768,788,837,913],8075}. +{[768,788,837,919],8091}. +{[768,788,837,937],8107}. +{[768,788,837,945],8067}. +{[768,788,837,951],8083}. +{[768,788,837,969],8099}. +{[768,788,913],7947}. +{[768,788,917],7963}. +{[768,788,919],7979}. +{[768,788,921],7995}. +{[768,788,927],8011}. +{[768,788,933],8027}. +{[768,788,937],8043}. +{[768,788,945],7939}. +{[768,788,949],7955}. +{[768,788,951],7971}. +{[768,788,953],7987}. +{[768,788,959],8003}. +{[768,788,965],8019}. +{[768,788,969],8035}. +{[768,795,79],7900}. +{[768,795,85],7914}. +{[768,795,111],7901}. +{[768,795,117],7915}. +{[768,837,945],8114}. +{[768,837,951],8130}. +{[768,837,969],8178}. +{[768,913],8122}. +{[768,917],8136}. +{[768,919],8138}. +{[768,921],8154}. +{[768,927],8184}. +{[768,933],8170}. +{[768,937],8186}. +{[768,945],8048}. +{[768,949],8050}. +{[768,951],8052}. +{[768,953],8054}. +{[768,959],8056}. +{[768,965],8058}. +{[768,969],8060}. +{[768,8127],8141}. +{[768,8190],8157}. +{[769],833}. +{[769,65],193}. +{[769,67],262}. +{[769,69],201}. +{[769,71],500}. +{[769,73],205}. +{[769,75],7728}. +{[769,76],313}. +{[769,77],7742}. +{[769,78],323}. +{[769,79],211}. +{[769,80],7764}. +{[769,82],340}. +{[769,83],346}. +{[769,85],218}. +{[769,87],7810}. +{[769,89],221}. +{[769,90],377}. +{[769,97],225}. +{[769,99],263}. +{[769,101],233}. +{[769,103],501}. +{[769,105],237}. +{[769,107],7729}. +{[769,108],314}. +{[769,109],7743}. +{[769,110],324}. +{[769,111],243}. +{[769,112],7765}. +{[769,114],341}. +{[769,115],347}. +{[769,117],250}. +{[769,119],7811}. +{[769,121],253}. +{[769,122],378}. +{[769,168],8174}. +{[769,198],508}. +{[769,216],510}. +{[769,230],509}. +{[769,248],511}. +{[769,770,65],7844}. +{[769,770,69],7870}. +{[769,770,79],7888}. +{[769,770,97],7845}. +{[769,770,101],7871}. +{[769,770,111],7889}. +{[769,771,79],7756}. +{[769,771,85],7800}. +{[769,771,111],7757}. +{[769,771,117],7801}. +{[769,772,69],7702}. +{[769,772,79],7762}. +{[769,772,101],7703}. +{[769,772,111],7763}. +{[769,774,65],7854}. +{[769,774,97],7855}. +{[769,776,73],7726}. +{[769,776,85],471}. +{[769,776,105],7727}. +{[769,776,117],472}. +{[769,776,953],8147}. +{[769,776,965],8163}. +{[769,778,65],506}. +{[769,778,97],507}. +{[769,787,837,913],8076}. +{[769,787,837,919],8092}. +{[769,787,837,937],8108}. +{[769,787,837,945],8068}. +{[769,787,837,951],8084}. +{[769,787,837,969],8100}. +{[769,787,913],7948}. +{[769,787,917],7964}. +{[769,787,919],7980}. +{[769,787,921],7996}. +{[769,787,927],8012}. +{[769,787,937],8044}. +{[769,787,945],7940}. +{[769,787,949],7956}. +{[769,787,951],7972}. +{[769,787,953],7988}. +{[769,787,959],8004}. +{[769,787,965],8020}. +{[769,787,969],8036}. +{[769,788,837,913],8077}. +{[769,788,837,919],8093}. +{[769,788,837,937],8109}. +{[769,788,837,945],8069}. +{[769,788,837,951],8085}. +{[769,788,837,969],8101}. +{[769,788,913],7949}. +{[769,788,917],7965}. +{[769,788,919],7981}. +{[769,788,921],7997}. +{[769,788,927],8013}. +{[769,788,933],8029}. +{[769,788,937],8045}. +{[769,788,945],7941}. +{[769,788,949],7957}. +{[769,788,951],7973}. +{[769,788,953],7989}. +{[769,788,959],8005}. +{[769,788,965],8021}. +{[769,788,969],8037}. +{[769,795,79],7898}. +{[769,795,85],7912}. +{[769,795,111],7899}. +{[769,795,117],7913}. +{[769,807,67],7688}. +{[769,807,99],7689}. +{[769,837,945],8116}. +{[769,837,951],8132}. +{[769,837,959],8180}. +{[769,913],8123}. +{[769,917],8137}. +{[769,919],8139}. +{[769,921],8155}. +{[769,927],8185}. +{[769,933],8171}. +{[769,937],8187}. +{[769,945],8049}. +{[769,949],8051}. +{[769,951],8053}. +{[769,953],8055}. +{[769,959],8057}. +{[769,965],8059}. +{[769,969],8061}. +{[769,1043],1027}. +{[769,1050],1036}. +{[769,1075],1107}. +{[769,1082],1116}. +{[769,8127],8142}. +{[769,8190],8158}. +{[770,65],194}. +{[770,67],264}. +{[770,69],202}. +{[770,71],284}. +{[770,72],292}. +{[770,73],206}. +{[770,74],308}. +{[770,79],212}. +{[770,83],348}. +{[770,85],219}. +{[770,87],372}. +{[770,89],374}. +{[770,90],7824}. +{[770,97],226}. +{[770,99],265}. +{[770,101],234}. +{[770,103],285}. +{[770,104],293}. +{[770,105],238}. +{[770,106],309}. +{[770,111],244}. +{[770,115],349}. +{[770,117],251}. +{[770,119],373}. +{[770,121],375}. +{[770,122],7825}. +{[770,803,65],7852}. +{[770,803,69],7878}. +{[770,803,79],7896}. +{[770,803,97],7853}. +{[770,803,101],7879}. +{[770,803,111],7897}. +{[771,65],195}. +{[771,69],7868}. +{[771,73],296}. +{[771,78],209}. +{[771,79],213}. +{[771,85],360}. +{[771,86],7804}. +{[771,89],7928}. +{[771,97],227}. +{[771,101],7869}. +{[771,105],297}. +{[771,110],241}. +{[771,111],245}. +{[771,117],361}. +{[771,118],7805}. +{[771,121],7929}. +{[771,770,65],7850}. +{[771,770,69],7876}. +{[771,770,79],7894}. +{[771,770,97],7851}. +{[771,770,101],7877}. +{[771,770,111],7895}. +{[771,774,65],7860}. +{[771,774,97],7861}. +{[771,795,79],7904}. +{[771,795,85],7918}. +{[771,795,111],7905}. +{[771,795,117],7919}. +{[772,65],256}. +{[772,69],274}. +{[772,71],7712}. +{[772,73],298}. +{[772,79],332}. +{[772,85],362}. +{[772,97],257}. +{[772,101],275}. +{[772,103],7713}. +{[772,105],299}. +{[772,111],333}. +{[772,117],363}. +{[772,198],482}. +{[772,230],483}. +{[772,775,65],480}. +{[772,775,97],481}. +{[772,776,65],478}. +{[772,776,85],469}. +{[772,776,97],479}. +{[772,776,117],470}. +{[772,803,76],7736}. +{[772,803,82],7772}. +{[772,803,108],7737}. +{[772,803,114],7773}. +{[772,808,79],492}. +{[772,808,111],493}. +{[772,913],8121}. +{[772,921],8153}. +{[772,933],8169}. +{[772,945],8113}. +{[772,953],8145}. +{[772,965],8161}. +{[772,1048],1250}. +{[772,1059],1262}. +{[772,1080],1251}. +{[772,1091],1263}. +{[774,65],258}. +{[774,69],276}. +{[774,71],286}. +{[774,73],300}. +{[774,79],334}. +{[774,85],364}. +{[774,97],259}. +{[774,101],277}. +{[774,103],287}. +{[774,105],301}. +{[774,111],335}. +{[774,117],365}. +{[774,803,65],7862}. +{[774,803,97],7863}. +{[774,807,69],7708}. +{[774,807,101],7709}. +{[774,913],8120}. +{[774,921],8152}. +{[774,933],8168}. +{[774,945],8112}. +{[774,953],8144}. +{[774,965],8160}. +{[774,1040],1232}. +{[774,1045],1238}. +{[774,1046],1217}. +{[774,1048],1049}. +{[774,1059],1038}. +{[774,1072],1233}. +{[774,1077],1239}. +{[774,1078],1218}. +{[774,1080],1081}. +{[774,1091],1118}. +{[775,66],7682}. +{[775,67],266}. +{[775,68],7690}. +{[775,69],278}. +{[775,70],7710}. +{[775,71],288}. +{[775,72],7714}. +{[775,73],304}. +{[775,77],7744}. +{[775,78],7748}. +{[775,80],7766}. +{[775,82],7768}. +{[775,83],7776}. +{[775,84],7786}. +{[775,87],7814}. +{[775,88],7818}. +{[775,89],7822}. +{[775,90],379}. +{[775,98],7683}. +{[775,99],267}. +{[775,100],7691}. +{[775,101],279}. +{[775,102],7711}. +{[775,103],289}. +{[775,104],7715}. +{[775,109],7745}. +{[775,110],7749}. +{[775,112],7767}. +{[775,114],7769}. +{[775,115],7777}. +{[775,116],7787}. +{[775,119],7815}. +{[775,120],7819}. +{[775,121],7823}. +{[775,122],380}. +{[775,383],7835}. +{[775,769,83],7780}. +{[775,769,115],7781}. +{[775,774],784}. +{[775,780,83],7782}. +{[775,780,115],7783}. +{[775,803,83],7784}. +{[775,803,115],7785}. +{[776,65],196}. +{[776,69],203}. +{[776,72],7718}. +{[776,73],207}. +{[776,79],214}. +{[776,85],220}. +{[776,87],7812}. +{[776,88],7820}. +{[776,89],376}. +{[776,97],228}. +{[776,101],235}. +{[776,104],7719}. +{[776,105],239}. +{[776,111],246}. +{[776,116],7831}. +{[776,117],252}. +{[776,119],7813}. +{[776,120],7821}. +{[776,121],255}. +{[776,399],1242}. +{[776,415],1258}. +{[776,601],1243}. +{[776,629],1259}. +{[776,771,79],7758}. +{[776,771,111],7759}. +{[776,772,85],7802}. +{[776,772,117],7803}. +{[776,921],938}. +{[776,933],939}. +{[776,953],970}. +{[776,965],971}. +{[776,978],980}. +{[776,1030],1031}. +{[776,1040],1234}. +{[776,1045],1025}. +{[776,1046],1244}. +{[776,1047],1246}. +{[776,1048],1252}. +{[776,1054],1254}. +{[776,1059],1264}. +{[776,1063],1268}. +{[776,1067],1272}. +{[776,1072],1235}. +{[776,1077],1105}. +{[776,1078],1245}. +{[776,1079],1247}. +{[776,1080],1253}. +{[776,1086],1255}. +{[776,1091],1265}. +{[776,1095],1269}. +{[776,1099],1273}. +{[776,1110],1111}. +{[777,65],7842}. +{[777,69],7866}. +{[777,73],7880}. +{[777,79],7886}. +{[777,85],7910}. +{[777,89],7926}. +{[777,97],7843}. +{[777,101],7867}. +{[777,105],7881}. +{[777,111],7887}. +{[777,117],7911}. +{[777,121],7927}. +{[777,770,65],7848}. +{[777,770,69],7874}. +{[777,770,79],7892}. +{[777,770,97],7849}. +{[777,770,101],7875}. +{[777,770,111],7893}. +{[777,774,65],7858}. +{[777,774,97],7859}. +{[777,795,79],7902}. +{[777,795,85],7916}. +{[777,795,111],7903}. +{[777,795,117],7917}. +{[778,65],197}. +{[778,85],366}. +{[778,97],229}. +{[778,117],367}. +{[778,119],7832}. +{[778,121],7833}. +{[779,79],336}. +{[779,85],368}. +{[779,111],337}. +{[779,117],369}. +{[779,1059],1266}. +{[779,1091],1267}. +{[780,65],461}. +{[780,67],268}. +{[780,68],270}. +{[780,69],282}. +{[780,71],486}. +{[780,73],463}. +{[780,75],488}. +{[780,76],317}. +{[780,78],327}. +{[780,79],465}. +{[780,82],344}. +{[780,83],352}. +{[780,84],356}. +{[780,85],467}. +{[780,90],381}. +{[780,97],462}. +{[780,99],269}. +{[780,100],271}. +{[780,101],283}. +{[780,103],487}. +{[780,105],464}. +{[780,106],496}. +{[780,107],489}. +{[780,108],318}. +{[780,110],328}. +{[780,111],466}. +{[780,114],345}. +{[780,115],353}. +{[780,116],357}. +{[780,117],468}. +{[780,122],382}. +{[780,439],494}. +{[780,658],495}. +{[780,776,85],473}. +{[780,776,117],474}. +{[781,168],901}. +{[781,776],836}. +{[781,776,953],912}. +{[781,776,965],944}. +{[781,913],902}. +{[781,917],904}. +{[781,919],905}. +{[781,921],906}. +{[781,927],908}. +{[781,933],910}. +{[781,937],911}. +{[781,945],940}. +{[781,949],941}. +{[781,951],942}. +{[781,953],943}. +{[781,959],972}. +{[781,965],973}. +{[781,969],974}. +{[781,978],979}. +{[783,65],512}. +{[783,69],516}. +{[783,73],520}. +{[783,79],524}. +{[783,82],528}. +{[783,85],532}. +{[783,97],513}. +{[783,101],517}. +{[783,105],521}. +{[783,111],525}. +{[783,114],529}. +{[783,117],533}. +{[783,1140],1142}. +{[783,1141],1143}. +{[785,65],514}. +{[785,69],518}. +{[785,73],522}. +{[785,79],526}. +{[785,82],530}. +{[785,85],534}. +{[785,97],515}. +{[785,101],519}. +{[785,105],523}. +{[785,111],527}. +{[785,114],531}. +{[785,117],535}. +{[787],835}. +{[787,837,913],8072}. +{[787,837,919],8088}. +{[787,837,937],8104}. +{[787,837,945],8064}. +{[787,837,951],8080}. +{[787,837,969],8096}. +{[787,913],7944}. +{[787,917],7960}. +{[787,919],7976}. +{[787,921],7992}. +{[787,927],8008}. +{[787,937],8040}. +{[787,945],7936}. +{[787,949],7952}. +{[787,951],7968}. +{[787,953],7984}. +{[787,959],8000}. +{[787,961],8164}. +{[787,965],8016}. +{[787,969],8032}. +{[788,837,913],8073}. +{[788,837,919],8089}. +{[788,837,937],8105}. +{[788,837,945],8065}. +{[788,837,951],8081}. +{[788,837,969],8097}. +{[788,913],7945}. +{[788,917],7961}. +{[788,919],7977}. +{[788,921],7993}. +{[788,927],8009}. +{[788,929],8172}. +{[788,933],8025}. +{[788,937],8041}. +{[788,945],7937}. +{[788,949],7953}. +{[788,951],7969}. +{[788,953],7985}. +{[788,959],8001}. +{[788,961],8165}. +{[788,965],8017}. +{[788,969],8033}. +{[795,79],416}. +{[795,85],431}. +{[795,111],417}. +{[795,117],432}. +{[803,65],7840}. +{[803,66],7684}. +{[803,68],7692}. +{[803,69],7864}. +{[803,72],7716}. +{[803,73],7882}. +{[803,75],7730}. +{[803,76],7734}. +{[803,77],7746}. +{[803,78],7750}. +{[803,79],7884}. +{[803,82],7770}. +{[803,83],7778}. +{[803,84],7788}. +{[803,85],7908}. +{[803,86],7806}. +{[803,87],7816}. +{[803,89],7924}. +{[803,90],7826}. +{[803,97],7841}. +{[803,98],7685}. +{[803,100],7693}. +{[803,101],7865}. +{[803,104],7717}. +{[803,105],7883}. +{[803,107],7731}. +{[803,108],7735}. +{[803,109],7747}. +{[803,110],7751}. +{[803,111],7885}. +{[803,114],7771}. +{[803,115],7779}. +{[803,116],7789}. +{[803,117],7909}. +{[803,118],7807}. +{[803,119],7817}. +{[803,121],7925}. +{[803,122],7827}. +{[803,795,79],7906}. +{[803,795,85],7920}. +{[803,795,111],7907}. +{[803,795,117],7921}. +{[804,85],7794}. +{[804,117],7795}. +{[805,65],7680}. +{[805,97],7681}. +{[807,67],199}. +{[807,68],7696}. +{[807,71],290}. +{[807,72],7720}. +{[807,75],310}. +{[807,76],315}. +{[807,78],325}. +{[807,82],342}. +{[807,83],350}. +{[807,84],354}. +{[807,99],231}. +{[807,100],7697}. +{[807,103],291}. +{[807,104],7721}. +{[807,107],311}. +{[807,108],316}. +{[807,110],326}. +{[807,114],343}. +{[807,115],351}. +{[807,116],355}. +{[808,65],260}. +{[808,69],280}. +{[808,73],302}. +{[808,79],490}. +{[808,85],370}. +{[808,97],261}. +{[808,101],281}. +{[808,105],303}. +{[808,111],491}. +{[808,117],371}. +{[813,68],7698}. +{[813,69],7704}. +{[813,76],7740}. +{[813,78],7754}. +{[813,84],7792}. +{[813,85],7798}. +{[813,100],7699}. +{[813,101],7705}. +{[813,108],7741}. +{[813,110],7755}. +{[813,116],7793}. +{[813,117],7799}. +{[814,72],7722}. +{[814,104],7723}. +{[816,69],7706}. +{[816,73],7724}. +{[816,85],7796}. +{[816,101],7707}. +{[816,105],7725}. +{[816,117],7797}. +{[817,66],7686}. +{[817,68],7694}. +{[817,75],7732}. +{[817,76],7738}. +{[817,78],7752}. +{[817,82],7774}. +{[817,84],7790}. +{[817,90],7828}. +{[817,98],7687}. +{[817,100],7695}. +{[817,104],7830}. +{[817,107],7733}. +{[817,108],7739}. +{[817,110],7753}. +{[817,114],7775}. +{[817,116],7791}. +{[817,122],7829}. +{[834,168],8129}. +{[834,776,953],8151}. +{[834,776,965],8167}. +{[834,787,837,913],8078}. +{[834,787,837,919],8094}. +{[834,787,837,937],8110}. +{[834,787,837,945],8070}. +{[834,787,837,951],8086}. +{[834,787,837,969],8102}. +{[834,787,913],7950}. +{[834,787,919],7982}. +{[834,787,921],7998}. +{[834,787,937],8046}. +{[834,787,945],7942}. +{[834,787,951],7974}. +{[834,787,953],7990}. +{[834,787,965],8022}. +{[834,787,969],8038}. +{[834,788,837,913],8079}. +{[834,788,837,919],8095}. +{[834,788,837,937],8111}. +{[834,788,837,945],8071}. +{[834,788,837,951],8087}. +{[834,788,837,969],8103}. +{[834,788,913],7951}. +{[834,788,919],7983}. +{[834,788,921],7999}. +{[834,788,933],8031}. +{[834,788,937],8047}. +{[834,788,945],7943}. +{[834,788,951],7975}. +{[834,788,953],7991}. +{[834,788,965],8023}. +{[834,788,969],8039}. +{[834,837,945],8119}. +{[834,837,951],8135}. +{[834,837,969],8183}. +{[834,945],8118}. +{[834,951],8134}. +{[834,953],8150}. +{[834,965],8166}. +{[834,969],8182}. +{[834,8127],8143}. +{[834,8190],8159}. +{[837,913],8124}. +{[837,919],8140}. +{[837,937],8188}. +{[837,945],8115}. +{[837,951],8131}. +{[837,969],8179}. +{[953],8126}. +{[1463,1488],64302}. +{[1463,1522],64287}. +{[1464,1488],64303}. +{[1465,1493],64331}. +{[1468,1488],64304}. +{[1468,1489],64305}. +{[1468,1490],64306}. +{[1468,1491],64307}. +{[1468,1492],64308}. +{[1468,1493],64309}. +{[1468,1494],64310}. +{[1468,1496],64312}. +{[1468,1497],64313}. +{[1468,1498],64314}. +{[1468,1499],64315}. +{[1468,1500],64316}. +{[1468,1502],64318}. +{[1468,1504],64320}. +{[1468,1505],64321}. +{[1468,1507],64323}. +{[1468,1508],64324}. +{[1468,1510],64326}. +{[1468,1511],64327}. +{[1468,1512],64328}. +{[1468,1513],64329}. +{[1468,1514],64330}. +{[1471,1489],64332}. +{[1471,1499],64333}. +{[1471,1508],64334}. +{[1473,1468,1513],64300}. +{[1473,1513],64298}. +{[1474,1468,1513],64301}. +{[1474,1513],64299}. +{[2364,2325],2392}. +{[2364,2326],2393}. +{[2364,2327],2394}. +{[2364,2332],2395}. +{[2364,2337],2396}. +{[2364,2338],2397}. +{[2364,2344],2345}. +{[2364,2347],2398}. +{[2364,2351],2399}. +{[2364,2352],2353}. +{[2364,2355],2356}. +{[2492,2465],2524}. +{[2492,2466],2525}. +{[2492,2476],2480}. +{[2492,2479],2527}. +{[2494,2503],2507}. +{[2519,2503],2508}. +{[2620,2582],2649}. +{[2620,2583],2650}. +{[2620,2588],2651}. +{[2620,2593],2652}. +{[2620,2603],2654}. +{[2876,2849],2908}. +{[2876,2850],2909}. +{[2876,2863],2911}. +{[2878,2887],2891}. +{[2902,2887],2888}. +{[2903,2887],2892}. +{[3006,3014],3018}. +{[3006,3015],3019}. +{[3031,2962],2964}. +{[3031,3014],3020}. +{[3158,3142],3144}. +{[3266,3270],3274}. +{[3285,3263],3264}. +{[3285,3266,3270],3275}. +{[3285,3270],3271}. +{[3286,3270],3272}. +{[3390,3398],3402}. +{[3390,3399],3403}. +{[3415,3398],3404}. +{[3634,3661],3635}. +{[3762,3789],3763}. +{[3953,3954],3955}. +{[3953,3956],3957}. +{[3953,3968],3969}. +{[3953,3968,4018],3959}. +{[3953,3968,4019],3961}. +{[3968,4018],3958}. +{[3968,4019],3960}. +{[4021,3904],3945}. +{[4021,3984],4025}. +{[4023,3906],3907}. +{[4023,3916],3917}. +{[4023,3921],3922}. +{[4023,3926],3927}. +{[4023,3931],3932}. +{[4023,3986],3987}. +{[4023,3996],3997}. +{[4023,4001],4002}. +{[4023,4006],4007}. +{[4023,4011],4012}. +{[12441,12358],12436}. +{[12441,12363],12364}. +{[12441,12365],12366}. +{[12441,12367],12368}. +{[12441,12369],12370}. +{[12441,12371],12372}. +{[12441,12373],12374}. +{[12441,12375],12376}. +{[12441,12377],12378}. +{[12441,12379],12380}. +{[12441,12381],12382}. +{[12441,12383],12384}. +{[12441,12385],12386}. +{[12441,12388],12389}. +{[12441,12390],12391}. +{[12441,12392],12393}. +{[12441,12399],12400}. +{[12441,12402],12403}. +{[12441,12405],12406}. +{[12441,12408],12409}. +{[12441,12411],12412}. +{[12441,12445],12446}. +{[12441,12454],12532}. +{[12441,12459],12460}. +{[12441,12461],12462}. +{[12441,12463],12464}. +{[12441,12465],12466}. +{[12441,12467],12468}. +{[12441,12469],12470}. +{[12441,12471],12472}. +{[12441,12473],12474}. +{[12441,12475],12476}. +{[12441,12477],12478}. +{[12441,12479],12480}. +{[12441,12481],12482}. +{[12441,12484],12485}. +{[12441,12486],12487}. +{[12441,12488],12489}. +{[12441,12495],12496}. +{[12441,12498],12499}. +{[12441,12501],12502}. +{[12441,12504],12505}. +{[12441,12507],12508}. +{[12441,12527],12535}. +{[12441,12528],12536}. +{[12441,12529],12537}. +{[12441,12530],12538}. +{[12441,12541],12542}. +{[12442,12399],12401}. +{[12442,12402],12404}. +{[12442,12405],12407}. +{[12442,12408],12410}. +{[12442,12411],12413}. +{[12442,12495],12497}. +{[12442,12498],12500}. +{[12442,12501],12503}. +{[12442,12504],12506}. +{[12442,12507],12509}. diff --git a/erts/emulator/internal_doc/dec.erl b/erts/emulator/internal_doc/dec.erl new file mode 100644 index 0000000000..0315f2a52d --- /dev/null +++ b/erts/emulator/internal_doc/dec.erl @@ -0,0 +1,237 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2000-2010. All Rights Reserved. +%% +%% The contents of this file are subject to the Erlang Public License, +%% Version 1.1, (the "License"); you may not use this file except in +%% compliance with the License. You should have received a copy of the +%% Erlang Public License along with this software. If not, it can be +%% retrieved online at http://www.erlang.org/. +%% +%% Software distributed under the License is distributed on an "AS IS" +%% basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See +%% the License for the specific language governing rights and limitations +%% under the License. +%% +%% %CopyrightEnd% +%% + +%% This program is used to generate a header file with data for +%% normalizing denormalized unicode. + +%% The C header is generated from a text file containing tuples in the +%% following format: +%% {RevList,Translation} +%% Where 'RevList' is a reversed list of the denormalized repressentation of +%% the character 'Translation'. An example would be the swedish character +%% '�', which would be represented in the file as: +%% {[776,111],246}, as the denormalized representation of codepoint 246 +%% is [111,776] (i.e an 'o' followed by the "double dot accent character 776), +%% while '�' instead is represented as {[776,97],228}, as the denormalized +%% form would be [97,776] (same accent but an 'a' instead). +%% The datafile is generated from the table on Apple's developer connection +%% http://developer.apple.com/library/mac/#technotes/tn/tn1150table.html +%% The generating is done whenever new data is present (i.e. dec.dat has +%% to be changed) and not for every build. The product (the C header) is copied +%% to $ERL_TOP/erts/beam after generation and checked in. +%% The program and the data file is included for reference. + +-module(dec). + +-compile(export_all). + +-define(HASH_SIZE_FACTOR,2). +-define(BIG_PREFIX_SIZE,392). + +-define(INPUT_FILE_NAME,"dec.dat"). +-define(OUTPUT_FILE_NAME,"erl_unicode_normalize.h"). + +read(FName) -> + {ok,L} = file:consult(FName), + [{A,B} || {A,B} <- L, + length(A) > 1% , hd(A) < 769 + ]. + +dec() -> + L = read(?INPUT_FILE_NAME), + G = group(L), + {ok,Out} = file:open(?OUTPUT_FILE_NAME,[write]), + io:format + (Out, + "/*~n" + "* %CopyrightBegin%~n" + "*~n" + "* Copyright Ericsson AB 1999-2010. All Rights Reserved.~n" + "*~n" + "* The contents of this file are subject to the Erlang Public License,~n" + "* Version 1.1, (the \"License\"); you may not use this file except in~n" + "* compliance with the License. You should have received a copy of the~n" + "* Erlang Public License along with this software. If not, it can be~n" + "* retrieved online at http://www.erlang.org/.~n" + "*~n" + "* Software distributed under the License is distributed on an " + "\"AS IS\"~n" + "* basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See~n" + "* the License for the specific language governing rights and " + "limitations~n" + "* under the License.~n" + "*~n" + "* %CopyrightEnd%~n" + "*/~n" + "/*~n" + "* This file is automatically generated by ~p.erl, " + "do not edit manually~n" + "*/~n", + [?MODULE]), + + io:format(Out, + "#define HASH_SIZE_FACTOR ~w~n" + "typedef struct _compose_entry {~n" + " Uint16 c;~n" + " Uint16 res;~n" + " Uint16 num_subs;~n" + " struct _compose_entry *subs;~n" + " int *hash;~n" + "} CompEntry;~n~n" + "static int compose_tab_size = ~p;~n", + [?HASH_SIZE_FACTOR,length(G)]), + d(Out,G,[],0), + PreTab = tuple_to_list(make_prefix_table(G,erlang:make_tuple(102,0))), + dump_prefixes(Out,PreTab), +%% Using this cuts down on the searching in the +%% actual implementation, but wastes memory with little real gain.. +%% LL = lists:flatten([PartList || {PartList,_} <- L]), +%% BigPreTab = tuple_to_list( +%% make_big_prefixes(LL, +%% erlang:make_tuple(?BIG_PREFIX_SIZE,0))), +%% dump_big_prefixes(Out,BigPreTab), + file:close(Out), + ok. + + + +d(Out,List,D,C) -> + d_sub(Out,List,D,C), + d_top_hash(Out,List,D,C), + d_top(Out,List,D,C). +d_sub(_Out,[],_D,_C) -> + ok; +d_sub(Out,[{_CP,[],_Res}|T],D,C) -> + d_sub(Out,T,D,C+1); +d_sub(Out,[{_CP,Subs,_Res0}|T],D,C) -> + d(Out,Subs,[C|D],0), + d_sub(Out,T,D,C+1). +d_top(Out,L,D,C) -> + io:format(Out,"static CompEntry ~s[] = {~n",[format_depth(D)]), + d_top_1(Out,L,D,C), + io:format(Out,"}; /* ~s */ ~n",[format_depth(D)]). + +d_top_1(_Out,[],_D,_C) -> + ok; +d_top_1(Out,[{CP,[],Res}|T],D,C) -> + io:format(Out, + "{~w, ~w, 0, NULL, NULL}",[CP,Res]), + if + T =:= [] -> + io:format(Out,"~n",[]); + true -> + io:format(Out,",~n",[]) + end, + d_top_1(Out,T,D,C+1); +d_top_1(Out,[{CP,Subs,_Res}|T],D,C) -> + io:format(Out, + "{~w, 0, ~w, ~s, ~s}",[CP,length(Subs), + format_depth([C|D]), + "hash_"++format_depth([C|D])]), + if + T =:= [] -> + io:format(Out,"~n",[]); + true -> + io:format(Out,",~n",[]) + end, + d_top_1(Out,T,D,C+1). + + +d_top_hash(Out,List,D,_C) -> + HSize = length(List)*?HASH_SIZE_FACTOR, + io:format(Out,"static int ~s[~p] = ~n",["hash_"++format_depth(D),HSize]), + Tup = d_top_hash_1(List,0,erlang:make_tuple(HSize,-1),HSize), + io:format(Out,"~p; /* ~s */ ~n",[Tup,"hash_"++format_depth(D)]). + +d_top_hash_1([],_,Hash,_HSize) -> + Hash; +d_top_hash_1([{CP,_,_}|T],Index,Hash,HSize) -> + Bucket = hash_search(Hash,HSize,CP rem HSize), + d_top_hash_1(T,Index+1,erlang:setelement(Bucket+1,Hash,Index),HSize). + +hash_search(Hash,_HSize,Bucket) when element(Bucket+1,Hash) =:= -1 -> + Bucket; +hash_search(Hash,HSize,Bucket) -> + hash_search(Hash,HSize,(Bucket + 1) rem HSize). + +format_depth(D) -> + lists:reverse(tl(lists:reverse(lists:flatten(["compose_tab_",[ integer_to_list(X) ++ "_" || X <- lists:reverse(D) ]])))). + + + + +make_prefix_table([],Table) -> + Table; +make_prefix_table([{C,_,_}|T],Table) when C =< 4023 -> + Index = (C div 32) + 1 - 24, + Pos = C rem 32, + X = element(Index,Table), + Y = X bor (1 bsl Pos), + NewTab = setelement(Index,Table,Y), + make_prefix_table(T,NewTab); +make_prefix_table([_|T],Tab) -> + make_prefix_table(T,Tab). + +dump_prefixes(Out,L) -> + io:format(Out,"#define COMP_CANDIDATE_MAP_OFFSET 24~n",[]), + io:format(Out,"static Uint32 comp_candidate_map[] = {~n",[]), + dump_prefixes_1(Out,L). +dump_prefixes_1(Out,[H]) -> + io:format(Out," 0x~8.16.0BU~n",[H]), + io:format(Out,"};~n",[]); +dump_prefixes_1(Out,[H|T]) -> + io:format(Out," 0x~8.16.0BU,~n",[H]), + dump_prefixes_1(Out,T). + +%% make_big_prefixes([],Table) -> +%% Table; +%% make_big_prefixes([C|T],Table) -> +%% Index = (C div 32) + 1, +%% Pos = C rem 32, +%% X = element(Index,Table), +%% Y = X bor (1 bsl Pos), +%% NewTab = setelement(Index,Table,Y), +%% make_big_prefixes(T,NewTab). + +%% dump_big_prefixes(Out,L) -> +%% io:format(Out,"#define BIG_COMP_CANDIDATE_SIZE ~w~n", [?BIG_PREFIX_SIZE]), +%% io:format(Out,"static Uint32 big_comp_candidate_map[] = {~n",[]), +%% dump_prefixes_1(Out,L). + +pick([],_,Acc) -> + {lists:reverse(Acc),[]}; +pick([{[H|TT],N}|T],H,Acc) -> + pick(T,H,[{TT,N}|Acc]); +pick([{[H|_],_}|_]=L,M,Acc) when H =/= M -> + {lists:reverse(Acc),L}. + + +group([]) -> + []; +group([{[H],N}|T]) -> + {Part,Rest} = pick(T,H,[]), + [{H,group(Part),N}| group(Rest)]; +group([{[H|_],_}|_]=L) -> + {Part,Rest} = pick(L,H,[]), + [{H,group(Part),0}| group(Rest)]. + + + + + diff --git a/erts/emulator/sys/common/erl_sys_common_misc.c b/erts/emulator/sys/common/erl_sys_common_misc.c new file mode 100644 index 0000000000..461e763f03 --- /dev/null +++ b/erts/emulator/sys/common/erl_sys_common_misc.c @@ -0,0 +1,107 @@ +/* + * %CopyrightBegin% + * + * Copyright Ericsson AB 2006-2010. All Rights Reserved. + * + * The contents of this file are subject to the Erlang Public License, + * Version 1.1, (the "License"); you may not use this file except in + * compliance with the License. You should have received a copy of the + * Erlang Public License along with this software. If not, it can be + * retrieved online at http://www.erlang.org/. + * + * Software distributed under the License is distributed on an "AS IS" + * basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See + * the License for the specific language governing rights and limitations + * under the License. + * + * %CopyrightEnd% + */ + + + +/* + * Darwin needs conversion! + * http://developer.apple.com/library/mac/#qa/qa2001/qa1235.html + */ + +#ifdef HAVE_CONFIG_H +# include "config.h" +#endif + +#include "sys.h" +#include "global.h" + +#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__) +#define __DARWIN__ 1 +#endif + +#if !defined(__WIN32__) +#include <locale.h> +#if !defined(HAVE_SETLOCALE) || !defined(HAVE_NL_LANGINFO) || !defined(HAVE_LANGINFO_H) +#define PRIMITIVE_UTF8_CHECK 1 +#else +#include <langinfo.h> +#endif +#endif + +/* Written once and only once */ + +static int filename_encoding = ERL_FILENAME_UNKNOWN; +#if defined(__WIN32__) || defined(__DARWIN__) +static int user_filename_encoding = ERL_FILENAME_UTF8; /* Default unicode on windows */ +#else +static int user_filename_encoding = ERL_FILENAME_LATIN1; +#endif +void erts_set_user_requested_filename_encoding(int encoding) +{ + user_filename_encoding = encoding; +} + +int erts_get_user_requested_filename_encoding(void) +{ + return user_filename_encoding; +} + +void erts_init_sys_common_misc(void) +{ +#if defined(__WIN32__) + /* win_efile will totally fail if this is not set. */ + filename_encoding = ERL_FILENAME_WIN_WCHAR; +#else + if (user_filename_encoding != ERL_FILENAME_UNKNOWN) { + filename_encoding = user_filename_encoding; + } else { + char *l; + filename_encoding = ERL_FILENAME_LATIN1; +# ifdef PRIMITIVE_UTF8_CHECK + setlocale(LC_CTYPE, ""); /* Set international environment, + ignore result */ + if (((l = getenv("LC_ALL")) && *l) || + ((l = getenv("LC_CTYPE")) && *l) || + ((l = getenv("LANG")) && *l)) { + if (strstr(l, "UTF-8")) { + filename_encoding = ERL_FILENAME_UTF8; + } + } + +# else + l = setlocale(LC_CTYPE, ""); /* Set international environment */ + if (l != NULL) { + if (strcmp(nl_langinfo(CODESET), "UTF-8") == 0) { + filename_encoding = ERL_FILENAME_UTF8; + } + } +# endif + } +# if defined(__DARWIN__) + if (filename_encoding == ERL_FILENAME_UTF8) { + filename_encoding = ERL_FILENAME_UTF8_MAC; + } +# endif +#endif +} + +int erts_get_native_filename_encoding(void) +{ + return filename_encoding; +} diff --git a/erts/emulator/sys/win32/sys.c b/erts/emulator/sys/win32/sys.c index 39b04b26a9..37041ed987 100644 --- a/erts/emulator/sys/win32/sys.c +++ b/erts/emulator/sys/win32/sys.c @@ -67,14 +67,17 @@ static void async_read_file(struct async_io* aio, LPVOID buf, DWORD numToRead); static int async_write_file(struct async_io* aio, LPVOID buf, DWORD numToWrite); static int get_overlapped_result(struct async_io* aio, LPDWORD pBytesRead, BOOL wait); -static BOOL CreateChildProcess(char *, HANDLE, HANDLE, +static BOOL create_child_process(char *, HANDLE, HANDLE, HANDLE, LPHANDLE, BOOL, LPVOID, LPTSTR, unsigned, char **, int *); static int create_pipe(LPHANDLE, LPHANDLE, BOOL, BOOL); -static int ApplicationType(const char* originalName, char fullPath[MAX_PATH], +static int application_type(const char* originalName, char fullPath[MAX_PATH], BOOL search_in_path, BOOL handle_quotes, int *error_return); +static int application_type_w(const char* originalName, WCHAR fullPath[MAX_PATH], + BOOL search_in_path, BOOL handle_quotes, + int *error_return); HANDLE erts_service_event; @@ -87,7 +90,7 @@ static erts_smp_atomic_t pipe_creation_counter; static erts_smp_mtx_t sys_driver_data_lock; -/* Results from ApplicationType is one of */ +/* Results from application_type(_w) is one of */ #define APPL_NONE 0 #define APPL_DOS 1 #define APPL_WIN3X 2 @@ -1235,8 +1238,10 @@ spawn_start(ErlDrvPort port_num, char* name, SysDriverOpts* opts) */ DEBUGF(("Spawning \"%s\"\n", name)); - envir = win_build_environment(envir); - ok = CreateChildProcess(name, + envir = win_build_environment(envir); /* Still an ansi environment, could be + converted to unicode for spawn_executable, but + that is not done (yet) */ + ok = create_child_process(name, hChildStdin, hChildStdout, hChildStderr, @@ -1315,7 +1320,7 @@ create_file_thread(AsyncIo* aio, int mode) } /* - * A helper function used by CreateChildProcess(). + * A helper function used by create_child_process(). * Parses a command line with arguments and returns the length of the * first part containing the program name. * Example: input = "\"Program Files\"\\erl arg1 arg2" @@ -1356,24 +1361,25 @@ int parse_command(char* cmd){ return i; } -BOOL need_quotes(char *str) +static BOOL need_quotes(WCHAR *str) { int in_quote = 0; int backslashed = 0; int naked_space = 0; - while (*str != '\0') { + + while (*str != L'\0') { switch (*str) { - case '\\' : + case L'\\' : backslashed = !backslashed; break; - case '"': + case L'"': if (backslashed) { backslashed=0; } else { in_quote = !in_quote; } break; - case ' ': + case L' ': backslashed = 0; if (!(backslashed || in_quote)) { naked_space++; @@ -1392,7 +1398,7 @@ BOOL need_quotes(char *str) /* *---------------------------------------------------------------------- * - * CreateChildProcess -- + * create_child_process -- * * Create a child process that has pipes as its * standard input, output, and error. The child process runs @@ -1417,7 +1423,7 @@ BOOL need_quotes(char *str) */ static BOOL -CreateChildProcess +create_child_process ( char *origcmd, /* Command line for child process (including * name of executable). Or whole executable if st is @@ -1436,14 +1442,12 @@ CreateChildProcess ) { PROCESS_INFORMATION piProcInfo = {0}; - STARTUPINFO siStartInfo = {0}; BOOL ok = FALSE; int applType; /* Not to be changed for different types of executables */ int staticCreateFlags = GetPriorityClass(GetCurrentProcess()); int createFlags = DETACHED_PROCESS; char *newcmdline = NULL; - char execPath[MAX_PATH]; int cmdlength; char* thecommand; LPTSTR appname = NULL; @@ -1451,14 +1455,17 @@ CreateChildProcess *errno_return = -1; - siStartInfo.cb = sizeof(STARTUPINFO); - siStartInfo.dwFlags = STARTF_USESTDHANDLES; - siStartInfo.hStdInput = hStdin; - siStartInfo.hStdOutput = hStdout; - siStartInfo.hStdError = hStderr; - if (st != ERTS_SPAWN_EXECUTABLE) { + STARTUPINFO siStartInfo = {0}; + char execPath[MAX_PATH]; + + siStartInfo.cb = sizeof(STARTUPINFO); + siStartInfo.dwFlags = STARTF_USESTDHANDLES; + siStartInfo.hStdInput = hStdin; + siStartInfo.hStdOutput = hStdout; + siStartInfo.hStdError = hStderr; + /* * Parse out the program name from the command line (it can be quoted and * contain spaces). @@ -1470,9 +1477,9 @@ CreateChildProcess thecommand[cmdlength] = '\0'; DEBUGF(("spawn command: %s\n", thecommand)); - applType = ApplicationType(thecommand, execPath, TRUE, + applType = application_type(thecommand, execPath, TRUE, TRUE, errno_return); - DEBUGF(("ApplicationType returned for (%s) is %d\n", thecommand, applType)); + DEBUGF(("application_type returned for (%s) is %d\n", thecommand, applType)); erts_free(ERTS_ALC_T_TMP, (void *) thecommand); if (applType == APPL_NONE) { erts_free(ERTS_ALC_T_TMP,newcmdline); @@ -1501,126 +1508,147 @@ CreateChildProcess strcat(newcmdline, execPath); strcat(newcmdline, origcmd+cmdlength); - } else { /* ERTS_SPAWN_EXECUTABLE */ + DEBUGF(("Creating child process: %s, createFlags = %d\n", newcmdline, createFlags)); + ok = CreateProcessA(appname, + newcmdline, + NULL, + NULL, + TRUE, + createFlags | staticCreateFlags, + env, + wd, + &siStartInfo, + &piProcInfo); + + } else { /* ERTS_SPAWN_EXECUTABLE, filename and args are in unicode ({utf16,little}) */ int run_cmd = 0; - applType = ApplicationType(origcmd, execPath, FALSE, FALSE, - errno_return); + STARTUPINFOW siStartInfo = {0}; + WCHAR execPath[MAX_PATH]; + + + siStartInfo.cb = sizeof(STARTUPINFOW); + siStartInfo.dwFlags = STARTF_USESTDHANDLES; + siStartInfo.hStdInput = hStdin; + siStartInfo.hStdOutput = hStdout; + siStartInfo.hStdError = hStderr; + + applType = application_type_w(origcmd, (char *) execPath, FALSE, FALSE, + errno_return); if (applType == APPL_NONE) { return FALSE; } if (applType == APPL_DOS) { - /* - * See comment above - */ + /* + * See comment above + */ - siStartInfo.wShowWindow = SW_HIDE; - siStartInfo.dwFlags |= STARTF_USESHOWWINDOW; - createFlags = CREATE_NEW_CONSOLE; - run_cmd = 1; + siStartInfo.wShowWindow = SW_HIDE; + siStartInfo.dwFlags |= STARTF_USESHOWWINDOW; + createFlags = CREATE_NEW_CONSOLE; + run_cmd = 1; } else if (hide) { - DEBUGF(("hiding window\n")); - siStartInfo.wShowWindow = SW_HIDE; - siStartInfo.dwFlags |= STARTF_USESHOWWINDOW; - createFlags = 0; + DEBUGF(("hiding window\n")); + siStartInfo.wShowWindow = SW_HIDE; + siStartInfo.dwFlags |= STARTF_USESHOWWINDOW; + createFlags = 0; } if (run_cmd) { - char cmdPath[MAX_PATH]; + WCHAR cmdPath[MAX_PATH]; int cmdType; - cmdType = ApplicationType("cmd.exe", cmdPath, TRUE, FALSE, errno_return); + cmdType = application_type_w((char *) L"cmd.exe", (char *) cmdPath, TRUE, FALSE, errno_return); if (cmdType == APPL_NONE || cmdType == APPL_DOS) { return FALSE; } - appname = (char *) erts_alloc(ERTS_ALC_T_TMP, strlen(cmdPath)+1); - strcpy(appname,cmdPath); + appname = (char *) erts_alloc(ERTS_ALC_T_TMP, (wcslen(cmdPath)+1)*sizeof(WCHAR)); + wcscpy((WCHAR *) appname,cmdPath); } else { - appname = (char *) erts_alloc(ERTS_ALC_T_TMP, strlen(execPath)+1); - strcpy(appname,execPath); + appname = (char *) erts_alloc(ERTS_ALC_T_TMP, (wcslen(execPath)+1)*sizeof(WCHAR)); + wcscpy((WCHAR *) appname, execPath); } - if (argv == NULL) { + if (argv == NULL) { BOOL orig_need_q = need_quotes(execPath); - char *ptr; - int ocl = strlen(execPath); + WCHAR *ptr; + int ocl = wcslen(execPath); if (run_cmd) { newcmdline = (char *) erts_alloc(ERTS_ALC_T_TMP, - ocl + ((orig_need_q) ? 3 : 1) - + 11); - memcpy(newcmdline,"cmd.exe /c ",11); - ptr = newcmdline + 11; + (ocl + ((orig_need_q) ? 3 : 1) + + 11)*sizeof(WCHAR)); + memcpy(newcmdline,L"cmd.exe /c ",11*sizeof(WCHAR)); + ptr = (WCHAR *) (newcmdline + (11*sizeof(WCHAR))); } else { newcmdline = (char *) erts_alloc(ERTS_ALC_T_TMP, - ocl + ((orig_need_q) ? 3 : 1)); - ptr = newcmdline; + (ocl + ((orig_need_q) ? 3 : 1))*sizeof(WCHAR)); + ptr = (WCHAR *) newcmdline; } if (orig_need_q) { - *ptr++ = '"'; + *ptr++ = L'"'; } - memcpy(ptr,execPath,ocl); + memcpy(ptr,execPath,ocl*sizeof(WCHAR)); ptr += ocl; if (orig_need_q) { - *ptr++ = '"'; + *ptr++ = L'"'; } - *ptr = '\0'; + *ptr = L'\0'; } else { int sum = 1; /* '\0' */ - char **ar = argv; - char *n; + WCHAR **ar = (WCHAR **) argv; + WCHAR *n; char *save_arg0 = NULL; if (argv[0] == erts_default_arg0 || run_cmd) { save_arg0 = argv[0]; - argv[0] = execPath; + argv[0] = (char *) execPath; } if (run_cmd) { sum += 11; /* cmd.exe /c */ } while (*ar != NULL) { - sum += strlen(*ar); + sum += wcslen(*ar); if (need_quotes(*ar)) { sum += 2; /* quotes */ } sum++; /* space */ ++ar; } - ar = argv; - newcmdline = erts_alloc(ERTS_ALC_T_TMP, sum); - n = newcmdline; + ar = (WCHAR **) argv; + newcmdline = erts_alloc(ERTS_ALC_T_TMP, sum*sizeof(WCHAR)); + n = (WCHAR *) newcmdline; if (run_cmd) { - memcpy(n,"cmd.exe /c ",11); + memcpy(n,L"cmd.exe /c ",11*sizeof(WCHAR)); n += 11; } while (*ar != NULL) { int q = need_quotes(*ar); - sum = strlen(*ar); + sum = wcslen(*ar); if (q) { - *n++ = '"'; + *n++ = L'"'; } - memcpy(n,*ar,sum); + memcpy(n,*ar,sum*sizeof(WCHAR)); n += sum; if (q) { - *n++ = '"'; + *n++ = L'"'; } - *n++ = ' '; + *n++ = L' '; ++ar; } - ASSERT(n > newcmdline); - *(n-1) = '\0'; + *(n-1) = L'\0'; if (save_arg0 != NULL) { argv[0] = save_arg0; } } - } - DEBUGF(("Creating child process: %s, createFlags = %d\n", newcmdline, createFlags)); - ok = CreateProcess(appname, - newcmdline, - NULL, - NULL, - TRUE, - createFlags | staticCreateFlags, - env, - wd, - &siStartInfo, - &piProcInfo); - + DEBUGF(("Creating child process: %s, createFlags = %d\n", newcmdline, createFlags)); + ok = CreateProcessW((WCHAR *) appname, + (WCHAR *) newcmdline, + NULL, + NULL, + TRUE, + createFlags | staticCreateFlags, + env, + (WCHAR *) wd, + &siStartInfo, + &piProcInfo); + + } /* end SPAWN_EXECUTABLE */ if (newcmdline != NULL) { erts_free(ERTS_ALC_T_TMP,newcmdline); } @@ -1739,7 +1767,7 @@ static int create_pipe(HANDLE *phRead, HANDLE *phWrite, BOOL inheritRead, BOOL o -static int ApplicationType +static int application_type ( const char *originalName, /* Name of the application to find. */ char fullPath[MAX_PATH], /* Filled with complete path to @@ -1893,6 +1921,146 @@ static int ApplicationType return applType; } +static int application_type_w (const char *originalName, /* Name of the application to find. */ + WCHAR wfullpath[MAX_PATH],/* Filled with complete path to + * application. */ + BOOL search_in_path, /* If we should search the system wide path */ + BOOL handle_quotes, /* If we should handle quotes around executable */ + int *error_return) /* A place to put an error code */ +{ + int applType, i; + HANDLE hFile; + WCHAR *ext, *rest; + char buf[2]; + DWORD read; + IMAGE_DOS_HEADER header; + static WCHAR extensions[][5] = {L"", L".com", L".exe", L".bat"}; + int is_quoted; + int len; + WCHAR *wname = (WCHAR *) originalName; + WCHAR xfullpath[MAX_PATH]; + + len = wcslen(wname); + is_quoted = handle_quotes && len > 0 && wname[0] == L'"' && + wname[len-1] == L'"'; + + applType = APPL_NONE; + *error_return = ENOENT; + for (i = 0; i < (int) (sizeof(extensions) / sizeof(extensions[0])); i++) { + if(is_quoted) { + lstrcpynW(xfullpath, wname+1, MAX_PATH - 7); /* Cannot start using StringCchCopy yet, we support + older platforms */ + len = wcslen(xfullpath); + if(len > 0) { + xfullpath[len-1] = L'\0'; + } + } else { + lstrcpynW(xfullpath, wname, MAX_PATH - 5); + } + wcscat(xfullpath, extensions[i]); + /* It seems that the Unicode version does not allow in and out parameter to overlap. */ + SearchPathW((search_in_path) ? NULL : L".", xfullpath, NULL, MAX_PATH, wfullpath, &rest); + + /* + * Ignore matches on directories or data files, return if identified + * a known type. + */ + + if (GetFileAttributesW(wfullpath) & FILE_ATTRIBUTE_DIRECTORY) { + continue; + } + + ext = wcsrchr(wfullpath, L'.'); + if ((ext != NULL) && (_wcsicmp(ext, L".bat") == 0)) { + *error_return = EACCES; + applType = APPL_DOS; + break; + } + + hFile = CreateFileW(wfullpath, GENERIC_READ, FILE_SHARE_READ, NULL, + OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL); + if (hFile == INVALID_HANDLE_VALUE) { + continue; + } + + *error_return = EACCES; /* If considered an error, + it's an access error */ + header.e_magic = 0; + ReadFile(hFile, (void *) &header, sizeof(header), &read, NULL); + if (header.e_magic != IMAGE_DOS_SIGNATURE) { + /* + * Doesn't have the magic number for relocatable executables. If + * filename ends with .com, assume it's a DOS application anyhow. + * Note that we didn't make this assumption at first, because some + * supposed .com files are really 32-bit executables with all the + * magic numbers and everything. + */ + + CloseHandle(hFile); + if ((ext != NULL) && (_wcsicmp(ext, L".com") == 0)) { + applType = APPL_DOS; + break; + } + continue; + } + if (header.e_lfarlc != sizeof(header)) { + /* + * All Windows 3.X and Win32 and some DOS programs have this value + * set here. If it doesn't, assume that since it already had the + * other magic number it was a DOS application. + */ + + CloseHandle(hFile); + applType = APPL_DOS; + break; + } + + /* + * The DWORD at header.e_lfanew points to yet another magic number. + */ + + buf[0] = '\0'; + SetFilePointer(hFile, header.e_lfanew, NULL, FILE_BEGIN); + ReadFile(hFile, (void *) buf, 2, &read, NULL); + CloseHandle(hFile); + + if ((buf[0] == 'L') && (buf[1] == 'E')) { + applType = APPL_DOS; + } else if ((buf[0] == 'N') && (buf[1] == 'E')) { + applType = APPL_WIN3X; + } else if ((buf[0] == 'P') && (buf[1] == 'E')) { + applType = APPL_WIN32; + } else { + continue; + } + break; + } + + if (applType == APPL_NONE) { + return APPL_NONE; + } + + if ((applType == APPL_DOS) || (applType == APPL_WIN3X)) { + /* + * Replace long path name of executable with short path name for + * 16-bit applications. Otherwise the application may not be able + * to correctly parse its own command line to separate off the + * application name from the arguments. + */ + + GetShortPathNameW(wfullpath, wfullpath, MAX_PATH); + } + if (is_quoted) { + /* restore quotes on quoted program name */ + len = wcslen(wfullpath); + memmove(wfullpath+1,wfullpath,len*sizeof(WCHAR)); + wfullpath[0]=L'"'; + wfullpath[len+1]=L'"'; + wfullpath[len+2]=L'\0'; + } + return applType; +} + /* * Thread function used to emulate overlapped reading. */ diff --git a/erts/emulator/zlib/zutil.h b/erts/emulator/zlib/zutil.h index d560382691..a8872e1c88 100644 --- a/erts/emulator/zlib/zutil.h +++ b/erts/emulator/zlib/zutil.h @@ -142,6 +142,7 @@ extern const char * const z_errmsg[10]; /* indexed by 2-zlib_error */ #ifdef WIN32 # ifndef __CYGWIN__ /* Cygwin is Unix, not Win32 */ # define OS_CODE 0x0b +# define F_OPEN(name, mode) _wfopen((WCHAR *)(name), (WCHAR *)(mode)) /* Unicode */ # endif #endif diff --git a/erts/etc/win32/nsis/Makefile b/erts/etc/win32/nsis/Makefile index ebb3ad9a96..981a232c69 100644 --- a/erts/etc/win32/nsis/Makefile +++ b/erts/etc/win32/nsis/Makefile @@ -45,6 +45,7 @@ WTARGET_DIR=$(shell (cygpath -d $(TARGET_DIR) 2>/dev/null || cygpath -d $(TARGET REDIST_FILE=$(shell (sh ./find_redist.sh || echo "")) REDIST_DLL_VERSION=$(shell (sh ./dll_version_helper.sh || echo "")) +REDIST_DLL_NAME=$(shell (sh ./dll_version_helper.sh -n || echo "")) release_spec: @NSIS_VER=`makensis /hdrinfo | head -1 | awk '{print $$2}'`; \ @@ -73,6 +74,7 @@ release_spec: cp $(REDIST_FILE) $(RELEASE_PATH)/vcredist_x86.exe;\ echo '!define HAVE_REDIST_FILE 1' >> $(VERSION_HEADER); \ echo '!define REDIST_DLL_VERSION "$(REDIST_DLL_VERSION)"' >> $(VERSION_HEADER);\ + echo '!define REDIST_DLL_NAME "$(REDIST_DLL_NAME)"' >> $(VERSION_HEADER);\ fi;\ if [ -f $(RELEASE_PATH)/docs/doc/index.html ];\ then \ diff --git a/erts/etc/win32/nsis/dll_version_helper.sh b/erts/etc/win32/nsis/dll_version_helper.sh index e0047dea8b..571ee3e39e 100755 --- a/erts/etc/win32/nsis/dll_version_helper.sh +++ b/erts/etc/win32/nsis/dll_version_helper.sh @@ -2,7 +2,7 @@ # # %CopyrightBegin% # -# Copyright Ericsson AB 2007-2009. All Rights Reserved. +# Copyright Ericsson AB 2007-2010. All Rights Reserved. # # The contents of this file are subject to the Erlang Public License, # Version 1.1, (the "License"); you may not use this file except in @@ -41,9 +41,15 @@ if [ '!' -f hello.exe.manifest ]; then exit 0 fi VERSION=`grep '<assemblyIdentity' hello.exe.manifest | sed 's,.*version=.\([0-9\.]*\).*,\1,g' | grep -v '<'` +NAME=`grep '<assemblyIdentity' hello.exe.manifest | sed 's,.*name=.[A-Za-z\.]*\([0-9]*\).*,msvcr\1.dll,g' | grep -v '<'` rm -f hello.c hello.obj hello.exe hello.exe.manifest -if [ -z "$VERSION" ]; then +if [ "$1" = "-n" ]; then + ASKEDFOR=$NAME +else + ASKEDFOR=$VERSION +fi +if [ -z "$ASKEDFOR" ]; then exit 1 fi -echo $VERSION +echo $ASKEDFOR exit 0 diff --git a/erts/etc/win32/nsis/erlang20.nsi b/erts/etc/win32/nsis/erlang20.nsi index 43e5d91604..941e8e6f5d 100644 --- a/erts/etc/win32/nsis/erlang20.nsi +++ b/erts/etc/win32/nsis/erlang20.nsi @@ -311,23 +311,23 @@ FunctionEnd Function .onInit
SectionGetFlags 0 $MYTEMP
-; MessageBox MB_YESNO "Found $SYSDIR\msvcr80.dll" IDYES FoundLbl
- IfFileExists $SYSDIR\msvcr80.dll MaybeFoundInSystemLbl
+ ;MessageBox MB_YESNO "Found $SYSDIR\${REDIST_DLL_NAME}" IDYES FoundLbl
+ IfFileExists $SYSDIR\${REDIST_DLL_NAME} MaybeFoundInSystemLbl
SearchSxsLbl:
FindFirst $0 $1 $WINDIR\WinSxS\x86*
LoopLbl:
StrCmp $1 "" NotFoundLbl
- IfFileExists $WINDIR\WinSxS\$1\msvcr80.dll MaybeFoundInSxsLbl
+ IfFileExists $WINDIR\WinSxS\$1\${REDIST_DLL_NAME} MaybeFoundInSxsLbl
FindNext $0 $1
Goto LoopLbl
MaybeFoundInSxsLbl:
- GetDllVersion $WINDIR\WinSxS\$1\msvcr80.dll $R0 $R1
+ GetDllVersion $WINDIR\WinSxS\$1\${REDIST_DLL_NAME} $R0 $R1
Call DllVersionGoodEnough
FindNext $0 $1
IntCmp 2 $R0 LoopLbl
Goto FoundLbl
MaybeFoundInSystemLbl:
- GetDllVersion $SYSDIR\msvcr80.dll $R0 $R1
+ GetDllVersion $SYSDIR\${REDIST_DLL_NAME} $R0 $R1
Call DllVersionGoodEnough
IntCmp 2 $R0 SearchSxSLbl
FoundLbl:
diff --git a/erts/etc/win32/nsis/find_redist.sh b/erts/etc/win32/nsis/find_redist.sh index c5572839c5..153977ded5 100755 --- a/erts/etc/win32/nsis/find_redist.sh +++ b/erts/etc/win32/nsis/find_redist.sh @@ -2,7 +2,7 @@ # # %CopyrightBegin% # -# Copyright Ericsson AB 2007-2009. All Rights Reserved. +# Copyright Ericsson AB 2007-2010. All Rights Reserved. # # The contents of this file are subject to the Erlang Public License, # Version 1.1, (the "License"); you may not use this file except in @@ -107,16 +107,56 @@ for x in cl bin vc; do fi BPATH="$NBPATH" done +BPATH_LIST=$BPATH + +# rc.exe is in the Microsoft SDK directory of VS2008 +RCPATH=`lookup_prog_in_path rc` +fail=false +if [ '!' -z "$RCPATH" ]; then + BPATH=$RCPATH + for x in rc bin v6.0A ; do + NBPATH=`remove_path_element $x "$BPATH"` + if [ "$NBPATH" = "$BPATH" ]; then + fail=true + break; + fi + BPATH="$NBPATH" + done + if [ $fail = false ]; then + BPATH_LIST="$BPATH_LIST $BPATH" + fi +fi + +# Frantic search through two roots with different +# version directories. We want to be very specific about the +# directory structures as we woildnt want to find the wrong +# redistributables... + #echo $BPATH -for x in sdk v2.0 bootstrapper packages vcredist_x86 vcredist_x86.exe; do - #echo "x=$x" - #echo "BPATH=$BPATH" - NBPATH=`add_path_element $x "$BPATH"` - if [ "$NBPATH" = "$BPATH" ]; then - echo "Failed to locate vcredist_x86.exe because directory structure was unexpected" >&2 - exit 3 +for BP in $BPATH_LIST; do + for verdir in "sdk v2.0" "sdk v3.5" "v6.0A"; do + BPATH=$BP + fail=false + for x in $verdir bootstrapper packages vcredist_x86 vcredist_x86.exe; do + #echo "x=$x" + #echo "BPATH=$BPATH" + NBPATH=`add_path_element $x "$BPATH"` + if [ "$NBPATH" = "$BPATH" ]; then + fail=true + break; + fi + BPATH="$NBPATH" + done + if [ $fail = false ]; then + break; + fi + done + if [ $fail = false ]; then + echo $BPATH + exit 0 fi - BPATH="$NBPATH" done -echo $BPATH -exit 0
\ No newline at end of file + +echo "Failed to locate vcredist_x86.exe because directory structure was unexpected" >&2 +exit 3 + diff --git a/erts/preloaded/ebin/prim_file.beam b/erts/preloaded/ebin/prim_file.beam Binary files differindex a3f300268f..8a1c22e5b5 100644 --- a/erts/preloaded/ebin/prim_file.beam +++ b/erts/preloaded/ebin/prim_file.beam diff --git a/erts/preloaded/src/prim_file.erl b/erts/preloaded/src/prim_file.erl index 7f24889bb2..10be852e92 100644 --- a/erts/preloaded/src/prim_file.erl +++ b/erts/preloaded/src/prim_file.erl @@ -109,6 +109,8 @@ -define(FILE_RESP_LDATA, 6). -define(FILE_RESP_N2DATA, 7). -define(FILE_RESP_EOF, 8). +-define(FILE_RESP_FNAME, 9). +-define(FILE_RESP_ALL_DATA, 10). %% Open modes for the driver's open function. -define(EFILE_MODE_READ, 1). @@ -153,7 +155,7 @@ %% Opens a file using the driver port Port. Returns {error, Reason} %% | {ok, FileDescriptor} open(Port, File, ModeList) when is_port(Port), - is_list(File), + (is_list(File) orelse is_binary(File)), is_list(ModeList) -> case open_mode(ModeList) of {Mode, _Portopts, _Setopts} -> @@ -165,10 +167,11 @@ open(_,_,_) -> {error, badarg}. %% Opens a file. Returns {error, Reason} | {ok, FileDescriptor}. -open(File, ModeList) when is_list(File), is_list(ModeList) -> +open(File, ModeList) when (is_list(File) orelse is_binary(File)), + is_list(ModeList) -> case open_mode(ModeList) of {Mode, Portopts, Setopts} -> - open_int({?FD_DRV, Portopts}, File, Mode, Setopts); + open_int({?FD_DRV, Portopts},File, Mode, Setopts); Reason -> {error, Reason} end; @@ -196,7 +199,7 @@ open_int({Driver, Portopts}, File, Mode, Setopts) -> end; open_int(Port, File, Mode, Setopts) -> M = Mode band ?EFILE_MODE_MASK, - case drv_command(Port, [<<?FILE_OPEN, M:32>>, File, 0]) of + case drv_command(Port, [<<?FILE_OPEN, M:32>>, pathname(File)]) of {ok, Number} -> open_int_setopts(Port, Number, Setopts); Error -> @@ -489,7 +492,7 @@ ipread_s32bu_p32bu(#file_descriptor{module = ?MODULE, data = {_, _}}, %% Returns {ok, Contents} | {error, Reason} -read_file(File) -> +read_file(File) when (is_list(File) orelse is_binary(File)) -> case drv_open(?FD_DRV, [binary]) of {ok, Port} -> Result = read_file(Port, File), @@ -497,11 +500,14 @@ read_file(File) -> Result; {error, _} = Error -> Error - end. + end; +read_file(_) -> + {error, badarg}. %% Takes a Port opened with open/1. -read_file(Port, File) when is_port(Port) -> - Cmd = [?FILE_READ_FILE | File], +read_file(Port, File) when is_port(Port), + (is_list(File) orelse is_binary(File))-> + Cmd = [?FILE_READ_FILE | pathname(File)], case drv_command(Port, Cmd) of {error, enomem} -> %% It could possibly help to do a @@ -512,12 +518,14 @@ read_file(Port, File) when is_port(Port) -> drv_command(Port, Cmd); Result -> Result - end. + end; +read_file(_,_) -> + {error, badarg}. %% Returns {error, Reason} | ok. -write_file(File, Bin) -> +write_file(File, Bin) when (is_list(File) orelse is_binary(File)) -> case open(File, [binary, write]) of {ok, Handle} -> Result = write(Handle, Bin), @@ -525,8 +533,10 @@ write_file(File, Bin) -> Result; Error -> Error - end. - + end; +write_file(_, _) -> + {error, badarg}. + %%%----------------------------------------------------------------- @@ -539,7 +549,7 @@ write_file(File, Bin) -> %% Returns {ok, Port}, the Port should be used as first argument in all %% the following functions. Returns {error, Reason} upon failure. start() -> - try erlang:open_port({spawn, atom_to_list(?DRV)}, []) of + try erlang:open_port({spawn, atom_to_list(?DRV)}, [binary]) of Port -> {ok, Port} catch @@ -596,7 +606,7 @@ get_cwd(_, _) -> {error, badarg}. get_cwd_int(Drive) -> - get_cwd_int({?DRV, []}, Drive). + get_cwd_int({?DRV, [binary]}, Drive). get_cwd_int(Port, Drive) -> drv_command(Port, <<?FILE_PWD, Drive>>). @@ -606,7 +616,7 @@ get_cwd_int(Port, Drive) -> %% set_cwd/{1,2} set_cwd(Dir) -> - set_cwd_int({?DRV, []}, Dir). + set_cwd_int({?DRV, [binary]}, Dir). set_cwd(Port, Dir) when is_port(Port) -> set_cwd_int(Port, Dir). @@ -632,89 +642,88 @@ set_cwd_int(Port, Dir0) -> end), %% Dir is now either a string or an EXIT tuple. %% An EXIT tuple will fail in the following catch. - drv_command(Port, [?FILE_CHDIR, Dir, 0]). + drv_command(Port, [?FILE_CHDIR, pathname(Dir)]). %% delete/{1,2} delete(File) -> - delete_int({?DRV, []}, File). + delete_int({?DRV, [binary]}, File). delete(Port, File) when is_port(Port) -> delete_int(Port, File). delete_int(Port, File) -> - drv_command(Port, [?FILE_DELETE, File, 0]). + drv_command(Port, [?FILE_DELETE, pathname(File)]). %% rename/{2,3} rename(From, To) -> - rename_int({?DRV, []}, From, To). + rename_int({?DRV, [binary]}, From, To). rename(Port, From, To) when is_port(Port) -> rename_int(Port, From, To). rename_int(Port, From, To) -> - drv_command(Port, [?FILE_RENAME, From, 0, To, 0]). + drv_command(Port, [?FILE_RENAME, pathname(From), pathname(To)]). %% make_dir/{1,2} make_dir(Dir) -> - make_dir_int({?DRV, []}, Dir). + make_dir_int({?DRV, [binary]}, Dir). make_dir(Port, Dir) when is_port(Port) -> make_dir_int(Port, Dir). make_dir_int(Port, Dir) -> - drv_command(Port, [?FILE_MKDIR, Dir, 0]). + drv_command(Port, [?FILE_MKDIR, pathname(Dir)]). %% del_dir/{1,2} del_dir(Dir) -> - del_dir_int({?DRV, []}, Dir). + del_dir_int({?DRV, [binary]}, Dir). del_dir(Port, Dir) when is_port(Port) -> del_dir_int(Port, Dir). del_dir_int(Port, Dir) -> - drv_command(Port, [?FILE_RMDIR, Dir, 0]). + drv_command(Port, [?FILE_RMDIR, pathname(Dir)]). %% read_file_info/{1,2} read_file_info(File) -> - read_file_info_int({?DRV, []}, File). + read_file_info_int({?DRV, [binary]}, File). read_file_info(Port, File) when is_port(Port) -> read_file_info_int(Port, File). read_file_info_int(Port, File) -> - drv_command(Port, [?FILE_FSTAT, File, 0]). + drv_command(Port, [?FILE_FSTAT, pathname(File)]). %% altname/{1,2} altname(File) -> - altname_int({?DRV, []}, File). + altname_int({?DRV, [binary]}, File). altname(Port, File) when is_port(Port) -> altname_int(Port, File). altname_int(Port, File) -> - drv_command(Port, [?FILE_ALTNAME, File, 0]). - + drv_command(Port, [?FILE_ALTNAME, pathname(File)]). %% write_file_info/{2,3} write_file_info(File, Info) -> - write_file_info_int({?DRV, []}, File, Info). + write_file_info_int({?DRV, [binary]}, File, Info). write_file_info(Port, File, Info) when is_port(Port) -> write_file_info_int(Port, File, Info). @@ -740,72 +749,72 @@ write_file_info_int(Port, date_to_bytes(Atime), date_to_bytes(Mtime), date_to_bytes(Ctime), - File, 0]). + pathname(File)]). %% make_link/{2,3} make_link(Old, New) -> - make_link_int({?DRV, []}, Old, New). + make_link_int({?DRV, [binary]}, Old, New). make_link(Port, Old, New) when is_port(Port) -> make_link_int(Port, Old, New). make_link_int(Port, Old, New) -> - drv_command(Port, [?FILE_LINK, Old, 0, New, 0]). + drv_command(Port, [?FILE_LINK, pathname(Old), pathname(New)]). %% make_symlink/{2,3} make_symlink(Old, New) -> - make_symlink_int({?DRV, []}, Old, New). + make_symlink_int({?DRV, [binary]}, Old, New). make_symlink(Port, Old, New) when is_port(Port) -> make_symlink_int(Port, Old, New). make_symlink_int(Port, Old, New) -> - drv_command(Port, [?FILE_SYMLINK, Old, 0, New, 0]). + drv_command(Port, [?FILE_SYMLINK, pathname(Old), pathname(New)]). %% read_link/{2,3} read_link(Link) -> - read_link_int({?DRV, []}, Link). + read_link_int({?DRV, [binary]}, Link). read_link(Port, Link) when is_port(Port) -> read_link_int(Port, Link). read_link_int(Port, Link) -> - drv_command(Port, [?FILE_READLINK, Link, 0]). + drv_command(Port, [?FILE_READLINK, pathname(Link)]). %% read_link_info/{2,3} read_link_info(Link) -> - read_link_info_int({?DRV, []}, Link). + read_link_info_int({?DRV, [binary]}, Link). read_link_info(Port, Link) when is_port(Port) -> read_link_info_int(Port, Link). read_link_info_int(Port, Link) -> - drv_command(Port, [?FILE_LSTAT, Link, 0]). + drv_command(Port, [?FILE_LSTAT, pathname(Link)]). %% list_dir/{1,2} list_dir(Dir) -> - list_dir_int({?DRV, []}, Dir). + list_dir_int({?DRV, [binary]}, Dir). list_dir(Port, Dir) when is_port(Port) -> list_dir_int(Port, Dir). list_dir_int(Port, Dir) -> - drv_command(Port, [?FILE_READDIR, Dir, 0], []). + drv_command(Port, [?FILE_READDIR, pathname(Dir)], []). @@ -1026,8 +1035,6 @@ lseek_position(_) -> translate_response(?FILE_RESP_OK, []) -> ok; -translate_response(?FILE_RESP_OK, Data) -> - {ok, Data}; translate_response(?FILE_RESP_ERROR, List) when is_list(List) -> {error, list_to_atom(List)}; translate_response(?FILE_RESP_NUMBER, List) -> @@ -1074,6 +1081,16 @@ translate_response(?FILE_RESP_N2DATA = X, L0) when is_list(L0) -> end; translate_response(?FILE_RESP_EOF, []) -> eof; +translate_response(?FILE_RESP_FNAME, []) -> + ok; +translate_response(?FILE_RESP_FNAME, Data) when is_binary(Data) -> + {ok, prim_file:internal_native2name(Data)}; +translate_response(?FILE_RESP_FNAME, Data) -> + {ok, Data}; + +translate_response(?FILE_RESP_ALL_DATA, Data) -> + {ok, Data}; + translate_response(X, Data) -> {error, {bad_response_from_port, [X | Data]}}. @@ -1209,3 +1226,9 @@ lists_split([Hd | Tl], N, Rev) -> reverse(X) -> lists:reverse(X, []). reverse(L, T) -> lists:reverse(L, T). + +% Will add zero termination too +% The 'EXIT' tuple from a bad argument will eventually generate an error +% in list_to_binary, which is caught and generates the {error,badarg} return +pathname(File) -> + (catch prim_file:internal_name2native(File)). diff --git a/erts/test/z_SUITE.erl b/erts/test/z_SUITE.erl index 8faddeb0d3..9f13a7083d 100644 --- a/erts/test/z_SUITE.erl +++ b/erts/test/z_SUITE.erl @@ -253,6 +253,8 @@ core_file_search(#core_search_conf{search_dir = Base, core_cand(Conf, Core, Cores); "core." ++ _ -> core_cand(Conf, Core, Cores); + Bin when is_binary(Bin) -> %Icky filename; ignore + Cores; BName -> case lists:suffix(".core", BName) of true -> core_cand(Conf, Core, Cores); |