diff options
Diffstat (limited to 'erts/emulator/beam')
| -rw-r--r-- | erts/emulator/beam/bif.tab | 7 | ||||
| -rw-r--r-- | erts/emulator/beam/binary.c | 10 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_bif_binary.c | 2 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_bif_port.c | 91 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_binary.h | 2 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_init.c | 23 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_unicode.c | 902 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_unicode_normalize.h | 1687 | ||||
| -rw-r--r-- | erts/emulator/beam/global.h | 13 | ||||
| -rw-r--r-- | erts/emulator/beam/sys.h | 16 |
10 files changed, 2652 insertions, 101 deletions
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 |