diff options
Diffstat (limited to 'erts')
| -rw-r--r-- | erts/doc/src/erl_ext_dist.xml | 15 | ||||
| -rw-r--r-- | erts/doc/src/erlang.xml | 35 | ||||
| -rw-r--r-- | erts/emulator/beam/atom.c | 35 | ||||
| -rw-r--r-- | erts/emulator/beam/atom.h | 3 | ||||
| -rw-r--r-- | erts/emulator/beam/beam_load.c | 66 | ||||
| -rw-r--r-- | erts/emulator/beam/bif.c | 14 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_unicode.c | 83 | ||||
| -rw-r--r-- | erts/emulator/beam/global.h | 1 | ||||
| -rw-r--r-- | erts/emulator/beam/utils.c | 62 | ||||
| -rw-r--r-- | erts/emulator/test/bif_SUITE.erl | 47 | ||||
| -rw-r--r-- | erts/emulator/test/code_SUITE.erl | 10 |
11 files changed, 242 insertions, 129 deletions
diff --git a/erts/doc/src/erl_ext_dist.xml b/erts/doc/src/erl_ext_dist.xml index 4f799f8f34..a436a9ca74 100644 --- a/erts/doc/src/erl_ext_dist.xml +++ b/erts/doc/src/erl_ext_dist.xml @@ -119,16 +119,11 @@ <tcaption>Compressed Data Format when Expanded</tcaption></table> <marker id="utf8_atoms"/> <note> - <p>As from ERTS 5.10 (OTP R16) support - for UTF-8 encoded atoms has been introduced in the external format. - However, only characters that can be encoded using Latin-1 (ISO-8859-1) - are currently supported in atoms. The support for UTF-8 encoded atoms - in the external format has been implemented to be able to support - all Unicode characters in atoms in <em>some future release</em>. - Until full Unicode support for atoms has been introduced, - it is an <em>error</em> to pass atoms containing - characters that cannot be encoded in Latin-1, and <em>the behavior is - undefined</em>.</p> + <p>As from ERTS 9.0 (OTP 20), UTF-8 encoded atoms may contain any Unicode + character. Although the support for UTF-8 encoded atoms in the external + format is available since ERTS 5.10 (OTP R16), passing atoms that cannot + be encoded in Latin-1 is an <em>error</em> in versions earlier than + Erlang/OTP 20, and <em>the behavior is undefined</em>.</p> <p>When distribution flag <seealso marker="erl_dist_protocol#dflags"> <c>DFLAG_UTF8_ATOMS</c></seealso> has been exchanged between both nodes in the <seealso marker="erl_dist_protocol#distribution_handshake"> diff --git a/erts/doc/src/erlang.xml b/erts/doc/src/erlang.xml index 2859f22bf4..a3401fe469 100644 --- a/erts/doc/src/erlang.xml +++ b/erts/doc/src/erlang.xml @@ -325,16 +325,11 @@ Z = erlang:adler32_combine(X,Y,iolist_size(Data2)).</code> is <c>latin1</c>, one byte exists for each character in the text representation. If <c><anno>Encoding</anno></c> is <c>utf8</c> or - <c>unicode</c>, the characters are encoded using UTF-8 - (that is, characters from 128 through 255 are - encoded in two bytes).</p> + <c>unicode</c>, the characters are encoded using UTF-8 where + characters may require multiple bytes.</p> <note> - <p><c>atom_to_binary(<anno>Atom</anno>, latin1)</c> never - fails, as the text representation of an atom can only - contain characters from 0 through 255. In a future release, - the text representation - of atoms can be allowed to contain any Unicode character and - <c>atom_to_binary(<anno>Atom</anno>, latin1)</c> then fails if the + <p>As from Erlang/OTP 20, atoms can contain any Unicode character + and <c>atom_to_binary(<anno>Atom</anno>, latin1)</c> may fail if the text representation for <c><anno>Atom</anno></c> contains a Unicode character > 255.</p> </note> @@ -402,13 +397,11 @@ Z = erlang:adler32_combine(X,Y,iolist_size(Data2)).</code> translation of bytes in the binary is done. If <c><anno>Encoding</anno></c> is <c>utf8</c> or <c>unicode</c>, the binary must contain - valid UTF-8 sequences. Only Unicode characters up - to 255 are allowed.</p> + valid UTF-8 sequences.</p> <note> - <p><c>binary_to_atom(<anno>Binary</anno>, utf8)</c> fails if - the binary contains Unicode characters > 255. - In a future release, such Unicode characters can be allowed and - <c>binary_to_atom(<anno>Binary</anno>, utf8)</c> does then not fail. + <p>As from Erlang/OTP 20, <c>binary_to_atom(<anno>Binary</anno>, utf8)</c> + is capable of encoding any Unicode character. Earlier versions would + fail if the binary contained Unicode characters > 255. For more information about Unicode support in atoms, see the <seealso marker="erl_ext_dist#utf8_atoms">note on UTF-8 encoded atoms</seealso> @@ -419,9 +412,7 @@ Z = erlang:adler32_combine(X,Y,iolist_size(Data2)).</code> > <input>binary_to_atom(<<"Erlang">>, latin1).</input> 'Erlang' > <input>binary_to_atom(<<1024/utf8>>, utf8).</input> -** exception error: bad argument - in function binary_to_atom/2 - called as binary_to_atom(<<208,128>>,utf8)</pre> +'Ѐ'</pre> </desc> </func> @@ -2381,10 +2372,10 @@ os_prompt%</pre> <desc> <p>Returns the atom whose text representation is <c><anno>String</anno></c>.</p> - <p><c><anno>String</anno></c> can only contain ISO-latin-1 - characters (that is, numbers < 256) as the implementation does not - allow Unicode characters equal to or above 256 in atoms. - For more information on Unicode support in atoms, see + <p>As from Erlang/OTP 20, <c><anno>String</anno></c> may contain + any Unicode character. Earlier versions allowed only ISO-latin-1 + characters as the implementation did not allow Unicode characters + above 255. For more information on Unicode support in atoms, see <seealso marker="erl_ext_dist#utf8_atoms">note on UTF-8 encoded atoms</seealso> in section "External Term Format" in the User's Guide.</p> diff --git a/erts/emulator/beam/atom.c b/erts/emulator/beam/atom.c index 2b5ad097a0..2055c29190 100644 --- a/erts/emulator/beam/atom.c +++ b/erts/emulator/beam/atom.c @@ -233,10 +233,10 @@ need_convertion: } /* - * erts_atom_put() may fail. If it fails THE_NON_VALUE is returned! + * erts_atom_put_index() may fail. Returns negative indexes for errors. */ -Eterm -erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) +int +erts_atom_put_index(const byte *name, int len, ErtsAtomEncoding enc, int trunc) { byte utf8_copy[MAX_ATOM_SZ_FROM_LATIN1]; const byte *text = name; @@ -253,7 +253,7 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) if (trunc) tlen = 0; else - return THE_NON_VALUE; + return ATOM_MAX_CHARS_ERROR; } switch (enc) { @@ -262,7 +262,7 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) if (trunc) tlen = MAX_ATOM_CHARACTERS; else - return THE_NON_VALUE; + return ATOM_MAX_CHARS_ERROR; } #ifdef DEBUG for (aix = 0; aix < len; aix++) { @@ -276,7 +276,7 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) if (trunc) tlen = MAX_ATOM_CHARACTERS; else - return THE_NON_VALUE; + return ATOM_MAX_CHARS_ERROR; } no_latin1_chars = tlen; latin1_to_utf8(utf8_copy, &text, &tlen); @@ -284,7 +284,7 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) case ERTS_ATOM_ENC_UTF8: /* First sanity check; need to verify later */ if (tlen > MAX_ATOM_SZ_LIMIT && !trunc) - return THE_NON_VALUE; + return ATOM_MAX_CHARS_ERROR; break; } @@ -295,7 +295,7 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) atom_read_unlock(); if (aix >= 0) { /* Already in table no need to verify it */ - return make_atom(aix); + return aix; } if (enc == ERTS_ATOM_ENC_UTF8) { @@ -314,13 +314,13 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) case ERTS_UTF8_OK_MAX_CHARS: /* Truncated... */ if (!trunc) - return THE_NON_VALUE; + return ATOM_MAX_CHARS_ERROR; ASSERT(no_chars == MAX_ATOM_CHARACTERS); tlen = err_pos - text; break; default: /* Bad utf8... */ - return THE_NON_VALUE; + return ATOM_BAD_ENCODING_ERROR; } } @@ -333,7 +333,20 @@ erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) atom_write_lock(); aix = index_put(&erts_atom_table, (void*) &a); atom_write_unlock(); - return make_atom(aix); + return aix; +} + +/* + * erts_atom_put() may fail. If it fails THE_NON_VALUE is returned! + */ +Eterm +erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc) +{ + int aix = erts_atom_put_index(name, len, enc, trunc); + if (aix >= 0) + return make_atom(aix); + else + return THE_NON_VALUE; } Eterm diff --git a/erts/emulator/beam/atom.h b/erts/emulator/beam/atom.h index abd3b44993..be998a46bd 100644 --- a/erts/emulator/beam/atom.h +++ b/erts/emulator/beam/atom.h @@ -29,6 +29,8 @@ #define MAX_ATOM_SZ_LIMIT (4*MAX_ATOM_CHARACTERS) /* theoretical byte limit */ #define ATOM_LIMIT (1024*1024) #define MIN_ATOM_TABLE_SIZE 8192 +#define ATOM_BAD_ENCODING_ERROR -1 +#define ATOM_MAX_CHARS_ERROR -2 #ifndef ARCH_32 /* Internal atom cache needs MAX_ATOM_TABLE_SIZE to be less than an @@ -133,6 +135,7 @@ int atom_table_sz(void); /* table size in bytes, excluding stored objects */ Eterm am_atom_put(const char*, int); /* ONLY 7-bit ascii! */ Eterm erts_atom_put(const byte *name, int len, ErtsAtomEncoding enc, int trunc); +int erts_atom_put_index(const byte *name, int len, ErtsAtomEncoding enc, int trunc); void init_atom_table(void); void atom_info(fmtfn_t, void *); void dump_atoms(fmtfn_t, void *); diff --git a/erts/emulator/beam/beam_load.c b/erts/emulator/beam/beam_load.c index 309465bcd3..d298a4ff26 100644 --- a/erts/emulator/beam/beam_load.c +++ b/erts/emulator/beam/beam_load.c @@ -157,13 +157,15 @@ typedef struct { #define STR_CHUNK 2 #define IMP_CHUNK 3 #define EXP_CHUNK 4 -#define NUM_MANDATORY 5 +#define MIN_MANDATORY 1 +#define MAX_MANDATORY 5 #define LAMBDA_CHUNK 5 #define LITERAL_CHUNK 6 #define ATTR_CHUNK 7 #define COMPILE_CHUNK 8 #define LINE_CHUNK 9 +#define UTF8_ATOM_CHUNK 10 #define NUM_CHUNK_TYPES (sizeof(chunk_types)/sizeof(chunk_types[0])) @@ -173,9 +175,13 @@ typedef struct { static Uint chunk_types[] = { /* - * Mandatory chunk types -- these MUST be present. + * Atom chunk types -- Atom or AtU8 MUST be present. */ MakeIffId('A', 't', 'o', 'm'), /* 0 */ + + /* + * Mandatory chunk types -- these MUST be present. + */ MakeIffId('C', 'o', 'd', 'e'), /* 1 */ MakeIffId('S', 't', 'r', 'T'), /* 2 */ MakeIffId('I', 'm', 'p', 'T'), /* 3 */ @@ -189,6 +195,7 @@ static Uint chunk_types[] = { MakeIffId('A', 't', 't', 'r'), /* 7 */ MakeIffId('C', 'I', 'n', 'f'), /* 8 */ MakeIffId('L', 'i', 'n', 'e'), /* 9 */ + MakeIffId('A', 't', 'U', '8'), /* 10 */ }; /* @@ -490,9 +497,9 @@ static Eterm stub_insert_new_code(Process *c_p, ErtsProcLocks c_p_locks, #endif static int init_iff_file(LoaderState* stp, byte* code, Uint size); static int scan_iff_file(LoaderState* stp, Uint* chunk_types, - Uint num_types, Uint num_mandatory); + Uint num_types); static int verify_chunks(LoaderState* stp); -static int load_atom_table(LoaderState* stp); +static int load_atom_table(LoaderState* stp, ErtsAtomEncoding enc); static int load_import_table(LoaderState* stp); static int read_export_table(LoaderState* stp); static int is_bif(Eterm mod, Eterm func, unsigned arity); @@ -629,7 +636,7 @@ erts_prepare_loading(Binary* magic, Process *c_p, Eterm group_leader, CHKALLOC(); CHKBLK(ERTS_ALC_T_CODE,stp->code); if (!init_iff_file(stp, code, unloaded_size) || - !scan_iff_file(stp, chunk_types, NUM_CHUNK_TYPES, NUM_MANDATORY) || + !scan_iff_file(stp, chunk_types, NUM_CHUNK_TYPES) || !verify_chunks(stp)) { goto load_error; } @@ -674,9 +681,16 @@ erts_prepare_loading(Binary* magic, Process *c_p, Eterm group_leader, */ CHKBLK(ERTS_ALC_T_CODE,stp->code); - define_file(stp, "atom table", ATOM_CHUNK); - if (!load_atom_table(stp)) { - goto load_error; + if (stp->chunks[UTF8_ATOM_CHUNK].size > 0) { + define_file(stp, "utf8 atom table", UTF8_ATOM_CHUNK); + if (!load_atom_table(stp, ERTS_ATOM_ENC_UTF8)) { + goto load_error; + } + } else { + define_file(stp, "atom table", ATOM_CHUNK); + if (!load_atom_table(stp, ERTS_ATOM_ENC_LATIN1)) { + goto load_error; + } } /* @@ -1212,7 +1226,7 @@ init_iff_file(LoaderState* stp, byte* code, Uint size) * Scan the IFF file. The header should have been verified by init_iff_file(). */ static int -scan_iff_file(LoaderState* stp, Uint* chunk_types, Uint num_types, Uint num_mandatory) +scan_iff_file(LoaderState* stp, Uint* chunk_types, Uint num_types) { Uint count; Uint id; @@ -1291,7 +1305,16 @@ verify_chunks(LoaderState* stp) MD5_CTX context; MD5Init(&context); - for (i = 0; i < NUM_MANDATORY; i++) { + + if (stp->chunks[UTF8_ATOM_CHUNK].start != NULL) { + MD5Update(&context, stp->chunks[UTF8_ATOM_CHUNK].start, stp->chunks[UTF8_ATOM_CHUNK].size); + } else if (stp->chunks[ATOM_CHUNK].start != NULL) { + MD5Update(&context, stp->chunks[ATOM_CHUNK].start, stp->chunks[ATOM_CHUNK].size); + } else { + LoadError0(stp, "mandatory chunk of type 'Atom' or 'AtU8' not found\n"); + } + + for (i = MIN_MANDATORY; i < MAX_MANDATORY; i++) { if (stp->chunks[i].start != NULL) { MD5Update(&context, stp->chunks[i].start, stp->chunks[i].size); } else { @@ -1352,7 +1375,7 @@ verify_chunks(LoaderState* stp) } static int -load_atom_table(LoaderState* stp) +load_atom_table(LoaderState* stp, ErtsAtomEncoding enc) { unsigned int i; @@ -1371,7 +1394,7 @@ load_atom_table(LoaderState* stp) GetByte(stp, n); GetString(stp, atom, n); - stp->atom[i] = erts_atom_put(atom, n, ERTS_ATOM_ENC_LATIN1, 1); + stp->atom[i] = erts_atom_put(atom, n, enc, 1); } /* @@ -5938,7 +5961,7 @@ code_get_chunk_2(BIF_ALIST_2) goto error; } if (!init_iff_file(stp, start, binary_size(Bin)) || - !scan_iff_file(stp, &chunk, 1, 1) || + !scan_iff_file(stp, &chunk, 1) || stp->chunks[0].start == NULL) { res = am_undefined; goto done; @@ -5987,7 +6010,7 @@ code_module_md5_1(BIF_ALIST_1) } stp->module = THE_NON_VALUE; /* Suppress diagnostiscs */ if (!init_iff_file(stp, bytes, binary_size(Bin)) || - !scan_iff_file(stp, chunk_types, NUM_CHUNK_TYPES, NUM_MANDATORY) || + !scan_iff_file(stp, chunk_types, NUM_CHUNK_TYPES) || !verify_chunks(stp)) { res = am_undefined; goto done; @@ -6336,7 +6359,7 @@ erts_make_stub_module(Process* p, Eterm hipe_magic_bin, Eterm Beam, Eterm Info) if (!init_iff_file(stp, bytes, size)) { goto error; } - if (!scan_iff_file(stp, chunk_types, NUM_CHUNK_TYPES, NUM_MANDATORY) || + if (!scan_iff_file(stp, chunk_types, NUM_CHUNK_TYPES) || !verify_chunks(stp)) { goto error; } @@ -6344,9 +6367,16 @@ erts_make_stub_module(Process* p, Eterm hipe_magic_bin, Eterm Beam, Eterm Info) if (!read_code_header(stp)) { goto error; } - define_file(stp, "atom table", ATOM_CHUNK); - if (!load_atom_table(stp)) { - goto error; + if (stp->chunks[UTF8_ATOM_CHUNK].size > 0) { + define_file(stp, "utf8 atom table", UTF8_ATOM_CHUNK); + if (!load_atom_table(stp, ERTS_ATOM_ENC_UTF8)) { + goto error; + } + } else { + define_file(stp, "atom table", ATOM_CHUNK); + if (!load_atom_table(stp, ERTS_ATOM_ENC_LATIN1)) { + goto error; + } } define_file(stp, "export table", EXP_CHUNK); if (!stub_read_export_table(stp)) { diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c index a3acf87000..c35002cf06 100644 --- a/erts/emulator/beam/bif.c +++ b/erts/emulator/beam/bif.c @@ -3022,8 +3022,8 @@ BIF_RETTYPE atom_to_list_1(BIF_ALIST_1) BIF_RETTYPE list_to_atom_1(BIF_ALIST_1) { Eterm res; - char *buf = (char *) erts_alloc(ERTS_ALC_T_TMP, MAX_ATOM_CHARACTERS); - Sint i = intlist_to_buf(BIF_ARG_1, buf, MAX_ATOM_CHARACTERS); + byte *buf = (byte *) erts_alloc(ERTS_ALC_T_TMP, MAX_ATOM_SZ_LIMIT); + Sint i = erts_unicode_list_to_buf(BIF_ARG_1, buf, MAX_ATOM_CHARACTERS); if (i < 0) { erts_free(ERTS_ALC_T_TMP, (void *) buf); @@ -3033,7 +3033,7 @@ BIF_RETTYPE list_to_atom_1(BIF_ALIST_1) } BIF_ERROR(BIF_P, BADARG); } - res = erts_atom_put((byte *) buf, i, ERTS_ATOM_ENC_LATIN1, 1); + res = erts_atom_put(buf, i, ERTS_ATOM_ENC_UTF8, 1); ASSERT(is_atom(res)); erts_free(ERTS_ALC_T_TMP, (void *) buf); BIF_RET(res); @@ -3043,17 +3043,17 @@ BIF_RETTYPE list_to_atom_1(BIF_ALIST_1) BIF_RETTYPE list_to_existing_atom_1(BIF_ALIST_1) { - Sint i; - char *buf = (char *) erts_alloc(ERTS_ALC_T_TMP, MAX_ATOM_CHARACTERS); + byte *buf = (byte *) erts_alloc(ERTS_ALC_T_TMP, MAX_ATOM_SZ_LIMIT); + Sint i = erts_unicode_list_to_buf(BIF_ARG_1, buf, MAX_ATOM_CHARACTERS); - if ((i = intlist_to_buf(BIF_ARG_1, buf, MAX_ATOM_CHARACTERS)) < 0) { + if (i < 0) { error: erts_free(ERTS_ALC_T_TMP, (void *) buf); BIF_ERROR(BIF_P, BADARG); } else { Eterm a; - if (erts_atom_get(buf, i, &a, ERTS_ATOM_ENC_LATIN1)) { + if (erts_atom_get((char *) buf, i, &a, ERTS_ATOM_ENC_UTF8)) { erts_free(ERTS_ALC_T_TMP, (void *) buf); BIF_RET(a); } else { diff --git a/erts/emulator/beam/erl_unicode.c b/erts/emulator/beam/erl_unicode.c index bd5e1482fb..8919898181 100644 --- a/erts/emulator/beam/erl_unicode.c +++ b/erts/emulator/beam/erl_unicode.c @@ -1890,74 +1890,57 @@ binary_to_atom(Process* proc, Eterm bin, Eterm enc, int must_exist) byte* bytes; byte *temp_alloc = NULL; Uint bin_size; + Eterm a; if ((bytes = erts_get_aligned_binary_bytes(bin, &temp_alloc)) == 0) { BIF_ERROR(proc, BADARG); } bin_size = binary_size(bin); + if (enc == am_latin1) { - Eterm a; - if (bin_size > MAX_ATOM_CHARACTERS) { - system_limit: - erts_free_aligned_binary_bytes(temp_alloc); - BIF_ERROR(proc, SYSTEM_LIMIT); - } if (!must_exist) { - a = erts_atom_put((byte *) bytes, - bin_size, - ERTS_ATOM_ENC_LATIN1, - 0); - erts_free_aligned_binary_bytes(temp_alloc); - if (is_non_value(a)) - goto badarg; - BIF_RET(a); - } else if (erts_atom_get((char *)bytes, bin_size, &a, ERTS_ATOM_ENC_LATIN1)) { - erts_free_aligned_binary_bytes(temp_alloc); - BIF_RET(a); - } else { + int lix = erts_atom_put_index((byte *) bytes, + bin_size, + ERTS_ATOM_ENC_LATIN1, + 0); + if (lix == ATOM_BAD_ENCODING_ERROR) { + badarg: + erts_free_aligned_binary_bytes(temp_alloc); + BIF_ERROR(proc, BADARG); + } else if (lix == ATOM_MAX_CHARS_ERROR) { + system_limit: + erts_free_aligned_binary_bytes(temp_alloc); + BIF_ERROR(proc, SYSTEM_LIMIT); + } + + a = make_atom(lix); + } else if (!erts_atom_get((char *)bytes, bin_size, &a, ERTS_ATOM_ENC_LATIN1)) { goto badarg; } - } else if (enc == am_utf8 || enc == am_unicode) { - Eterm res; - Uint num_chars = 0; - const byte* p = bytes; - Uint left = bin_size; - while (left) { - if (++num_chars > MAX_ATOM_CHARACTERS) { + } else if (enc == am_utf8 || enc == am_unicode) { + if (!must_exist) { + int uix = erts_atom_put_index((byte *) bytes, + bin_size, + ERTS_ATOM_ENC_UTF8, + 0); + if (uix == ATOM_BAD_ENCODING_ERROR) { + goto badarg; + } else if (uix == ATOM_MAX_CHARS_ERROR) { goto system_limit; } - if ((p[0] & 0x80) == 0) { - ++p; - --left; - } - else if (left >= 2 - && (p[0] & 0xFE) == 0xC2 /* only allow latin1 subset */ - && (p[1] & 0xC0) == 0x80) { - p += 2; - left -= 2; - } - else goto badarg; - } - if (!must_exist) { - res = erts_atom_put((byte *) bytes, - bin_size, - ERTS_ATOM_ENC_UTF8, - 0); + a = make_atom(uix); } - else if (!erts_atom_get((char*)bytes, bin_size, &res, ERTS_ATOM_ENC_UTF8)) { + else if (!erts_atom_get((char*)bytes, bin_size, &a, ERTS_ATOM_ENC_UTF8)) { goto badarg; } - erts_free_aligned_binary_bytes(temp_alloc); - if (is_non_value(res)) - goto badarg; - BIF_RET(res); } else { - badarg: - erts_free_aligned_binary_bytes(temp_alloc); - BIF_ERROR(proc, BADARG); + goto badarg; } + + erts_free_aligned_binary_bytes(temp_alloc); + BIF_RET(a); } BIF_RETTYPE binary_to_atom_2(BIF_ALIST_2) diff --git a/erts/emulator/beam/global.h b/erts/emulator/beam/global.h index c39ac2f7ec..b6568e85ea 100644 --- a/erts/emulator/beam/global.h +++ b/erts/emulator/beam/global.h @@ -1380,6 +1380,7 @@ int erts_utf8_to_latin1(byte* dest, const byte* source, int slen); void bin_write(fmtfn_t, void*, byte*, size_t); Sint intlist_to_buf(Eterm, char*, Sint); /* most callers pass plain char*'s */ +Sint erts_unicode_list_to_buf(Eterm list, byte *buf, Sint len); struct Sint_buf { #if defined(ARCH_64) diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c index 74a0681dd5..323f61c9f3 100644 --- a/erts/emulator/beam/utils.c +++ b/erts/emulator/beam/utils.c @@ -3671,6 +3671,68 @@ intlist_to_buf(Eterm list, char *buf, Sint len) return -2; /* not enough space */ } +/* Fill buf with the contents of the unicode list. + * Return the number of bytes in the buffer, + * or -1 for type error, + * or -2 for not enough buffer space (buffer contains truncated result). + */ +Sint +erts_unicode_list_to_buf(Eterm list, byte *buf, Sint len) +{ + Eterm* listptr; + Sint sz = 0; + + if (is_nil(list)) { + return 0; + } + if (is_not_list(list)) { + return -1; + } + listptr = list_val(list); + + while (len-- > 0) { + Sint val; + + if (is_not_small(CAR(listptr))) { + return -1; + } + val = signed_val(CAR(listptr)); + if (0 <= val && val < 0x80) { + buf[sz] = val; + sz++; + } else if (val < 0x800) { + buf[sz+0] = 0xC0 | (val >> 6); + buf[sz+1] = 0x80 | (val & 0x3F); + sz += 2; + } else if (val < 0x10000UL) { + if (0xD800 <= val && val <= 0xDFFF) { + return -1; + } + buf[sz+0] = 0xE0 | (val >> 12); + buf[sz+1] = 0x80 | ((val >> 6) & 0x3F); + buf[sz+2] = 0x80 | (val & 0x3F); + sz += 3; + } else if (val < 0x110000) { + buf[sz+0] = 0xF0 | (val >> 18); + buf[sz+1] = 0x80 | ((val >> 12) & 0x3F); + buf[sz+2] = 0x80 | ((val >> 6) & 0x3F); + buf[sz+3] = 0x80 | (val & 0x3F); + sz += 4; + } else { + return -1; + } + list = CDR(listptr); + if (is_nil(list)) { + return sz; + } + if (is_not_list(list)) { + return -1; + } + listptr = list_val(list); + } + return -2; /* not enough space */ +} + /* ** Convert an integer to a byte list ** return pointer to converted stuff (need not to be at start of buf!) diff --git a/erts/emulator/test/bif_SUITE.erl b/erts/emulator/test/bif_SUITE.erl index f70fb0e501..339c827602 100644 --- a/erts/emulator/test/bif_SUITE.erl +++ b/erts/emulator/test/bif_SUITE.erl @@ -26,7 +26,7 @@ -export([all/0, suite/0, display/1, display_huge/0, erl_bif_types/1,guard_bifs_in_erl_bif_types/1, - shadow_comments/1, + shadow_comments/1,list_to_utf8_atom/1, specs/1,improper_bif_stubs/1,auto_imports/1, t_list_to_existing_atom/1,os_env/1,otp_7526/1, binary_to_atom/1,binary_to_existing_atom/1, @@ -43,7 +43,7 @@ all() -> [erl_bif_types, guard_bifs_in_erl_bif_types, shadow_comments, specs, improper_bif_stubs, auto_imports, t_list_to_existing_atom, os_env, otp_7526, - display, + display, list_to_utf8_atom, atom_to_binary, binary_to_atom, binary_to_existing_atom, erl_crash_dump_bytes, min_max, erlang_halt, is_builtin, error_stacktrace, error_stacktrace_during_call_trace]. @@ -339,6 +339,38 @@ check_stub({_,F,A}, B) -> ct:fail(invalid_body) end. +list_to_utf8_atom(Config) when is_list(Config) -> + 'hello' = atom_roundtrip("hello"), + 'こんにちは' = atom_roundtrip("こんにちは"), + + %% Test all edge cases. + _ = atom_roundtrip([16#80]), + _ = atom_roundtrip([16#7F]), + _ = atom_roundtrip([16#FF]), + _ = atom_roundtrip([16#100]), + _ = atom_roundtrip([16#7FF]), + _ = atom_roundtrip([16#800]), + _ = atom_roundtrip([16#D7FF]), + atom_badarg([16#D800]), + atom_badarg([16#DFFF]), + _ = atom_roundtrip([16#E000]), + _ = atom_roundtrip([16#FFFF]), + _ = atom_roundtrip([16#1000]), + _ = atom_roundtrip([16#10FFFF]), + atom_badarg([16#110000]), + ok. + +atom_roundtrip(String) -> + Atom = list_to_atom(String), + Atom = list_to_existing_atom(String), + String = atom_to_list(Atom), + Atom. + +atom_badarg(String) -> + {'EXIT',{badarg,_}} = (catch list_to_atom(String)), + {'EXIT',{badarg,_}} = (catch list_to_existing_atom(String)), + ok. + t_list_to_existing_atom(Config) when is_list(Config) -> all = list_to_existing_atom("all"), ?MODULE = list_to_existing_atom(?MODULE_STRING), @@ -429,6 +461,8 @@ binary_to_atom(Config) when is_list(Config) -> Long = lists:seq(0, 254), LongAtom = list_to_atom(Long), LongBin = list_to_binary(Long), + UnicodeLongAtom = list_to_atom([$é || _ <- lists:seq(0, 254)]), + UnicodeLongBin = << <<"é"/utf8>> || _ <- lists:seq(0, 254)>>, %% latin1 '' = test_binary_to_atom(<<>>, latin1), @@ -440,12 +474,17 @@ binary_to_atom(Config) when is_list(Config) -> '' = test_binary_to_atom(<<>>, utf8), HalfLongAtom = test_binary_to_atom(HalfLongBin, utf8), HalfLongAtom = test_binary_to_atom(HalfLongBin, unicode), + UnicodeLongAtom = test_binary_to_atom(UnicodeLongBin, utf8), + UnicodeLongAtom = test_binary_to_atom(UnicodeLongBin, unicode), [] = [C || C <- lists:seq(128, 255), begin list_to_atom([C]) =/= test_binary_to_atom(<<C/utf8>>, utf8) end], + <<"こんにちは"/utf8>> = + atom_to_binary(test_binary_to_atom(<<"こんにちは"/utf8>>, utf8), utf8), + %% badarg failures. fail_binary_to_atom(atom), fail_binary_to_atom(42), @@ -464,10 +503,6 @@ binary_to_atom(Config) when is_list(Config) -> ?BADARG(binary_to_atom(id(<<255>>), utf8)), ?BADARG(binary_to_atom(id(<<255,0>>), utf8)), ?BADARG(binary_to_atom(id(<<16#C0,16#80>>), utf8)), %Overlong 0. - [?BADARG(binary_to_atom(<<C/utf8>>, utf8)) || C <- lists:seq(256, 16#D7FF)], - [?BADARG(binary_to_atom(<<C/utf8>>, utf8)) || C <- lists:seq(16#E000, 16#FFFD)], - [?BADARG(binary_to_atom(<<C/utf8>>, utf8)) || C <- lists:seq(16#10000, 16#8FFFF)], - [?BADARG(binary_to_atom(<<C/utf8>>, utf8)) || C <- lists:seq(16#90000, 16#10FFFF)], %% system_limit failures. ?SYS_LIMIT(binary_to_atom(id(<<0:512/unit:8,255>>), utf8)), diff --git a/erts/emulator/test/code_SUITE.erl b/erts/emulator/test/code_SUITE.erl index b29520ab9f..d07166ed98 100644 --- a/erts/emulator/test/code_SUITE.erl +++ b/erts/emulator/test/code_SUITE.erl @@ -296,16 +296,16 @@ get_chunk(Config) when is_list(Config) -> {ok,my_code_test,Code} = compile:file(File, [binary]), %% Should work. - Chunk = get_chunk_ok("Atom", Code), - Chunk = get_chunk_ok("Atom", make_sub_binary(Code)), - Chunk = get_chunk_ok("Atom", make_unaligned_sub_binary(Code)), + Chunk = get_chunk_ok("AtU8", Code), + Chunk = get_chunk_ok("AtU8", make_sub_binary(Code)), + Chunk = get_chunk_ok("AtU8", make_unaligned_sub_binary(Code)), %% Should fail. - {'EXIT',{badarg,_}} = (catch code:get_chunk(bit_sized_binary(Code), "Atom")), + {'EXIT',{badarg,_}} = (catch code:get_chunk(bit_sized_binary(Code), "AtU8")), {'EXIT',{badarg,_}} = (catch code:get_chunk(Code, "bad chunk id")), %% Invalid beam code or missing chunk should return 'undefined'. - undefined = code:get_chunk(<<"not a beam module">>, "Atom"), + undefined = code:get_chunk(<<"not a beam module">>, "AtU8"), undefined = code:get_chunk(Code, "XXXX"), ok. |