Ruby: encoding.c Source File

00001 /**********************************************************************
00002 
00003   encoding.c -
00004 
00005   $Author: naruse $
00006   created at: Thu May 24 17:23:27 JST 2007
00007 
00008   Copyright (C) 2007 Yukihiro Matsumoto
00009 
00010 **********************************************************************/
00011 
00012 #include "ruby/ruby.h"
00013 #include "ruby/encoding.h"
00014 #include "internal.h"
00015 #include "regenc.h"
00016 #include <ctype.h>
00017 #ifndef NO_LOCALE_CHARMAP
00018 #ifdef __CYGWIN__
00019 #include <windows.h>
00020 #endif
00021 #ifdef HAVE_LANGINFO_H
00022 #include <langinfo.h>
00023 #endif
00024 #endif
00025 #include "ruby/util.h"
00026 
00027 #if defined __GNUC__ && __GNUC__ >= 4
00028 #pragma GCC visibility push(default)
00029 int rb_enc_register(const char *name, rb_encoding *encoding);
00030 void rb_enc_set_base(const char *name, const char *orig);
00031 void rb_encdb_declare(const char *name);
00032 int rb_encdb_replicate(const char *name, const char *orig);
00033 int rb_encdb_dummy(const char *name);
00034 int rb_encdb_alias(const char *alias, const char *orig);
00035 #pragma GCC visibility pop
00036 #endif
00037 
00038 static ID id_encoding;
00039 VALUE rb_cEncoding;
00040 static VALUE rb_encoding_list;
00041 
00042 struct rb_encoding_entry {
00043     const char *name;
00044     rb_encoding *enc;
00045     rb_encoding *base;
00046 };
00047 
00048 static struct {
00049     struct rb_encoding_entry *list;
00050     int count;
00051     int size;
00052     st_table *names;
00053 } enc_table;
00054 
00055 void rb_enc_init(void);
00056 
00057 #define ENCODING_COUNT ENCINDEX_BUILTIN_MAX
00058 #define UNSPECIFIED_ENCODING INT_MAX
00059 
00060 #define ENCODING_NAMELEN_MAX 63
00061 #define valid_encoding_name_p(name) ((name) && strlen(name) <= ENCODING_NAMELEN_MAX)
00062 
00063 #define enc_autoload_p(enc) (!rb_enc_mbmaxlen(enc))
00064 
00065 static int load_encoding(const char *name);
00066 
00067 static size_t
00068 enc_memsize(const void *p)
00069 {
00070     return 0;
00071 }
00072 
00073 static const rb_data_type_t encoding_data_type = {
00074     "encoding",
00075     {0, 0, enc_memsize,},
00076 };
00077 
00078 #define is_data_encoding(obj) (RTYPEDDATA_P(obj) && RTYPEDDATA_TYPE(obj) == &encoding_data_type)
00079 
00080 static VALUE
00081 enc_new(rb_encoding *encoding)
00082 {
00083     return TypedData_Wrap_Struct(rb_cEncoding, &encoding_data_type, encoding);
00084 }
00085 
00086 static VALUE
00087 rb_enc_from_encoding_index(int idx)
00088 {
00089     VALUE list, enc;
00090 
00091     if (!(list = rb_encoding_list)) {
00092         rb_bug("rb_enc_from_encoding_index(%d): no rb_encoding_list", idx);
00093     }
00094     enc = rb_ary_entry(list, idx);
00095     if (NIL_P(enc)) {
00096         rb_bug("rb_enc_from_encoding_index(%d): not created yet", idx);
00097     }
00098     return enc;
00099 }
00100 
00101 VALUE
00102 rb_enc_from_encoding(rb_encoding *encoding)
00103 {
00104     int idx;
00105     if (!encoding) return Qnil;
00106     idx = ENC_TO_ENCINDEX(encoding);
00107     return rb_enc_from_encoding_index(idx);
00108 }
00109 
00110 static int enc_autoload(rb_encoding *);
00111 
00112 static int
00113 check_encoding(rb_encoding *enc)
00114 {
00115     int index = rb_enc_to_index(enc);
00116     if (rb_enc_from_index(index) != enc)
00117         return -1;
00118     if (enc_autoload_p(enc)) {
00119         index = enc_autoload(enc);
00120     }
00121     return index;
00122 }
00123 
00124 static int
00125 enc_check_encoding(VALUE obj)
00126 {
00127     if (SPECIAL_CONST_P(obj) || !rb_typeddata_is_kind_of(obj, &encoding_data_type)) {
00128         return -1;
00129     }
00130     return check_encoding(RDATA(obj)->data);
00131 }
00132 
00133 static int
00134 must_encoding(VALUE enc)
00135 {
00136     int index = enc_check_encoding(enc);
00137     if (index < 0) {
00138         rb_raise(rb_eTypeError, "wrong argument type %s (expected Encoding)",
00139                  rb_obj_classname(enc));
00140     }
00141     return index;
00142 }
00143 
00144 int
00145 rb_to_encoding_index(VALUE enc)
00146 {
00147     int idx;
00148 
00149     idx = enc_check_encoding(enc);
00150     if (idx >= 0) {
00151         return idx;
00152     }
00153     else if (NIL_P(enc = rb_check_string_type(enc))) {
00154         return -1;
00155     }
00156     if (!rb_enc_asciicompat(rb_enc_get(enc))) {
00157         return -1;
00158     }
00159     return rb_enc_find_index(StringValueCStr(enc));
00160 }
00161 
00162 static rb_encoding *
00163 to_encoding(VALUE enc)
00164 {
00165     int idx;
00166 
00167     StringValue(enc);
00168     if (!rb_enc_asciicompat(rb_enc_get(enc))) {
00169         rb_raise(rb_eArgError, "invalid name encoding (non ASCII)");
00170     }
00171     idx = rb_enc_find_index(StringValueCStr(enc));
00172     if (idx < 0) {
00173         rb_raise(rb_eArgError, "unknown encoding name - %s", RSTRING_PTR(enc));
00174     }
00175     return rb_enc_from_index(idx);
00176 }
00177 
00178 rb_encoding *
00179 rb_to_encoding(VALUE enc)
00180 {
00181     if (enc_check_encoding(enc) >= 0) return RDATA(enc)->data;
00182     return to_encoding(enc);
00183 }
00184 
00185 void
00186 rb_gc_mark_encodings(void)
00187 {
00188 }
00189 
00190 static int
00191 enc_table_expand(int newsize)
00192 {
00193     struct rb_encoding_entry *ent;
00194     int count = newsize;
00195 
00196     if (enc_table.size >= newsize) return newsize;
00197     newsize = (newsize + 7) / 8 * 8;
00198     ent = realloc(enc_table.list, sizeof(*enc_table.list) * newsize);
00199     if (!ent) return -1;
00200     memset(ent + enc_table.size, 0, sizeof(*ent)*(newsize - enc_table.size));
00201     enc_table.list = ent;
00202     enc_table.size = newsize;
00203     return count;
00204 }
00205 
00206 static int
00207 enc_register_at(int index, const char *name, rb_encoding *encoding)
00208 {
00209     struct rb_encoding_entry *ent = &enc_table.list[index];
00210     VALUE list;
00211 
00212     if (!valid_encoding_name_p(name)) return -1;
00213     if (!ent->name) {
00214         ent->name = name = strdup(name);
00215     }
00216     else if (STRCASECMP(name, ent->name)) {
00217         return -1;
00218     }
00219     if (!ent->enc) {
00220         ent->enc = xmalloc(sizeof(rb_encoding));
00221     }
00222     if (encoding) {
00223         *ent->enc = *encoding;
00224     }
00225     else {
00226         memset(ent->enc, 0, sizeof(*ent->enc));
00227     }
00228     encoding = ent->enc;
00229     encoding->name = name;
00230     encoding->ruby_encoding_index = index;
00231     st_insert(enc_table.names, (st_data_t)name, (st_data_t)index);
00232     list = rb_encoding_list;
00233     if (list && NIL_P(rb_ary_entry(list, index))) {
00234         /* initialize encoding data */
00235         rb_ary_store(list, index, enc_new(encoding));
00236     }
00237     return index;
00238 }
00239 
00240 static int
00241 enc_register(const char *name, rb_encoding *encoding)
00242 {
00243     int index = enc_table.count;
00244 
00245     if ((index = enc_table_expand(index + 1)) < 0) return -1;
00246     enc_table.count = index;
00247     return enc_register_at(index - 1, name, encoding);
00248 }
00249 
00250 static void set_encoding_const(const char *, rb_encoding *);
00251 int rb_enc_registered(const char *name);
00252 
00253 int
00254 rb_enc_register(const char *name, rb_encoding *encoding)
00255 {
00256     int index = rb_enc_registered(name);
00257 
00258     if (index >= 0) {
00259         rb_encoding *oldenc = rb_enc_from_index(index);
00260         if (STRCASECMP(name, rb_enc_name(oldenc))) {
00261             index = enc_register(name, encoding);
00262         }
00263         else if (enc_autoload_p(oldenc) || !ENC_DUMMY_P(oldenc)) {
00264             enc_register_at(index, name, encoding);
00265         }
00266         else {
00267             rb_raise(rb_eArgError, "encoding %s is already registered", name);
00268         }
00269     }
00270     else {
00271         index = enc_register(name, encoding);
00272         set_encoding_const(name, rb_enc_from_index(index));
00273     }
00274     return index;
00275 }
00276 
00277 void
00278 rb_encdb_declare(const char *name)
00279 {
00280     int idx = rb_enc_registered(name);
00281     if (idx < 0) {
00282         idx = enc_register(name, 0);
00283     }
00284     set_encoding_const(name, rb_enc_from_index(idx));
00285 }
00286 
00287 static void
00288 enc_check_duplication(const char *name)
00289 {
00290     if (rb_enc_registered(name) >= 0) {
00291         rb_raise(rb_eArgError, "encoding %s is already registered", name);
00292     }
00293 }
00294 
00295 static rb_encoding*
00296 set_base_encoding(int index, rb_encoding *base)
00297 {
00298     rb_encoding *enc = enc_table.list[index].enc;
00299 
00300     enc_table.list[index].base = base;
00301     if (rb_enc_dummy_p(base)) ENC_SET_DUMMY(enc);
00302     return enc;
00303 }
00304 
00305 /* for encdb.h
00306  * Set base encoding for encodings which are not replicas
00307  * but not in their own files.
00308  */
00309 void
00310 rb_enc_set_base(const char *name, const char *orig)
00311 {
00312     int idx = rb_enc_registered(name);
00313     int origidx = rb_enc_registered(orig);
00314     set_base_encoding(idx, rb_enc_from_index(origidx));
00315 }
00316 
00317 int
00318 rb_enc_replicate(const char *name, rb_encoding *encoding)
00319 {
00320     int idx;
00321 
00322     enc_check_duplication(name);
00323     idx = enc_register(name, encoding);
00324     set_base_encoding(idx, encoding);
00325     set_encoding_const(name, rb_enc_from_index(idx));
00326     return idx;
00327 }
00328 
00329 /*
00330  * call-seq:
00331  *   enc.replicate(name) -> encoding
00332  *
00333  * Returns a replicated encoding of _enc_ whose name is _name_.
00334  * The new encoding should have the same byte structure of _enc_.
00335  * If _name_ is used by another encoding, raise ArgumentError.
00336  *
00337  */
00338 static VALUE
00339 enc_replicate(VALUE encoding, VALUE name)
00340 {
00341     return rb_enc_from_encoding_index(
00342         rb_enc_replicate(StringValueCStr(name),
00343                          rb_to_encoding(encoding)));
00344 }
00345 
00346 static int
00347 enc_replicate_with_index(const char *name, rb_encoding *origenc, int idx)
00348 {
00349     if (idx < 0) {
00350         idx = enc_register(name, origenc);
00351     }
00352     else {
00353         idx = enc_register_at(idx, name, origenc);
00354     }
00355     if (idx >= 0) {
00356         set_base_encoding(idx, origenc);
00357         set_encoding_const(name, rb_enc_from_index(idx));
00358     }
00359     return idx;
00360 }
00361 
00362 int
00363 rb_encdb_replicate(const char *name, const char *orig)
00364 {
00365     int origidx = rb_enc_registered(orig);
00366     int idx = rb_enc_registered(name);
00367 
00368     if (origidx < 0) {
00369         origidx = enc_register(orig, 0);
00370     }
00371     return enc_replicate_with_index(name, rb_enc_from_index(origidx), idx);
00372 }
00373 
00374 int
00375 rb_define_dummy_encoding(const char *name)
00376 {
00377     int index = rb_enc_replicate(name, rb_ascii8bit_encoding());
00378     rb_encoding *enc = enc_table.list[index].enc;
00379 
00380     ENC_SET_DUMMY(enc);
00381     return index;
00382 }
00383 
00384 int
00385 rb_encdb_dummy(const char *name)
00386 {
00387     int index = enc_replicate_with_index(name, rb_ascii8bit_encoding(),
00388                                          rb_enc_registered(name));
00389     rb_encoding *enc = enc_table.list[index].enc;
00390 
00391     ENC_SET_DUMMY(enc);
00392     return index;
00393 }
00394 
00395 /*
00396  * call-seq:
00397  *   enc.dummy? -> true or false
00398  *
00399  * Returns true for dummy encodings.
00400  * A dummy encoding is an encoding for which character handling is not properly
00401  * implemented.
00402  * It is used for stateful encodings.
00403  *
00404  *   Encoding::ISO_2022_JP.dummy?       #=> true
00405  *   Encoding::UTF_8.dummy?             #=> false
00406  *
00407  */
00408 static VALUE
00409 enc_dummy_p(VALUE enc)
00410 {
00411     return ENC_DUMMY_P(enc_table.list[must_encoding(enc)].enc) ? Qtrue : Qfalse;
00412 }
00413 
00414 /*
00415  * call-seq:
00416  *   enc.ascii_compatible? -> true or false
00417  *
00418  * Returns whether ASCII-compatible or not.
00419  *
00420  *   Encoding::UTF_8.ascii_compatible?     #=> true
00421  *   Encoding::UTF_16BE.ascii_compatible?  #=> false
00422  *
00423  */
00424 static VALUE
00425 enc_ascii_compatible_p(VALUE enc)
00426 {
00427     return rb_enc_asciicompat(enc_table.list[must_encoding(enc)].enc) ? Qtrue : Qfalse;
00428 }
00429 
00430 /*
00431  * Returns 1 when the encoding is Unicode series other than UTF-7 else 0.
00432  */
00433 int
00434 rb_enc_unicode_p(rb_encoding *enc)
00435 {
00436     const char *name = rb_enc_name(enc);
00437     return name[0] == 'U' && name[1] == 'T' && name[2] == 'F' && name[4] != '7';
00438 }
00439 
00440 /*
00441  * Returns copied alias name when the key is added for st_table,
00442  * else returns NULL.
00443  */
00444 static int
00445 enc_alias_internal(const char *alias, int idx)
00446 {
00447     return st_insert2(enc_table.names, (st_data_t)alias, (st_data_t)idx,
00448             (st_data_t(*)(st_data_t))strdup);
00449 }
00450 
00451 static int
00452 enc_alias(const char *alias, int idx)
00453 {
00454     if (!valid_encoding_name_p(alias)) return -1;
00455     if (!enc_alias_internal(alias, idx))
00456         set_encoding_const(alias, rb_enc_from_index(idx));
00457     return idx;
00458 }
00459 
00460 int
00461 rb_enc_alias(const char *alias, const char *orig)
00462 {
00463     int idx;
00464 
00465     enc_check_duplication(alias);
00466     if (!enc_table.list) {
00467         rb_enc_init();
00468     }
00469     if ((idx = rb_enc_find_index(orig)) < 0) {
00470         return -1;
00471     }
00472     return enc_alias(alias, idx);
00473 }
00474 
00475 int
00476 rb_encdb_alias(const char *alias, const char *orig)
00477 {
00478     int idx = rb_enc_registered(orig);
00479 
00480     if (idx < 0) {
00481         idx = enc_register(orig, 0);
00482     }
00483     return enc_alias(alias, idx);
00484 }
00485 
00486 enum {
00487     ENCINDEX_ASCII,
00488     ENCINDEX_UTF_8,
00489     ENCINDEX_US_ASCII,
00490     ENCINDEX_BUILTIN_MAX
00491 };
00492 
00493 extern rb_encoding OnigEncodingUTF_8;
00494 extern rb_encoding OnigEncodingUS_ASCII;
00495 
00496 void
00497 rb_enc_init(void)
00498 {
00499     enc_table_expand(ENCODING_COUNT + 1);
00500     if (!enc_table.names) {
00501         enc_table.names = st_init_strcasetable();
00502     }
00503 #define ENC_REGISTER(enc) enc_register_at(ENCINDEX_##enc, rb_enc_name(&OnigEncoding##enc), &OnigEncoding##enc)
00504     ENC_REGISTER(ASCII);
00505     ENC_REGISTER(UTF_8);
00506     ENC_REGISTER(US_ASCII);
00507 #undef ENC_REGISTER
00508     enc_table.count = ENCINDEX_BUILTIN_MAX;
00509 }
00510 
00511 rb_encoding *
00512 rb_enc_from_index(int index)
00513 {
00514     if (!enc_table.list) {
00515         rb_enc_init();
00516     }
00517     if (index < 0 || enc_table.count <= index) {
00518         return 0;
00519     }
00520     return enc_table.list[index].enc;
00521 }
00522 
00523 int
00524 rb_enc_registered(const char *name)
00525 {
00526     st_data_t idx = 0;
00527 
00528     if (!name) return -1;
00529     if (!enc_table.list) return -1;
00530     if (st_lookup(enc_table.names, (st_data_t)name, &idx)) {
00531         return (int)idx;
00532     }
00533     return -1;
00534 }
00535 
00536 static VALUE
00537 require_enc(VALUE enclib)
00538 {
00539     int safe = rb_safe_level();
00540     return rb_require_safe(enclib, safe > 3 ? 3 : safe);
00541 }
00542 
00543 static int
00544 load_encoding(const char *name)
00545 {
00546     VALUE enclib = rb_sprintf("enc/%s.so", name);
00547     VALUE verbose = ruby_verbose;
00548     VALUE debug = ruby_debug;
00549     VALUE loaded;
00550     char *s = RSTRING_PTR(enclib) + 4, *e = RSTRING_END(enclib) - 3;
00551     int idx;
00552 
00553     while (s < e) {
00554         if (!ISALNUM(*s)) *s = '_';
00555         else if (ISUPPER(*s)) *s = TOLOWER(*s);
00556         ++s;
00557     }
00558     FL_UNSET(enclib, FL_TAINT|FL_UNTRUSTED);
00559     OBJ_FREEZE(enclib);
00560     ruby_verbose = Qfalse;
00561     ruby_debug = Qfalse;
00562     loaded = rb_protect(require_enc, enclib, 0);
00563     ruby_verbose = verbose;
00564     ruby_debug = debug;
00565     rb_set_errinfo(Qnil);
00566     if (NIL_P(loaded)) return -1;
00567     if ((idx = rb_enc_registered(name)) < 0) return -1;
00568     if (enc_autoload_p(enc_table.list[idx].enc)) return -1;
00569     return idx;
00570 }
00571 
00572 static int
00573 enc_autoload(rb_encoding *enc)
00574 {
00575     int i;
00576     rb_encoding *base = enc_table.list[ENC_TO_ENCINDEX(enc)].base;
00577 
00578     if (base) {
00579         i = 0;
00580         do {
00581             if (i >= enc_table.count) return -1;
00582         } while (enc_table.list[i].enc != base && (++i, 1));
00583         if (enc_autoload_p(base)) {
00584             if (enc_autoload(base) < 0) return -1;
00585         }
00586         i = ENC_TO_ENCINDEX(enc);
00587         enc_register_at(i, rb_enc_name(enc), base);
00588     }
00589     else {
00590         i = load_encoding(rb_enc_name(enc));
00591     }
00592     return i;
00593 }
00594 
00595 int
00596 rb_enc_find_index(const char *name)
00597 {
00598     int i = rb_enc_registered(name);
00599     rb_encoding *enc;
00600 
00601     if (i < 0) {
00602         i = load_encoding(name);
00603     }
00604     else if (!(enc = rb_enc_from_index(i))) {
00605         if (i != UNSPECIFIED_ENCODING) {
00606             rb_raise(rb_eArgError, "encoding %s is not registered", name);
00607         }
00608     }
00609     else if (enc_autoload_p(enc)) {
00610         if (enc_autoload(enc) < 0) {
00611             rb_warn("failed to load encoding (%s); use ASCII-8BIT instead",
00612                     name);
00613             return 0;
00614         }
00615     }
00616     return i;
00617 }
00618 
00619 rb_encoding *
00620 rb_enc_find(const char *name)
00621 {
00622     int idx = rb_enc_find_index(name);
00623     if (idx < 0) idx = 0;
00624     return rb_enc_from_index(idx);
00625 }
00626 
00627 static inline int
00628 enc_capable(VALUE obj)
00629 {
00630     if (SPECIAL_CONST_P(obj)) return SYMBOL_P(obj);
00631     switch (BUILTIN_TYPE(obj)) {
00632       case T_STRING:
00633       case T_REGEXP:
00634       case T_FILE:
00635         return TRUE;
00636       case T_DATA:
00637         if (is_data_encoding(obj)) return TRUE;
00638       default:
00639         return FALSE;
00640     }
00641 }
00642 
00643 ID
00644 rb_id_encoding(void)
00645 {
00646     CONST_ID(id_encoding, "encoding");
00647     return id_encoding;
00648 }
00649 
00650 int
00651 rb_enc_get_index(VALUE obj)
00652 {
00653     int i = -1;
00654     VALUE tmp;
00655 
00656     if (SPECIAL_CONST_P(obj)) {
00657         if (!SYMBOL_P(obj)) return -1;
00658         obj = rb_id2str(SYM2ID(obj));
00659     }
00660     switch (BUILTIN_TYPE(obj)) {
00661       as_default:
00662       default:
00663       case T_STRING:
00664       case T_REGEXP:
00665         i = ENCODING_GET_INLINED(obj);
00666         if (i == ENCODING_INLINE_MAX) {
00667             VALUE iv;
00668 
00669             iv = rb_ivar_get(obj, rb_id_encoding());
00670             i = NUM2INT(iv);
00671         }
00672         break;
00673       case T_FILE:
00674         tmp = rb_funcall(obj, rb_intern("internal_encoding"), 0, 0);
00675         if (NIL_P(tmp)) obj = rb_funcall(obj, rb_intern("external_encoding"), 0, 0);
00676         else obj = tmp;
00677         if (NIL_P(obj)) break;
00678       case T_DATA:
00679         if (is_data_encoding(obj)) {
00680             i = enc_check_encoding(obj);
00681         }
00682         else {
00683             goto as_default;
00684         }
00685         break;
00686     }
00687     return i;
00688 }
00689 
00690 static void
00691 enc_set_index(VALUE obj, int idx)
00692 {
00693     if (idx < ENCODING_INLINE_MAX) {
00694         ENCODING_SET_INLINED(obj, idx);
00695         return;
00696     }
00697     ENCODING_SET_INLINED(obj, ENCODING_INLINE_MAX);
00698     rb_ivar_set(obj, rb_id_encoding(), INT2NUM(idx));
00699 }
00700 
00701 void
00702 rb_enc_set_index(VALUE obj, int idx)
00703 {
00704     rb_check_frozen(obj);
00705     enc_set_index(obj, idx);
00706 }
00707 
00708 VALUE
00709 rb_enc_associate_index(VALUE obj, int idx)
00710 {
00711 /*    enc_check_capable(obj);*/
00712     rb_check_frozen(obj);
00713     if (rb_enc_get_index(obj) == idx)
00714         return obj;
00715     if (SPECIAL_CONST_P(obj)) {
00716         rb_raise(rb_eArgError, "cannot set encoding");
00717     }
00718     if (!ENC_CODERANGE_ASCIIONLY(obj) ||
00719         !rb_enc_asciicompat(rb_enc_from_index(idx))) {
00720         ENC_CODERANGE_CLEAR(obj);
00721     }
00722     enc_set_index(obj, idx);
00723     return obj;
00724 }
00725 
00726 VALUE
00727 rb_enc_associate(VALUE obj, rb_encoding *enc)
00728 {
00729     return rb_enc_associate_index(obj, rb_enc_to_index(enc));
00730 }
00731 
00732 rb_encoding*
00733 rb_enc_get(VALUE obj)
00734 {
00735     return rb_enc_from_index(rb_enc_get_index(obj));
00736 }
00737 
00738 rb_encoding*
00739 rb_enc_check(VALUE str1, VALUE str2)
00740 {
00741     rb_encoding *enc = rb_enc_compatible(str1, str2);
00742     if (!enc)
00743         rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s",
00744                  rb_enc_name(rb_enc_get(str1)),
00745                  rb_enc_name(rb_enc_get(str2)));
00746     return enc;
00747 }
00748 
00749 rb_encoding*
00750 rb_enc_compatible(VALUE str1, VALUE str2)
00751 {
00752     int idx1, idx2;
00753     rb_encoding *enc1, *enc2;
00754     int isstr1, isstr2;
00755 
00756     idx1 = rb_enc_get_index(str1);
00757     idx2 = rb_enc_get_index(str2);
00758 
00759     if (idx1 < 0 || idx2 < 0)
00760         return 0;
00761 
00762     if (idx1 == idx2) {
00763         return rb_enc_from_index(idx1);
00764     }
00765     enc1 = rb_enc_from_index(idx1);
00766     enc2 = rb_enc_from_index(idx2);
00767 
00768     isstr2 = RB_TYPE_P(str2, T_STRING);
00769     if (isstr2 && RSTRING_LEN(str2) == 0)
00770         return enc1;
00771     isstr1 = RB_TYPE_P(str1, T_STRING);
00772     if (isstr1 && RSTRING_LEN(str1) == 0)
00773         return (rb_enc_asciicompat(enc1) && rb_enc_str_asciionly_p(str2)) ? enc1 : enc2;
00774     if (!rb_enc_asciicompat(enc1) || !rb_enc_asciicompat(enc2)) {
00775         return 0;
00776     }
00777 
00778     /* objects whose encoding is the same of contents */
00779     if (!isstr2 && idx2 == ENCINDEX_US_ASCII)
00780         return enc1;
00781     if (!isstr1 && idx1 == ENCINDEX_US_ASCII)
00782         return enc2;
00783 
00784     if (!isstr1) {
00785         VALUE tmp = str1;
00786         int idx0 = idx1;
00787         str1 = str2;
00788         str2 = tmp;
00789         idx1 = idx2;
00790         idx2 = idx0;
00791         idx0 = isstr1;
00792         isstr1 = isstr2;
00793         isstr2 = idx0;
00794     }
00795     if (isstr1) {
00796         int cr1, cr2;
00797 
00798         cr1 = rb_enc_str_coderange(str1);
00799         if (isstr2) {
00800             cr2 = rb_enc_str_coderange(str2);
00801             if (cr1 != cr2) {
00802                 /* may need to handle ENC_CODERANGE_BROKEN */
00803                 if (cr1 == ENC_CODERANGE_7BIT) return enc2;
00804                 if (cr2 == ENC_CODERANGE_7BIT) return enc1;
00805             }
00806             if (cr2 == ENC_CODERANGE_7BIT) {
00807                 return enc1;
00808             }
00809         }
00810         if (cr1 == ENC_CODERANGE_7BIT)
00811             return enc2;
00812     }
00813     return 0;
00814 }
00815 
00816 void
00817 rb_enc_copy(VALUE obj1, VALUE obj2)
00818 {
00819     rb_enc_associate_index(obj1, rb_enc_get_index(obj2));
00820 }
00821 
00822 
00823 /*
00824  *  call-seq:
00825  *     obj.encoding   -> encoding
00826  *
00827  *  Returns the Encoding object that represents the encoding of obj.
00828  */
00829 
00830 VALUE
00831 rb_obj_encoding(VALUE obj)
00832 {
00833     rb_encoding *enc = rb_enc_get(obj);
00834     if (!enc) {
00835         rb_raise(rb_eTypeError, "unknown encoding");
00836     }
00837     return rb_enc_from_encoding(enc);
00838 }
00839 
00840 int
00841 rb_enc_fast_mbclen(const char *p, const char *e, rb_encoding *enc)
00842 {
00843     return ONIGENC_MBC_ENC_LEN(enc, (UChar*)p, (UChar*)e);
00844 }
00845 
00846 int
00847 rb_enc_mbclen(const char *p, const char *e, rb_encoding *enc)
00848 {
00849     int n = ONIGENC_PRECISE_MBC_ENC_LEN(enc, (UChar*)p, (UChar*)e);
00850     if (MBCLEN_CHARFOUND_P(n) && MBCLEN_CHARFOUND_LEN(n) <= e-p)
00851         return MBCLEN_CHARFOUND_LEN(n);
00852     else {
00853         int min = rb_enc_mbminlen(enc);
00854         return min <= e-p ? min : (int)(e-p);
00855     }
00856 }
00857 
00858 int
00859 rb_enc_precise_mbclen(const char *p, const char *e, rb_encoding *enc)
00860 {
00861     int n;
00862     if (e <= p)
00863         return ONIGENC_CONSTRUCT_MBCLEN_NEEDMORE(1);
00864     n = ONIGENC_PRECISE_MBC_ENC_LEN(enc, (UChar*)p, (UChar*)e);
00865     if (e-p < n)
00866         return ONIGENC_CONSTRUCT_MBCLEN_NEEDMORE(n-(int)(e-p));
00867     return n;
00868 }
00869 
00870 int
00871 rb_enc_ascget(const char *p, const char *e, int *len, rb_encoding *enc)
00872 {
00873     unsigned int c, l;
00874     if (e <= p)
00875         return -1;
00876     if (rb_enc_asciicompat(enc)) {
00877         c = (unsigned char)*p;
00878         if (!ISASCII(c))
00879             return -1;
00880         if (len) *len = 1;
00881         return c;
00882     }
00883     l = rb_enc_precise_mbclen(p, e, enc);
00884     if (!MBCLEN_CHARFOUND_P(l))
00885         return -1;
00886     c = rb_enc_mbc_to_codepoint(p, e, enc);
00887     if (!rb_enc_isascii(c, enc))
00888         return -1;
00889     if (len) *len = l;
00890     return c;
00891 }
00892 
00893 unsigned int
00894 rb_enc_codepoint_len(const char *p, const char *e, int *len_p, rb_encoding *enc)
00895 {
00896     int r;
00897     if (e <= p)
00898         rb_raise(rb_eArgError, "empty string");
00899     r = rb_enc_precise_mbclen(p, e, enc);
00900     if (MBCLEN_CHARFOUND_P(r)) {
00901         if (len_p) *len_p = MBCLEN_CHARFOUND_LEN(r);
00902         return rb_enc_mbc_to_codepoint(p, e, enc);
00903     }
00904     else
00905         rb_raise(rb_eArgError, "invalid byte sequence in %s", rb_enc_name(enc));
00906 }
00907 
00908 #undef rb_enc_codepoint
00909 unsigned int
00910 rb_enc_codepoint(const char *p, const char *e, rb_encoding *enc)
00911 {
00912     return rb_enc_codepoint_len(p, e, 0, enc);
00913 }
00914 
00915 int
00916 rb_enc_codelen(int c, rb_encoding *enc)
00917 {
00918     int n = ONIGENC_CODE_TO_MBCLEN(enc,c);
00919     if (n == 0) {
00920         rb_raise(rb_eArgError, "invalid codepoint 0x%x in %s", c, rb_enc_name(enc));
00921     }
00922     return n;
00923 }
00924 
00925 int
00926 rb_enc_toupper(int c, rb_encoding *enc)
00927 {
00928     return (ONIGENC_IS_ASCII_CODE(c)?ONIGENC_ASCII_CODE_TO_UPPER_CASE(c):(c));
00929 }
00930 
00931 int
00932 rb_enc_tolower(int c, rb_encoding *enc)
00933 {
00934     return (ONIGENC_IS_ASCII_CODE(c)?ONIGENC_ASCII_CODE_TO_LOWER_CASE(c):(c));
00935 }
00936 
00937 /*
00938  * call-seq:
00939  *   enc.inspect -> string
00940  *
00941  * Returns a string which represents the encoding for programmers.
00942  *
00943  *   Encoding::UTF_8.inspect       #=> "#<Encoding:UTF-8>"
00944  *   Encoding::ISO_2022_JP.inspect #=> "#<Encoding:ISO-2022-JP (dummy)>"
00945  */
00946 static VALUE
00947 enc_inspect(VALUE self)
00948 {
00949     VALUE str = rb_sprintf("#<%s:%s%s>", rb_obj_classname(self),
00950                       rb_enc_name((rb_encoding*)DATA_PTR(self)),
00951                       (enc_dummy_p(self) ? " (dummy)" : ""));
00952     ENCODING_CODERANGE_SET(str, rb_usascii_encindex(), ENC_CODERANGE_7BIT);
00953     return str;
00954 }
00955 
00956 /*
00957  * call-seq:
00958  *   enc.name -> string
00959  *
00960  * Returns the name of the encoding.
00961  *
00962  *   Encoding::UTF_8.name      #=> "UTF-8"
00963  */
00964 static VALUE
00965 enc_name(VALUE self)
00966 {
00967     return rb_usascii_str_new2(rb_enc_name((rb_encoding*)DATA_PTR(self)));
00968 }
00969 
00970 static int
00971 enc_names_i(st_data_t name, st_data_t idx, st_data_t args)
00972 {
00973     VALUE *arg = (VALUE *)args;
00974 
00975     if ((int)idx == (int)arg[0]) {
00976         VALUE str = rb_usascii_str_new2((char *)name);
00977         OBJ_FREEZE(str);
00978         rb_ary_push(arg[1], str);
00979     }
00980     return ST_CONTINUE;
00981 }
00982 
00983 /*
00984  * call-seq:
00985  *   enc.names -> array
00986  *
00987  * Returns the list of name and aliases of the encoding.
00988  *
00989  *   Encoding::WINDOWS_31J.names  #=> ["Windows-31J", "CP932", "csWindows31J"]
00990  */
00991 static VALUE
00992 enc_names(VALUE self)
00993 {
00994     VALUE args[2];
00995 
00996     args[0] = (VALUE)rb_to_encoding_index(self);
00997     args[1] = rb_ary_new2(0);
00998     st_foreach(enc_table.names, enc_names_i, (st_data_t)args);
00999     return args[1];
01000 }
01001 
01002 /*
01003  * call-seq:
01004  *   Encoding.list -> [enc1, enc2, ...]
01005  *
01006  * Returns the list of loaded encodings.
01007  *
01008  *   Encoding.list
01009  *   #=> [#<Encoding:ASCII-8BIT>, #<Encoding:UTF-8>,
01010  *         #<Encoding:ISO-2022-JP (dummy)>]
01011  *
01012  *   Encoding.find("US-ASCII")
01013  *   #=> #<Encoding:US-ASCII>
01014  *
01015  *   Encoding.list
01016  *   #=> [#<Encoding:ASCII-8BIT>, #<Encoding:UTF-8>,
01017  *         #<Encoding:US-ASCII>, #<Encoding:ISO-2022-JP (dummy)>]
01018  *
01019  */
01020 static VALUE
01021 enc_list(VALUE klass)
01022 {
01023     VALUE ary = rb_ary_new2(0);
01024     rb_ary_replace(ary, rb_encoding_list);
01025     return ary;
01026 }
01027 
01028 /*
01029  * call-seq:
01030  *   Encoding.find(string) -> enc
01031  *   Encoding.find(symbol) -> enc
01032  *
01033  * Search the encoding with specified <i>name</i>.
01034  * <i>name</i> should be a string or symbol.
01035  *
01036  *   Encoding.find("US-ASCII")  #=> #<Encoding:US-ASCII>
01037  *   Encoding.find(:Shift_JIS)  #=> #<Encoding:Shift_JIS>
01038  *
01039  * Names which this method accept are encoding names and aliases
01040  * including following special aliases
01041  *
01042  * "external"::   default external encoding
01043  * "internal"::   default internal encoding
01044  * "locale"::     locale encoding
01045  * "filesystem":: filesystem encoding
01046  *
01047  * An ArgumentError is raised when no encoding with <i>name</i>.
01048  * Only <code>Encoding.find("internal")</code> however returns nil
01049  * when no encoding named "internal", in other words, when Ruby has no
01050  * default internal encoding.
01051  */
01052 static VALUE
01053 enc_find(VALUE klass, VALUE enc)
01054 {
01055     return rb_enc_from_encoding(rb_to_encoding(enc));
01056 }
01057 
01058 /*
01059  * call-seq:
01060  *   Encoding.compatible?(obj1, obj2) -> enc or nil
01061  *
01062  * Checks the compatibility of two objects.
01063  *
01064  * If the objects are both strings they are compatible when they are
01065  * concatenatable.  The encoding of the concatenated string will be returned
01066  * if they are compatible, nil if they are not.
01067  *
01068  *   Encoding.compatible?("\xa1".force_encoding("iso-8859-1"), "b")
01069  *   #=> #<Encoding:ISO-8859-1>
01070  *
01071  *   Encoding.compatible?(
01072  *     "\xa1".force_encoding("iso-8859-1"),
01073  *     "\xa1\xa1".force_encoding("euc-jp"))
01074  *   #=> nil
01075  *
01076  * If the objects are non-strings their encodings are compatible when they
01077  * have an encoding and:
01078  * * Either encoding is US-ASCII compatible
01079  * * One of the encodings is a 7-bit encoding
01080  *
01081  */
01082 static VALUE
01083 enc_compatible_p(VALUE klass, VALUE str1, VALUE str2)
01084 {
01085     rb_encoding *enc;
01086 
01087     if (!enc_capable(str1)) return Qnil;
01088     if (!enc_capable(str2)) return Qnil;
01089     enc = rb_enc_compatible(str1, str2);
01090     if (!enc) return Qnil;
01091     return rb_enc_from_encoding(enc);
01092 }
01093 
01094 /* :nodoc: */
01095 static VALUE
01096 enc_dump(int argc, VALUE *argv, VALUE self)
01097 {
01098     rb_scan_args(argc, argv, "01", 0);
01099     return enc_name(self);
01100 }
01101 
01102 /* :nodoc: */
01103 static VALUE
01104 enc_load(VALUE klass, VALUE str)
01105 {
01106     return enc_find(klass, str);
01107 }
01108 
01109 rb_encoding *
01110 rb_ascii8bit_encoding(void)
01111 {
01112     if (!enc_table.list) {
01113         rb_enc_init();
01114     }
01115     return enc_table.list[ENCINDEX_ASCII].enc;
01116 }
01117 
01118 int
01119 rb_ascii8bit_encindex(void)
01120 {
01121     return ENCINDEX_ASCII;
01122 }
01123 
01124 rb_encoding *
01125 rb_utf8_encoding(void)
01126 {
01127     if (!enc_table.list) {
01128         rb_enc_init();
01129     }
01130     return enc_table.list[ENCINDEX_UTF_8].enc;
01131 }
01132 
01133 int
01134 rb_utf8_encindex(void)
01135 {
01136     return ENCINDEX_UTF_8;
01137 }
01138 
01139 rb_encoding *
01140 rb_usascii_encoding(void)
01141 {
01142     if (!enc_table.list) {
01143         rb_enc_init();
01144     }
01145     return enc_table.list[ENCINDEX_US_ASCII].enc;
01146 }
01147 
01148 int
01149 rb_usascii_encindex(void)
01150 {
01151     return ENCINDEX_US_ASCII;
01152 }
01153 
01154 int
01155 rb_locale_encindex(void)
01156 {
01157     VALUE charmap = rb_locale_charmap(rb_cEncoding);
01158     int idx;
01159 
01160     if (NIL_P(charmap))
01161         idx = rb_usascii_encindex();
01162     else if ((idx = rb_enc_find_index(StringValueCStr(charmap))) < 0)
01163         idx = rb_ascii8bit_encindex();
01164 
01165     if (rb_enc_registered("locale") < 0) enc_alias_internal("locale", idx);
01166 
01167     return idx;
01168 }
01169 
01170 rb_encoding *
01171 rb_locale_encoding(void)
01172 {
01173     return rb_enc_from_index(rb_locale_encindex());
01174 }
01175 
01176 static int
01177 enc_set_filesystem_encoding(void)
01178 {
01179     int idx;
01180 #if defined NO_LOCALE_CHARMAP
01181     idx = rb_enc_to_index(rb_default_external_encoding());
01182 #elif defined _WIN32 || defined __CYGWIN__
01183     char cp[sizeof(int) * 8 / 3 + 4];
01184     snprintf(cp, sizeof cp, "CP%d", AreFileApisANSI() ? GetACP() : GetOEMCP());
01185     idx = rb_enc_find_index(cp);
01186     if (idx < 0) idx = rb_ascii8bit_encindex();
01187 #else
01188     idx = rb_enc_to_index(rb_default_external_encoding());
01189 #endif
01190 
01191     enc_alias_internal("filesystem", idx);
01192     return idx;
01193 }
01194 
01195 int
01196 rb_filesystem_encindex(void)
01197 {
01198     int idx = rb_enc_registered("filesystem");
01199     if (idx < 0)
01200         idx = rb_ascii8bit_encindex();
01201     return idx;
01202 }
01203 
01204 rb_encoding *
01205 rb_filesystem_encoding(void)
01206 {
01207     return rb_enc_from_index(rb_filesystem_encindex());
01208 }
01209 
01210 struct default_encoding {
01211     int index;                  /* -2 => not yet set, -1 => nil */
01212     rb_encoding *enc;
01213 };
01214 
01215 static struct default_encoding default_external = {0};
01216 
01217 static int
01218 enc_set_default_encoding(struct default_encoding *def, VALUE encoding, const char *name)
01219 {
01220     int overridden = FALSE;
01221 
01222     if (def->index != -2)
01223         /* Already set */
01224         overridden = TRUE;
01225 
01226     if (NIL_P(encoding)) {
01227         def->index = -1;
01228         def->enc = 0;
01229         st_insert(enc_table.names, (st_data_t)strdup(name),
01230                   (st_data_t)UNSPECIFIED_ENCODING);
01231     }
01232     else {
01233         def->index = rb_enc_to_index(rb_to_encoding(encoding));
01234         def->enc = 0;
01235         enc_alias_internal(name, def->index);
01236     }
01237 
01238     if (def == &default_external)
01239         enc_set_filesystem_encoding();
01240 
01241     return overridden;
01242 }
01243 
01244 rb_encoding *
01245 rb_default_external_encoding(void)
01246 {
01247     if (default_external.enc) return default_external.enc;
01248 
01249     if (default_external.index >= 0) {
01250         default_external.enc = rb_enc_from_index(default_external.index);
01251         return default_external.enc;
01252     }
01253     else {
01254         return rb_locale_encoding();
01255     }
01256 }
01257 
01258 VALUE
01259 rb_enc_default_external(void)
01260 {
01261     return rb_enc_from_encoding(rb_default_external_encoding());
01262 }
01263 
01264 /*
01265  * call-seq:
01266  *   Encoding.default_external -> enc
01267  *
01268  * Returns default external encoding.
01269  *
01270  * The default external encoding is used by default for strings created from
01271  * the following locations:
01272  *
01273  * * CSV
01274  * * File data read from disk
01275  * * SDBM
01276  * * StringIO
01277  * * Zlib::GzipReader
01278  * * Zlib::GzipWriter
01279  * * String#inspect
01280  * * Regexp#inspect
01281  *
01282  * While strings created from these locations will have this encoding, the
01283  * encoding may not be valid.  Be sure to check String#valid_encoding?.
01284  *
01285  * File data written to disk will be transcoded to the default external
01286  * encoding when written.
01287  *
01288  * The default external encoding is initialized by the locale or -E option.
01289  */
01290 static VALUE
01291 get_default_external(VALUE klass)
01292 {
01293     return rb_enc_default_external();
01294 }
01295 
01296 void
01297 rb_enc_set_default_external(VALUE encoding)
01298 {
01299     if (NIL_P(encoding)) {
01300         rb_raise(rb_eArgError, "default external can not be nil");
01301     }
01302     enc_set_default_encoding(&default_external, encoding,
01303                             "external");
01304 }
01305 
01306 /*
01307  * call-seq:
01308  *   Encoding.default_external = enc
01309  *
01310  * Sets default external encoding.  You should not set
01311  * Encoding::default_external in ruby code as strings created before changing
01312  * the value may have a different encoding from strings created after thevalue
01313  * was changed., instead you should use <tt>ruby -E</tt> to invoke ruby with
01314  * the correct default_external.
01315  *
01316  * See Encoding::default_external for information on how the default external
01317  * encoding is used.
01318  */
01319 static VALUE
01320 set_default_external(VALUE klass, VALUE encoding)
01321 {
01322     rb_warning("setting Encoding.default_external");
01323     rb_enc_set_default_external(encoding);
01324     return encoding;
01325 }
01326 
01327 static struct default_encoding default_internal = {-2};
01328 
01329 rb_encoding *
01330 rb_default_internal_encoding(void)
01331 {
01332     if (!default_internal.enc && default_internal.index >= 0) {
01333         default_internal.enc = rb_enc_from_index(default_internal.index);
01334     }
01335     return default_internal.enc; /* can be NULL */
01336 }
01337 
01338 VALUE
01339 rb_enc_default_internal(void)
01340 {
01341     /* Note: These functions cope with default_internal not being set */
01342     return rb_enc_from_encoding(rb_default_internal_encoding());
01343 }
01344 
01345 /*
01346  * call-seq:
01347  *   Encoding.default_internal -> enc
01348  *
01349  * Returns default internal encoding.  Strings will be transcoded to the
01350  * default internal encoding in the following places if the default internal
01351  * encoding is not nil:
01352  *
01353  * * CSV
01354  * * Etc.sysconfdir and Etc.systmpdir
01355  * * File data read from disk
01356  * * File names from Dir
01357  * * Integer#chr
01358  * * String#inspect and Regexp#inspect
01359  * * Strings returned from Curses
01360  * * Strings returned from Readline
01361  * * Strings returned from SDBM
01362  * * Time#zone
01363  * * Values from ENV
01364  * * Values in ARGV including $PROGRAM_NAME
01365  * * __FILE__
01366  *
01367  * Additionally String#encode and String#encode! use the default internal
01368  * encoding if no encoding is given.
01369  *
01370  * The locale encoding (__ENCODING__), not default_internal, is used as the
01371  * encoding of created strings.
01372  *
01373  * Encoding::default_internal is initialized by the source file's
01374  * internal_encoding or -E option.
01375  */
01376 static VALUE
01377 get_default_internal(VALUE klass)
01378 {
01379     return rb_enc_default_internal();
01380 }
01381 
01382 void
01383 rb_enc_set_default_internal(VALUE encoding)
01384 {
01385     enc_set_default_encoding(&default_internal, encoding,
01386                             "internal");
01387 }
01388 
01389 /*
01390  * call-seq:
01391  *   Encoding.default_internal = enc or nil
01392  *
01393  * Sets default internal encoding or removes default internal encoding when
01394  * passed nil.  You should not set Encoding::default_internal in ruby code as
01395  * strings created before changing the value may have a different encoding
01396  * from strings created after the change.  Instead you should use
01397  * <tt>ruby -E</tt> to invoke ruby with the correct default_internal.
01398  *
01399  * See Encoding::default_internal for information on how the default internal
01400  * encoding is used.
01401  */
01402 static VALUE
01403 set_default_internal(VALUE klass, VALUE encoding)
01404 {
01405     rb_warning("setting Encoding.default_internal");
01406     rb_enc_set_default_internal(encoding);
01407     return encoding;
01408 }
01409 
01410 /*
01411  * call-seq:
01412  *   Encoding.locale_charmap -> string
01413  *
01414  * Returns the locale charmap name.
01415  * It returns nil if no appropriate information.
01416  *
01417  *   Debian GNU/Linux
01418  *     LANG=C
01419  *       Encoding.locale_charmap  #=> "ANSI_X3.4-1968"
01420  *     LANG=ja_JP.EUC-JP
01421  *       Encoding.locale_charmap  #=> "EUC-JP"
01422  *
01423  *   SunOS 5
01424  *     LANG=C
01425  *       Encoding.locale_charmap  #=> "646"
01426  *     LANG=ja
01427  *       Encoding.locale_charmap  #=> "eucJP"
01428  *
01429  * The result is highly platform dependent.
01430  * So Encoding.find(Encoding.locale_charmap) may cause an error.
01431  * If you need some encoding object even for unknown locale,
01432  * Encoding.find("locale") can be used.
01433  *
01434  */
01435 VALUE
01436 rb_locale_charmap(VALUE klass)
01437 {
01438 #if defined NO_LOCALE_CHARMAP
01439     return rb_usascii_str_new2("ASCII-8BIT");
01440 #elif defined _WIN32 || defined __CYGWIN__
01441     const char *nl_langinfo_codeset(void);
01442     const char *codeset = nl_langinfo_codeset();
01443     char cp[sizeof(int) * 3 + 4];
01444     if (!codeset) {
01445         UINT codepage = GetConsoleCP();
01446         if(!codepage) codepage = GetACP();
01447         snprintf(cp, sizeof(cp), "CP%d", codepage);
01448         codeset = cp;
01449     }
01450     return rb_usascii_str_new2(codeset);
01451 #elif defined HAVE_LANGINFO_H
01452     char *codeset;
01453     codeset = nl_langinfo(CODESET);
01454     return rb_usascii_str_new2(codeset);
01455 #else
01456     return Qnil;
01457 #endif
01458 }
01459 
01460 static void
01461 set_encoding_const(const char *name, rb_encoding *enc)
01462 {
01463     VALUE encoding = rb_enc_from_encoding(enc);
01464     char *s = (char *)name;
01465     int haslower = 0, hasupper = 0, valid = 0;
01466 
01467     if (ISDIGIT(*s)) return;
01468     if (ISUPPER(*s)) {
01469         hasupper = 1;
01470         while (*++s && (ISALNUM(*s) || *s == '_')) {
01471             if (ISLOWER(*s)) haslower = 1;
01472         }
01473     }
01474     if (!*s) {
01475         if (s - name > ENCODING_NAMELEN_MAX) return;
01476         valid = 1;
01477         rb_define_const(rb_cEncoding, name, encoding);
01478     }
01479     if (!valid || haslower) {
01480         size_t len = s - name;
01481         if (len > ENCODING_NAMELEN_MAX) return;
01482         if (!haslower || !hasupper) {
01483             do {
01484                 if (ISLOWER(*s)) haslower = 1;
01485                 if (ISUPPER(*s)) hasupper = 1;
01486             } while (*++s && (!haslower || !hasupper));
01487             len = s - name;
01488         }
01489         len += strlen(s);
01490         if (len++ > ENCODING_NAMELEN_MAX) return;
01491         MEMCPY(s = ALLOCA_N(char, len), name, char, len);
01492         name = s;
01493         if (!valid) {
01494             if (ISLOWER(*s)) *s = ONIGENC_ASCII_CODE_TO_UPPER_CASE((int)*s);
01495             for (; *s; ++s) {
01496                 if (!ISALNUM(*s)) *s = '_';
01497             }
01498             if (hasupper) {
01499                 rb_define_const(rb_cEncoding, name, encoding);
01500             }
01501         }
01502         if (haslower) {
01503             for (s = (char *)name; *s; ++s) {
01504                 if (ISLOWER(*s)) *s = ONIGENC_ASCII_CODE_TO_UPPER_CASE((int)*s);
01505             }
01506             rb_define_const(rb_cEncoding, name, encoding);
01507         }
01508     }
01509 }
01510 
01511 static int
01512 rb_enc_name_list_i(st_data_t name, st_data_t idx, st_data_t arg)
01513 {
01514     VALUE ary = (VALUE)arg;
01515     VALUE str = rb_usascii_str_new2((char *)name);
01516     OBJ_FREEZE(str);
01517     rb_ary_push(ary, str);
01518     return ST_CONTINUE;
01519 }
01520 
01521 /*
01522  * call-seq:
01523  *   Encoding.name_list -> ["enc1", "enc2", ...]
01524  *
01525  * Returns the list of available encoding names.
01526  *
01527  *   Encoding.name_list
01528  *   #=> ["US-ASCII", "ASCII-8BIT", "UTF-8",
01529  *         "ISO-8859-1", "Shift_JIS", "EUC-JP",
01530  *         "Windows-31J",
01531  *         "BINARY", "CP932", "eucJP"]
01532  *
01533  */
01534 
01535 static VALUE
01536 rb_enc_name_list(VALUE klass)
01537 {
01538     VALUE ary = rb_ary_new2(enc_table.names->num_entries);
01539     st_foreach(enc_table.names, rb_enc_name_list_i, (st_data_t)ary);
01540     return ary;
01541 }
01542 
01543 static int
01544 rb_enc_aliases_enc_i(st_data_t name, st_data_t orig, st_data_t arg)
01545 {
01546     VALUE *p = (VALUE *)arg;
01547     VALUE aliases = p[0], ary = p[1];
01548     int idx = (int)orig;
01549     VALUE key, str = rb_ary_entry(ary, idx);
01550 
01551     if (NIL_P(str)) {
01552         rb_encoding *enc = rb_enc_from_index(idx);
01553 
01554         if (!enc) return ST_CONTINUE;
01555         if (STRCASECMP((char*)name, rb_enc_name(enc)) == 0) {
01556             return ST_CONTINUE;
01557         }
01558         str = rb_usascii_str_new2(rb_enc_name(enc));
01559         OBJ_FREEZE(str);
01560         rb_ary_store(ary, idx, str);
01561     }
01562     key = rb_usascii_str_new2((char *)name);
01563     OBJ_FREEZE(key);
01564     rb_hash_aset(aliases, key, str);
01565     return ST_CONTINUE;
01566 }
01567 
01568 /*
01569  * call-seq:
01570  *   Encoding.aliases -> {"alias1" => "orig1", "alias2" => "orig2", ...}
01571  *
01572  * Returns the hash of available encoding alias and original encoding name.
01573  *
01574  *   Encoding.aliases
01575  *   #=> {"BINARY"=>"ASCII-8BIT", "ASCII"=>"US-ASCII", "ANSI_X3.4-1986"=>"US-ASCII",
01576  *         "SJIS"=>"Shift_JIS", "eucJP"=>"EUC-JP", "CP932"=>"Windows-31J"}
01577  *
01578  */
01579 
01580 static VALUE
01581 rb_enc_aliases(VALUE klass)
01582 {
01583     VALUE aliases[2];
01584     aliases[0] = rb_hash_new();
01585     aliases[1] = rb_ary_new();
01586     st_foreach(enc_table.names, rb_enc_aliases_enc_i, (st_data_t)aliases);
01587     return aliases[0];
01588 }
01589 
01590 void
01591 Init_Encoding(void)
01592 {
01593 #undef rb_intern
01594 #define rb_intern(str) rb_intern_const(str)
01595     VALUE list;
01596     int i;
01597 
01598     rb_cEncoding = rb_define_class("Encoding", rb_cObject);
01599     rb_undef_alloc_func(rb_cEncoding);
01600     rb_undef_method(CLASS_OF(rb_cEncoding), "new");
01601     rb_define_method(rb_cEncoding, "to_s", enc_name, 0);
01602     rb_define_method(rb_cEncoding, "inspect", enc_inspect, 0);
01603     rb_define_method(rb_cEncoding, "name", enc_name, 0);
01604     rb_define_method(rb_cEncoding, "names", enc_names, 0);
01605     rb_define_method(rb_cEncoding, "dummy?", enc_dummy_p, 0);
01606     rb_define_method(rb_cEncoding, "ascii_compatible?", enc_ascii_compatible_p, 0);
01607     rb_define_method(rb_cEncoding, "replicate", enc_replicate, 1);
01608     rb_define_singleton_method(rb_cEncoding, "list", enc_list, 0);
01609     rb_define_singleton_method(rb_cEncoding, "name_list", rb_enc_name_list, 0);
01610     rb_define_singleton_method(rb_cEncoding, "aliases", rb_enc_aliases, 0);
01611     rb_define_singleton_method(rb_cEncoding, "find", enc_find, 1);
01612     rb_define_singleton_method(rb_cEncoding, "compatible?", enc_compatible_p, 2);
01613 
01614     rb_define_method(rb_cEncoding, "_dump", enc_dump, -1);
01615     rb_define_singleton_method(rb_cEncoding, "_load", enc_load, 1);
01616 
01617     rb_define_singleton_method(rb_cEncoding, "default_external", get_default_external, 0);
01618     rb_define_singleton_method(rb_cEncoding, "default_external=", set_default_external, 1);
01619     rb_define_singleton_method(rb_cEncoding, "default_internal", get_default_internal, 0);
01620     rb_define_singleton_method(rb_cEncoding, "default_internal=", set_default_internal, 1);
01621     rb_define_singleton_method(rb_cEncoding, "locale_charmap", rb_locale_charmap, 0);
01622 
01623     list = rb_ary_new2(enc_table.count);
01624     RBASIC(list)->klass = 0;
01625     rb_encoding_list = list;
01626     rb_gc_register_mark_object(list);
01627 
01628     for (i = 0; i < enc_table.count; ++i) {
01629         rb_ary_push(list, enc_new(enc_table.list[i].enc));
01630     }
01631 }
01632 
01633 /* locale insensitive ctype functions */
01634 
01635 #define ctype_test(c, ctype) \
01636     (rb_isascii(c) && ONIGENC_IS_ASCII_CODE_CTYPE((c), (ctype)))
01637 
01638 int rb_isalnum(int c) { return ctype_test(c, ONIGENC_CTYPE_ALNUM); }
01639 int rb_isalpha(int c) { return ctype_test(c, ONIGENC_CTYPE_ALPHA); }
01640 int rb_isblank(int c) { return ctype_test(c, ONIGENC_CTYPE_BLANK); }
01641 int rb_iscntrl(int c) { return ctype_test(c, ONIGENC_CTYPE_CNTRL); }
01642 int rb_isdigit(int c) { return ctype_test(c, ONIGENC_CTYPE_DIGIT); }
01643 int rb_isgraph(int c) { return ctype_test(c, ONIGENC_CTYPE_GRAPH); }
01644 int rb_islower(int c) { return ctype_test(c, ONIGENC_CTYPE_LOWER); }
01645 int rb_isprint(int c) { return ctype_test(c, ONIGENC_CTYPE_PRINT); }
01646 int rb_ispunct(int c) { return ctype_test(c, ONIGENC_CTYPE_PUNCT); }
01647 int rb_isspace(int c) { return ctype_test(c, ONIGENC_CTYPE_SPACE); }
01648 int rb_isupper(int c) { return ctype_test(c, ONIGENC_CTYPE_UPPER); }
01649 int rb_isxdigit(int c) { return ctype_test(c, ONIGENC_CTYPE_XDIGIT); }
01650 
01651 int
01652 rb_tolower(int c)
01653 {
01654     return rb_isascii(c) ? ONIGENC_ASCII_CODE_TO_LOWER_CASE(c) : c;
01655 }
01656 
01657 int
01658 rb_toupper(int c)
01659 {
01660     return rb_isascii(c) ? ONIGENC_ASCII_CODE_TO_UPPER_CASE(c) : c;
01661 }
01662 
01663