зеркало из https://github.com/github/ruby.git
Regexp supports Unicoe 9.0.0's \X
* meta character \X matches Unicode 9.0.0 characters with some workarounds for UTR #51 Unicode Emoji, Version 4.0 emoji zwj sequences. [Feature #12831] [ruby-core:77586] The term "character" can have many meanings bytes, codepoints, combined characters, and so on. "grapheme cluster" is highest one of such words, which means user-perceived characters. Unicode Standard Annex #29 UNICODE TEXT SEGMENTATION specifies how to handle grapheme clusters (extended grapheme cluster). But some specs aren't updated to current situation because Unicode Emoji is rapidly extended without well definition. It breaks the precondition of UTR#29 "Grapheme cluster boundaries can be easily tested by looking at immediately adjacent characters". (the sentence will be removed in the next version) Though some of its detail are described in Unicode Technical Report #51 UNICODE EMOJI but it is not merged into UTR#29 yet. http://unicode.org/reports/tr29/ http://unicode.org/reports/tr51/ http://unicode.org/Public/emoji/4.0/ git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@56949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This commit is contained in:
naruse
Родитель
e680bfb1ab
Коммит
c11e648799
3
NEWS
3
NEWS
|
|
@ -132,6 +132,9 @@ with all sufficient information, see the ChangeLog file or Redmine
|
|||
* Regexp#match? [Feature #8110]
|
||||
This returns bool and doesn't save backref.
|
||||
|
||||
* meta character \X matches Unicode 9.0 characters with some workarounds
|
||||
for UTR #51 Unicode Emoji, Version 4.0 emoji zwj sequences.
|
||||
|
||||
* Regexp/String: Updated Unicode version from 8.0.0 to 9.0.0 [Feature #12513]
|
||||
|
||||
* RubyVM::Env
|
||||
|
|
|
|||
|
|
@ -1064,6 +1064,7 @@ UNICODE_PROPERTY_FILES = \
|
|||
$(UNICODE_SRC_DATA_DIR)/PropertyAliases.txt \
|
||||
$(UNICODE_SRC_DATA_DIR)/PropertyValueAliases.txt \
|
||||
$(UNICODE_SRC_DATA_DIR)/Scripts.txt \
|
||||
$(UNICODE_SRC_DATA_DIR)/auxiliary/GraphemeBreakProperty.txt \
|
||||
$(empty)
|
||||
|
||||
update-unicode: $(UNICODE_FILES)
|
||||
|
|
@ -1076,7 +1077,7 @@ UNICODE_DOWNLOAD = \
|
|||
|
||||
$(UNICODE_PROPERTY_FILES):
|
||||
$(ECHO) Downloading Unicode $(UNICODE_VERSION) property files...
|
||||
$(Q) $(MAKEDIRS) "$(UNICODE_SRC_DATA_DIR)"
|
||||
$(Q) $(MAKEDIRS) "$(UNICODE_SRC_DATA_DIR)/auxiliary"
|
||||
$(Q) $(UNICODE_DOWNLOAD) $(UNICODE_PROPERTY_FILES)
|
||||
|
||||
$(UNICODE_FILES):
|
||||
|
|
|
|||
Разница между файлами не показана из-за своего большого размера
Загрузить разницу
13
regcomp.c
13
regcomp.c
|
|
@ -6572,7 +6572,7 @@ print_compiled_byte_code_list(FILE* f, regex_t* reg)
|
|||
#endif /* ONIG_DEBUG_COMPILE */
|
||||
|
||||
#ifdef ONIG_DEBUG_PARSE_TREE
|
||||
static void
|
||||
void
|
||||
print_indent_tree(FILE* f, Node* node, int indent)
|
||||
{
|
||||
int i, type, container_p = 0;
|
||||
|
|
@ -6618,12 +6618,15 @@ print_indent_tree(FILE* f, Node* node, int indent)
|
|||
|
||||
case NT_CCLASS:
|
||||
fprintf(f, "<cclass:%"PRIxPTR">", (intptr_t )node);
|
||||
if (IS_NCCLASS_NOT(NCCLASS(node))) fputs(" not", f);
|
||||
if (IS_NCCLASS_NOT(NCCLASS(node))) fputs("not ", f);
|
||||
if (NCCLASS(node)->mbuf) {
|
||||
BBuf* bbuf = NCCLASS(node)->mbuf;
|
||||
for (i = 0; i < (int )bbuf->used; i++) {
|
||||
if (i > 0) fprintf(f, ",");
|
||||
fprintf(f, "%0x", bbuf->p[i]);
|
||||
OnigCodePoint* data = (OnigCodePoint*)bbuf->p;
|
||||
OnigCodePoint* end = (OnigCodePoint*)(bbuf->p + bbuf->used);
|
||||
fprintf(f, "%d", *data++);
|
||||
for (; data < end; data+=2) {
|
||||
fprintf(f, ",");
|
||||
fprintf(f, "%04x-%04x", data[0], data[1]);
|
||||
}
|
||||
}
|
||||
break;
|
||||
|
|
|
|||
730
regparse.c
730
regparse.c
|
|
@ -5784,61 +5784,703 @@ node_linebreak(Node** np, ScanEnv* env)
|
|||
return ONIGERR_MEMORY;
|
||||
}
|
||||
|
||||
static int
|
||||
propname2ctype(ScanEnv* env, const char* propname)
|
||||
{
|
||||
UChar* name = (UChar*)propname;
|
||||
int ctype = env->enc->property_name_to_ctype(ONIG_ENCODING_ASCII,
|
||||
name, name + strlen(propname));
|
||||
return ctype;
|
||||
}
|
||||
|
||||
static int
|
||||
node_extended_grapheme_cluster(Node** np, ScanEnv* env)
|
||||
{
|
||||
/* same as (?>\P{M}\p{M}*) */
|
||||
Node* np1 = NULL;
|
||||
Node* np2 = NULL;
|
||||
Node* qn = NULL;
|
||||
Node* list1 = NULL;
|
||||
Node* list = NULL;
|
||||
Node* list2 = NULL;
|
||||
Node* alt = NULL;
|
||||
Node* alt2 = NULL;
|
||||
int r = 0;
|
||||
|
||||
#ifdef USE_UNICODE_PROPERTIES
|
||||
if (ONIGENC_IS_UNICODE(env->enc)) {
|
||||
/* UTF-8, UTF-16BE/LE, UTF-32BE/LE */
|
||||
CClassNode* cc1;
|
||||
CClassNode* cc2;
|
||||
UChar* propname = (UChar* )"M";
|
||||
int ctype = env->enc->property_name_to_ctype(ONIG_ENCODING_ASCII,
|
||||
propname, propname + 1);
|
||||
if (ctype >= 0) {
|
||||
/* \P{M} */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc1 = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc1, ctype, 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
NCCLASS_SET_NOT(cc1);
|
||||
Node* tmp = NULL;
|
||||
int num1, num2;
|
||||
UChar buf[ONIGENC_CODE_TO_MBC_MAXLEN * 2];
|
||||
CClassNode* cc;
|
||||
OnigOptionType option;
|
||||
int extend = propname2ctype(env, "Grapheme_Cluster_Break=Extend");
|
||||
|
||||
/* \p{M}* */
|
||||
np2 = node_new_cclass();
|
||||
if (IS_NULL(np2)) goto err;
|
||||
cc2 = NCCLASS(np2);
|
||||
r = add_ctype_to_cc(cc2, ctype, 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
/* Prepend*
|
||||
* ( RI-sequence | Hangul-Syllable | !Control )
|
||||
* ( Grapheme_Extend | SpacingMark )* */
|
||||
|
||||
qn = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(qn)) goto err;
|
||||
NQTFR(qn)->target = np2;
|
||||
np2 = NULL;
|
||||
/* ( Grapheme_Extend | SpacingMark )* */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, extend, 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=SpacingMark"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x200D, 0x200D);
|
||||
if (r != 0) goto err;
|
||||
|
||||
/* \P{M}\p{M}* */
|
||||
list2 = node_new_list(qn, NULL_NODE);
|
||||
if (IS_NULL(list2)) goto err;
|
||||
qn = NULL;
|
||||
list1 = node_new_list(np1, list2);
|
||||
if (IS_NULL(list1)) goto err;
|
||||
np1 = NULL;
|
||||
list2 = NULL;
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
/* (?>...) */
|
||||
*np = node_new_enclose(ENCLOSE_STOP_BACKTRACK);
|
||||
if (IS_NULL(*np)) goto err;
|
||||
NENCLOSE(*np)->target = list1;
|
||||
return ONIG_NORMAL;
|
||||
}
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* ( RI-sequence | Hangul-Syllable | !Control ) */
|
||||
/* !Control */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=Control"), 1, 0, env);
|
||||
if (r != 0) goto err;
|
||||
BITSET_CLEAR_BIT(cc->bs, 0x0a);
|
||||
BITSET_CLEAR_BIT(cc->bs, 0x0d);
|
||||
|
||||
tmp = onig_node_new_alt(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* Hangul-Syllable
|
||||
* := L* V+ T*
|
||||
* | L* LV V* T*
|
||||
* | L* LVT T*
|
||||
* | L+
|
||||
* | T+ */
|
||||
|
||||
/* T+ */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=T"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(1, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = onig_node_new_alt(np1, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* L+ */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=L"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(1, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = onig_node_new_alt(np1, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* L* LVT T* */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=T"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=LVT"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=L"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* L* LV V* T* */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=T"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=V"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=LV"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=L"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* L* V+ T* */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=T"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=V"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(1, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=L"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* Emoji sequence := (E_Base | EBG) Extend* E_Modifier?
|
||||
* (ZWJ (Glue_After_Zwj | EBG Extend* E_Modifier?) )* */
|
||||
|
||||
/* ZWJ (Glue_After_Zwj | E_Base_GAZ Extend* E_Modifier?) */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Modifier"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, 1, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, extend, 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Base_GAZ"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt2 = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* Glue_After_Zwj */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, extend, 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F308, 0x1F308);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F33E, 0x1F33E);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F373, 0x1F373);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F393, 0x1F393);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3A4, 0x1F3A4);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3A8, 0x1F3A8);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3EB, 0x1F3EB);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3ED, 0x1F3ED);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F4BB, 0x1F4BC);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F527, 0x1F527);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F52C, 0x1F52C);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F680, 0x1F680);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F692, 0x1F692);
|
||||
if (r != 0) goto err;
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=Glue_After_Zwj"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt2 = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* Emoji variation sequence
|
||||
* http://unicode.org/Public/emoji/4.0/emoji-zwj-sequences.txt
|
||||
*/
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0xfe0f, 0xfe0f);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, 1, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x2640, 0x2640);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x2642, 0x2642);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x2695, 0x2696);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x2708, 0x2708);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt2 = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
tmp = node_new_list(alt2, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
alt2 = NULL;
|
||||
|
||||
/* ZWJ */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x200D, 0x200D);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = list2;
|
||||
np1 = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* E_Modifier? */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Modifier"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, 1, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* Extend* */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, extend, 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* (E_Base | EBG) */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3C2, 0x1F3C2);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3C7, 0x1F3C7);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3CC, 0x1F3CC);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F3F3, 0x1F3F3);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F441, 0x1F441);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F46F, 0x1F46F);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F574, 0x1F574);
|
||||
if (r != 0) goto err;
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F6CC, 0x1F6CC);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Base"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Base_GAZ"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* ZWJ (E_Base_GAZ | Glue_After_Zwj) E_Modifier? */
|
||||
/* a sequence starting with ZWJ seems artificial, but GraphemeBreakTest
|
||||
* has such examples.
|
||||
* http://www.unicode.org/Public/9.0.0/ucd/auxiliary/GraphemeBreakTest.html
|
||||
*/
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Modifier"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, 1, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=Glue_After_Zwj"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=E_Base_GAZ"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x200D, 0x200D);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
/* RI-Sequence := Regional_Indicator{2} */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x1F1E6, 0x1F1FF);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(2, 2, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
tmp = node_new_list(alt, list);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list = tmp;
|
||||
alt = NULL;
|
||||
|
||||
/* Prepend* */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=Prepend"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* PerlSyntax: (?s:.), RubySyntax: (?m:.) */
|
||||
np1 = node_new_anychar();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
|
||||
option = env->option;
|
||||
ONOFF(option, ONIG_OPTION_MULTILINE, 0);
|
||||
tmp = node_new_option(option);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NENCLOSE(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = onig_node_new_alt(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* Prepend+ */
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_code_range(&(cc->mbuf), env, 0x200D, 0x200D);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(0, 1, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, NULL_NODE);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
np1 = node_new_cclass();
|
||||
if (IS_NULL(np1)) goto err;
|
||||
cc = NCCLASS(np1);
|
||||
r = add_ctype_to_cc(cc, propname2ctype(env, "Grapheme_Cluster_Break=Prepend"), 0, 0, env);
|
||||
if (r != 0) goto err;
|
||||
|
||||
tmp = node_new_quantifier(1, REPEAT_INFINITE, 0);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
NQTFR(tmp)->target = np1;
|
||||
np1 = tmp;
|
||||
|
||||
tmp = node_new_list(np1, list2);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
list2 = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list2, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list2 = NULL;
|
||||
|
||||
tmp = onig_node_new_alt(list, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
list = NULL;
|
||||
|
||||
/* \x0D\x0A */
|
||||
num1 = ONIGENC_CODE_TO_MBC(env->enc, 0x0D, buf);
|
||||
if (num1 < 0) return num1;
|
||||
num2 = ONIGENC_CODE_TO_MBC(env->enc, 0x0A, buf + num1);
|
||||
if (num2 < 0) return num2;
|
||||
np1 = node_new_str_raw(buf, buf + num1 + num2);
|
||||
if (IS_NULL(np1)) goto err;
|
||||
|
||||
tmp = onig_node_new_alt(np1, alt);
|
||||
if (IS_NULL(tmp)) goto err;
|
||||
alt = tmp;
|
||||
np1 = NULL;
|
||||
|
||||
/* (?>...) */
|
||||
*np = node_new_enclose(ENCLOSE_STOP_BACKTRACK);
|
||||
if (IS_NULL(*np)) goto err;
|
||||
NENCLOSE(*np)->target = alt;
|
||||
return ONIG_NORMAL;
|
||||
}
|
||||
#endif /* USE_UNICODE_PROPERTIES */
|
||||
if (IS_NULL(*np)) {
|
||||
|
|
@ -5857,10 +6499,10 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
|
|||
|
||||
err:
|
||||
onig_node_free(np1);
|
||||
onig_node_free(np2);
|
||||
onig_node_free(qn);
|
||||
onig_node_free(list1);
|
||||
onig_node_free(list);
|
||||
onig_node_free(list2);
|
||||
onig_node_free(alt);
|
||||
onig_node_free(alt2);
|
||||
return (r == 0) ? ONIGERR_MEMORY : r;
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -900,6 +900,27 @@ class TestRegexp < Test::Unit::TestCase
|
|||
assert_no_match(/[[:ascii:]]/, "\x80\xFF")
|
||||
end
|
||||
|
||||
def test_cclass_R
|
||||
assert_match /\A\R\z/, "\r"
|
||||
assert_match /\A\R\z/, "\n"
|
||||
assert_match /\A\R\z/, "\r\n"
|
||||
end
|
||||
|
||||
def test_cclass_X
|
||||
assert_match /\A\X\z/, "\u{20 200d}"
|
||||
assert_match /\A\X\z/, "\u{600 600}"
|
||||
assert_match /\A\X\z/, "\u{600 20}"
|
||||
assert_match /\A\X\z/, "\u{261d 1F3FB}"
|
||||
assert_match /\A\X\z/, "\u{1f600}"
|
||||
assert_match /\A\X\z/, "\u{20 308}"
|
||||
assert_match /\A\X\X\z/, "\u{a 308}"
|
||||
assert_match /\A\X\X\z/, "\u{d 308}"
|
||||
assert_match /\A\X\z/, "\u{1F477 1F3FF 200D 2640 FE0F}"
|
||||
assert_match /\A\X\z/, "\u{1F468 200D 1F393}"
|
||||
assert_match /\A\X\z/, "\u{1F46F 200D 2642 FE0F}"
|
||||
assert_match /\A\X\z/, "\u{1f469 200d 2764 fe0f 200d 1f469}"
|
||||
end
|
||||
|
||||
def test_backward
|
||||
assert_equal(3, "foobar".rindex(/b.r/i))
|
||||
assert_equal(nil, "foovar".rindex(/b.r/i))
|
||||
|
|
|
|||
|
|
@ -228,7 +228,13 @@ if $0 == __FILE__
|
|||
dl = Downloader.const_get(dl)
|
||||
ARGV.shift
|
||||
ARGV.each do |name|
|
||||
name = "#{prefix}/#{File.basename(name)}" if prefix
|
||||
if prefix
|
||||
if name.include?('/auxiliary/')
|
||||
name = "#{prefix}/auxiliary/#{File.basename(name)}"
|
||||
else
|
||||
name = "#{prefix}/#{File.basename(name)}"
|
||||
end
|
||||
end
|
||||
dl.download(name, destdir, since, options)
|
||||
end
|
||||
else
|
||||
|
|
|
|||
|
|
@ -144,7 +144,7 @@ def parse_scripts(data, categories)
|
|||
categories[current] = file[:title]
|
||||
(names[file[:title]] ||= []) << current
|
||||
cps = []
|
||||
elsif /^([0-9a-fA-F]+)(?:..([0-9a-fA-F]+))?\s*;\s*(\w+)/ =~ line
|
||||
elsif /^([0-9a-fA-F]+)(?:\.\.([0-9a-fA-F]+))?\s*;\s*(\w+)/ =~ line
|
||||
current = $3
|
||||
$2 ? cps.concat(($1.to_i(16)..$2.to_i(16)).to_a) : cps.push($1.to_i(16))
|
||||
end
|
||||
|
|
@ -199,7 +199,26 @@ def parse_age(data)
|
|||
ages << current
|
||||
last_constname = constname
|
||||
cps = []
|
||||
elsif /^([0-9a-fA-F]+)(?:..([0-9a-fA-F]+))?\s*;\s*(\d+\.\d+)/ =~ line
|
||||
elsif /^([0-9a-fA-F]+)(?:\.\.([0-9a-fA-F]+))?\s*;\s*(\d+\.\d+)/ =~ line
|
||||
current = $3
|
||||
$2 ? cps.concat(($1.to_i(16)..$2.to_i(16)).to_a) : cps.push($1.to_i(16))
|
||||
end
|
||||
end
|
||||
ages
|
||||
end
|
||||
|
||||
def parse_GraphemeBreakProperty(data)
|
||||
current = nil
|
||||
cps = []
|
||||
ages = []
|
||||
data_foreach('GraphemeBreakProperty.txt') do |line|
|
||||
if /^# Total code points: / =~ line
|
||||
constname = constantize_Grapheme_Cluster_Break(current)
|
||||
data[constname] = cps
|
||||
make_const(constname, cps, "Grapheme_Cluster_Break=#{current}")
|
||||
ages << current
|
||||
cps = []
|
||||
elsif /^([0-9a-fA-F]+)(?:\.\.([0-9a-fA-F]+))?\s*;\s*(\w+)/ =~ line
|
||||
current = $3
|
||||
$2 ? cps.concat(($1.to_i(16)..$2.to_i(16)).to_a) : cps.push($1.to_i(16))
|
||||
end
|
||||
|
|
@ -209,7 +228,6 @@ end
|
|||
|
||||
def parse_block(data)
|
||||
current = nil
|
||||
last_constname = nil
|
||||
cps = []
|
||||
blocks = []
|
||||
data_foreach('Blocks.txt') do |line|
|
||||
|
|
@ -269,6 +287,10 @@ def constantize_agename(name)
|
|||
"Age_#{name.sub(/\./, '_')}"
|
||||
end
|
||||
|
||||
def constantize_Grapheme_Cluster_Break(name)
|
||||
"Grapheme_Cluster_Break_#{name}"
|
||||
end
|
||||
|
||||
def constantize_blockname(name)
|
||||
"In_#{name.gsub(/\W/, '_')}"
|
||||
end
|
||||
|
|
@ -381,6 +403,7 @@ end
|
|||
output.ifdef :USE_UNICODE_AGE_PROPERTIES
|
||||
ages = parse_age(data)
|
||||
output.endif :USE_UNICODE_AGE_PROPERTIES
|
||||
graphemeBreaks = parse_GraphemeBreakProperty(data)
|
||||
blocks = parse_block(data)
|
||||
output.endif :USE_UNICODE_PROPERTIES
|
||||
puts(<<'__HEREDOC')
|
||||
|
|
@ -393,6 +416,7 @@ props.each{|name| puts" CR_#{name},"}
|
|||
output.ifdef :USE_UNICODE_AGE_PROPERTIES
|
||||
ages.each{|name| puts" CR_#{constantize_agename(name)},"}
|
||||
output.endif :USE_UNICODE_AGE_PROPERTIES
|
||||
graphemeBreaks.each{|name| puts" CR_#{constantize_Grapheme_Cluster_Break(name)},"}
|
||||
blocks.each{|name|puts" CR_#{name},"}
|
||||
output.endif :USE_UNICODE_PROPERTIES
|
||||
|
||||
|
|
@ -437,6 +461,12 @@ ages.each do |name|
|
|||
puts "%-40s %3d" % [name + ',', i]
|
||||
end
|
||||
output.endif :USE_UNICODE_AGE_PROPERTIES
|
||||
graphemeBreaks.each do |name|
|
||||
i += 1
|
||||
name = "graphemeclusterbreak=#{name.delete('_').downcase}"
|
||||
name_to_index[name] = i
|
||||
puts "%-40s %3d" % [name + ',', i]
|
||||
end
|
||||
blocks.each do |name|
|
||||
i += 1
|
||||
name = normalize_propname(name)
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче