summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--regparse.c137
1 files changed, 59 insertions, 78 deletions
diff --git a/regparse.c b/regparse.c
index 48e5f511e3..32650e5c3e 100644
--- a/regparse.c
+++ b/regparse.c
@@ -5988,26 +5988,23 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
/* Unicode 10.0.0 */
/* Glue_After_Zwj */
- R_ERR(quantify_property_node(&np1, env, "Grapheme_Cluster_Break=Extend", '*'));
+ {
+ Node* seq[3];
- tmp = node_new_list(np1, NULL_NODE);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ seq[0] = node_new_cclass();
+ if (IS_NULL(seq[0])) goto err;
+ cc = NCCLASS(seq[0]);
+ {
+ const OnigCodePoint *ranges = onigenc_unicode_GCB_ranges_GAZ;
+ R_ERR(add_ctype_to_cc_by_range(cc, -1, 0, env, sb_out, ranges));
+ }
+ R_ERR(add_property_to_cc(cc, "Grapheme_Cluster_Break=Glue_After_Zwj", 0, env));
- np1 = node_new_cclass();
- if (IS_NULL(np1)) goto err;
- cc = NCCLASS(np1);
- {
- const OnigCodePoint *ranges = onigenc_unicode_GCB_ranges_GAZ;
- R_ERR(add_ctype_to_cc_by_range(cc, -1, 0, env, sb_out, ranges));
- }
- R_ERR(add_property_to_cc(cc, "Grapheme_Cluster_Break=Glue_After_Zwj", 0, env));
+ R_ERR(quantify_property_node(seq+1, env, "Grapheme_Cluster_Break=Extend", '*'));
- tmp = node_new_list(np1, list2);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ seq[2] = NULL_NODE;
+ R_ERR(create_sequence_node(&list2, seq));
+ }
tmp = onig_node_new_alt(list2, alt2);
if (IS_NULL(tmp)) goto err;
@@ -6018,31 +6015,27 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
/* Emoji variation sequence
* http://unicode.org/Public/emoji/4.0/emoji-zwj-sequences.txt
*/
- r = ONIGENC_CODE_TO_MBC(env->enc, 0xfe0f, buf);
- if (r < 0) goto err;
- np1 = node_new_str_raw(buf, buf + r);
- if (IS_NULL(np1)) goto err;
+ {
+ Node* seq[3];
- R_ERR(quantify_node(&np1, 0, 1));
+ seq[0] = node_new_cclass();
+ if (IS_NULL(seq[0])) goto err;
+ cc = NCCLASS(seq[0]);
+ {
+ const OnigCodePoint *ranges = onigenc_unicode_GCB_ranges_Emoji;
+ R_ERR(add_ctype_to_cc_by_range(cc, -1, 0, env, sb_out, ranges));
+ }
- tmp = node_new_list(np1, NULL_NODE);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ r = ONIGENC_CODE_TO_MBC(env->enc, 0xfe0f, buf); /* VARIATION SELECTOR-16 */
+ if (r < 0) goto err;
+ seq[1] = node_new_str_raw(buf, buf + r);
+ if (IS_NULL(seq[1])) goto err;
+ R_ERR(quantify_node(seq+1, 0, 1));
- np1 = node_new_cclass();
- if (IS_NULL(np1)) goto err;
- cc = NCCLASS(np1);
- {
- const OnigCodePoint *ranges = onigenc_unicode_GCB_ranges_Emoji;
- R_ERR(add_ctype_to_cc_by_range(cc, -1, 0, env, sb_out, ranges));
+ seq[2] = NULL_NODE;
+ R_ERR(create_sequence_node(&list2, seq));
}
- 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;
@@ -6054,7 +6047,7 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
alt2 = NULL;
/* ZWJ */
- r = ONIGENC_CODE_TO_MBC(env->enc, 0x200D, buf);
+ r = ONIGENC_CODE_TO_MBC(env->enc, 0x200D, buf); /* ZERO WIDTH JOINER (ZWJ) */
if (r < 0) goto err;
np1 = node_new_str_raw(buf, buf + r);
if (IS_NULL(np1)) goto err;
@@ -6111,38 +6104,30 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
list2 = NULL;
/* Unicode 10.0.0 */
- /* 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
*/
- R_ERR(quantify_property_node(&np1, env, "Grapheme_Cluster_Break=E_Modifier", '?'));
-
- tmp = node_new_list(np1, NULL_NODE);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ /* ZWJ (E_Base_GAZ | Glue_After_Zwj) E_Modifier? */
+ {
+ Node* seq[4];
- np1 = node_new_cclass();
- if (IS_NULL(np1)) goto err;
- cc = NCCLASS(np1);
- R_ERR(add_property_to_cc(cc, "Grapheme_Cluster_Break=Glue_After_Zwj", 0, env));
- R_ERR(add_property_to_cc(cc, "Grapheme_Cluster_Break=E_Base_GAZ", 0, env));
+ r = ONIGENC_CODE_TO_MBC(env->enc, 0x200D, buf); /* ZERO WIDTH JOINER (ZWJ) */
+ if (r < 0) goto err;
+ seq[0] = node_new_str_raw(buf, buf + r);
+ if (IS_NULL(seq[0])) goto err;
- tmp = node_new_list(np1, list2);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ seq[1] = node_new_cclass();
+ if (IS_NULL(seq[1])) goto err;
+ cc = NCCLASS(seq[1]);
+ R_ERR(add_property_to_cc(cc, "Grapheme_Cluster_Break=Glue_After_Zwj", 0, env));
+ R_ERR(add_property_to_cc(cc, "Grapheme_Cluster_Break=E_Base_GAZ", 0, env));
- r = ONIGENC_CODE_TO_MBC(env->enc, 0x200D, buf);
- if (r < 0) goto err;
- np1 = node_new_str_raw(buf, buf + r);
- if (IS_NULL(np1)) goto err;
+ R_ERR(quantify_property_node(seq+2, env, "Grapheme_Cluster_Break=E_Modifier", '?'));
- tmp = node_new_list(np1, list2);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ seq[3] = NULL_NODE;
+ R_ERR(create_sequence_node(&list2, seq));
+ }
tmp = onig_node_new_alt(list2, alt);
if (IS_NULL(tmp)) goto err;
@@ -6155,7 +6140,7 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
/* RI-Sequence := Regional_Indicator{2} */
R_ERR(quantify_property_node(&np1, env, "Regional_Indicator", '2'));
- tmp = node_new_list(np1, list2);
+ tmp = node_new_list(np1, list2); /* here, list2 should be guaranteed to be NULL */
if (IS_NULL(tmp)) goto err;
list2 = tmp;
np1 = NULL;
@@ -6195,24 +6180,20 @@ node_extended_grapheme_cluster(Node** np, ScanEnv* env)
np1 = NULL;
/* Prepend+ */
- r = ONIGENC_CODE_TO_MBC(env->enc, 0x200D, buf);
- if (r < 0) goto err;
- np1 = node_new_str_raw(buf, buf + r);
- if (IS_NULL(np1)) goto err;
-
- R_ERR(quantify_node(&np1, 0, 1));
+ {
+ Node* seq[3];
- tmp = node_new_list(np1, NULL_NODE);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ R_ERR(quantify_property_node(seq+0, env, "Grapheme_Cluster_Break=Prepend", '+'));
- R_ERR(quantify_property_node(&np1, env, "Grapheme_Cluster_Break=Prepend", '+'));
+ r = ONIGENC_CODE_TO_MBC(env->enc, 0x200D, buf); /* does this belong to Prepend?? */
+ if (r < 0) goto err;
+ seq[1] = node_new_str_raw(buf, buf + r);
+ if (IS_NULL(seq[1])) goto err;
+ R_ERR(quantify_node(seq+1, 0, 1));
- tmp = node_new_list(np1, list2);
- if (IS_NULL(tmp)) goto err;
- list2 = tmp;
- np1 = NULL;
+ seq[2] = NULL_NODE;
+ R_ERR(create_sequence_node(&list2, seq));
+ }
tmp = onig_node_new_alt(list2, alt);
if (IS_NULL(tmp)) goto err;
generated by cgit v1.2.3 (git 2.25.1) at 2025-05-14 08:12:57 +0000