From 0004565a919fde84def202432ef04cee0f7b689e Mon Sep 17 00:00:00 2001
From: Kevin Newton <kddnewton@gmail.com>
Date: Thu, 3 Aug 2023 13:25:38 -0400
Subject: [PATCH] [ruby/yarp] Use yp_memchr in regexp parsing

https://github.com/ruby/yarp/commit/08081dd24f
---
 yarp/extension.c |  2 +-
 yarp/regexp.c    | 15 ++++++++++-----
 yarp/regexp.h    |  4 +++-
 yarp/yarp.c      |  2 +-
 4 files changed, 15 insertions(+), 8 deletions(-)

diff --git a/yarp/extension.c b/yarp/extension.c
index d4ce3625d8..4f2065a81b 100644
--- a/yarp/extension.c
+++ b/yarp/extension.c
@@ -379,7 +379,7 @@ named_captures(VALUE self, VALUE source) {
     yp_string_list_t string_list;
     yp_string_list_init(&string_list);
 
-    if (!yp_regexp_named_capture_group_names(RSTRING_PTR(source), RSTRING_LEN(source), &string_list)) {
+    if (!yp_regexp_named_capture_group_names(RSTRING_PTR(source), RSTRING_LEN(source), &string_list, false, &yp_encoding_utf_8)) {
         yp_string_list_free(&string_list);
         return Qnil;
     }
diff --git a/yarp/regexp.c b/yarp/regexp.c
index 4855859442..40d85c5227 100644
--- a/yarp/regexp.c
+++ b/yarp/regexp.c
@@ -6,16 +6,20 @@ typedef struct {
     const char *cursor;
     const char *end;
     yp_string_list_t *named_captures;
+    bool encoding_changed;
+    yp_encoding_t *encoding;
 } yp_regexp_parser_t;
 
 // This initializes a new parser with the given source.
 static void
-yp_regexp_parser_init(yp_regexp_parser_t *parser, const char *start, const char *end, yp_string_list_t *named_captures) {
+yp_regexp_parser_init(yp_regexp_parser_t *parser, const char *start, const char *end, yp_string_list_t *named_captures, bool encoding_changed, yp_encoding_t *encoding) {
     *parser = (yp_regexp_parser_t) {
         .start = start,
         .cursor = start,
         .end = end,
-        .named_captures = named_captures
+        .named_captures = named_captures,
+        .encoding_changed = encoding_changed,
+        .encoding = encoding
     };
 }
 
@@ -60,7 +64,8 @@ yp_regexp_char_find(yp_regexp_parser_t *parser, char value) {
     if (yp_regexp_char_is_eof(parser)) {
         return false;
     }
-    const char *end = (const char *) memchr(parser->cursor, value, (size_t) (parser->end - parser->cursor));
+
+    const char *end = (const char *) yp_memchr(parser->cursor, value, (size_t) (parser->end - parser->cursor), parser->encoding_changed, parser->encoding);
     if (end == NULL) {
         return false;
     }
@@ -542,8 +547,8 @@ yp_regexp_parse_pattern(yp_regexp_parser_t *parser) {
 // Parse a regular expression and extract the names of all of the named capture
 // groups.
 YP_EXPORTED_FUNCTION bool
-yp_regexp_named_capture_group_names(const char *source, size_t size, yp_string_list_t *named_captures) {
+yp_regexp_named_capture_group_names(const char *source, size_t size, yp_string_list_t *named_captures, bool encoding_changed, yp_encoding_t *encoding) {
     yp_regexp_parser_t parser;
-    yp_regexp_parser_init(&parser, source, source + size, named_captures);
+    yp_regexp_parser_init(&parser, source, source + size, named_captures, encoding_changed, encoding);
     return yp_regexp_parse_pattern(&parser);
 }
diff --git a/yarp/regexp.h b/yarp/regexp.h
index cf624db6b8..5a2f13047e 100644
--- a/yarp/regexp.h
+++ b/yarp/regexp.h
@@ -3,6 +3,8 @@
 
 #include "yarp/defines.h"
 #include "yarp/parser.h"
+#include "yarp/enc/yp_encoding.h"
+#include "yarp/util/yp_memchr.h"
 #include "yarp/util/yp_string_list.h"
 #include "yarp/util/yp_string.h"
 
@@ -12,6 +14,6 @@
 
 // Parse a regular expression and extract the names of all of the named capture
 // groups.
-YP_EXPORTED_FUNCTION bool yp_regexp_named_capture_group_names(const char *source, size_t size, yp_string_list_t *named_captures);
+YP_EXPORTED_FUNCTION bool yp_regexp_named_capture_group_names(const char *source, size_t size, yp_string_list_t *named_captures, bool encoding_changed, yp_encoding_t *encoding);
 
 #endif
diff --git a/yarp/yarp.c b/yarp/yarp.c
index dd27c172f4..a8ff6c3859 100644
--- a/yarp/yarp.c
+++ b/yarp/yarp.c
@@ -12535,7 +12535,7 @@ parse_expression_infix(yp_parser_t *parser, yp_node_t *node, yp_binding_power_t
 
                 yp_location_t *content_loc = &((yp_regular_expression_node_t *) node)->content_loc;
 
-                if (yp_regexp_named_capture_group_names(content_loc->start, (size_t) (content_loc->end - content_loc->start), &named_captures)) {
+                if (yp_regexp_named_capture_group_names(content_loc->start, (size_t) (content_loc->end - content_loc->start), &named_captures, parser->encoding_changed, &parser->encoding)) {
                     for (size_t index = 0; index < named_captures.length; index++) {
                         yp_string_t *name = &named_captures.strings[index];
                         assert(name->type == YP_STRING_SHARED);