diff options
Diffstat (limited to 'libs/litehtml/src/gumbo')
| -rw-r--r-- | libs/litehtml/src/gumbo/CMakeLists.txt | 8 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/char_ref.c | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/char_ref.rl | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/error.c | 13 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo.h | 4 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/error.h | 6 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/tag_enum.h | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/tag_gperf.h | 418 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/tag_sizes.h | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/tag_strings.h | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/utf8.h | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/include/gumbo/util.h | 6 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/parser.c | 25 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/tag.c | 7 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/tag.in | 2 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/tokenizer.c | 9 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/utf8.c | 4 | ||||
| -rw-r--r-- | libs/litehtml/src/gumbo/vector.c | 2 |
18 files changed, 375 insertions, 141 deletions
diff --git a/libs/litehtml/src/gumbo/CMakeLists.txt b/libs/litehtml/src/gumbo/CMakeLists.txt index cd55186a15..7282604893 100644 --- a/libs/litehtml/src/gumbo/CMakeLists.txt +++ b/libs/litehtml/src/gumbo/CMakeLists.txt @@ -1,4 +1,4 @@ -cmake_minimum_required(VERSION 3.11) +cmake_minimum_required(VERSION 3.5) project(gumbo C) @@ -46,6 +46,10 @@ set(HEADER_GUMBO set(PROJECT_LIB_VERSION ${PROJECT_MAJOR}.${PROJECT_MINOR}.0) set(PROJECT_SO_VERSION ${PROJECT_MAJOR}) +if (MSVC) + add_compile_options(/wd4244 /wd4267) +endif() + add_library(${PROJECT_NAME} ${SOURCE_GUMBO}) set_target_properties(${PROJECT_NAME} PROPERTIES VERSION ${PROJECT_LIB_VERSION} SOVERSION ${PROJECT_SO_VERSION}) @@ -54,7 +58,7 @@ set_target_properties(${PROJECT_NAME} PROPERTIES PUBLIC_HEADER "${HEADER_GUMBO}" ) -if(MSVC) +if (MSVC) target_include_directories(${PROJECT_NAME} PRIVATE visualc/include) endif() diff --git a/libs/litehtml/src/gumbo/char_ref.c b/libs/litehtml/src/gumbo/char_ref.c index b3c5eccaf3..a1d74fd5df 100644 --- a/libs/litehtml/src/gumbo/char_ref.c +++ b/libs/litehtml/src/gumbo/char_ref.c @@ -23009,7 +23009,7 @@ _again: if (cs >= 7623) { assert(output->first != kGumboNoChar); char last_char = *(te - 1); - ptrdiff_t len = te - start; + int len = te - start; if (last_char == ';') { bool matched = utf8iterator_maybe_consume_match(input, start, len, true); assert(matched); diff --git a/libs/litehtml/src/gumbo/char_ref.rl b/libs/litehtml/src/gumbo/char_ref.rl index 7c1c410d93..139a4bbd33 100644 --- a/libs/litehtml/src/gumbo/char_ref.rl +++ b/libs/litehtml/src/gumbo/char_ref.rl @@ -2493,7 +2493,7 @@ static bool consume_named_ref( if (cs >= %%{ write first_final; }%%) { assert(output->first != kGumboNoChar); char last_char = *(te - 1); - ptrdiff_t len = te - start; + int len = te - start; if (last_char == ';') { bool matched = utf8iterator_maybe_consume_match(input, start, len, true); assert(matched); diff --git a/libs/litehtml/src/gumbo/error.c b/libs/litehtml/src/gumbo/error.c index 369e7c11e8..057e89ce29 100644 --- a/libs/litehtml/src/gumbo/error.c +++ b/libs/litehtml/src/gumbo/error.c @@ -33,7 +33,7 @@ static int print_message( GumboParser* parser, GumboStringBuffer* output, const char* format, ...) { va_list args; - size_t remaining_capacity = output->capacity - output->length; + int remaining_capacity = output->capacity - output->length; va_start(args, format); int bytes_written = vsnprintf( output->data + output->length, remaining_capacity, format, args); @@ -59,7 +59,7 @@ static int print_message( } #endif - if (bytes_written > remaining_capacity) { + if (bytes_written >= remaining_capacity) { gumbo_string_buffer_reserve( parser, output->capacity + bytes_written, output); remaining_capacity = output->capacity - output->length; @@ -79,7 +79,7 @@ static void print_tag_stack(GumboParser* parser, const GumboParserError* error, if (i) { print_message(parser, output, ", "); } - GumboTag tag = (GumboTag)(uintptr_t) error->tag_stack.data[i]; + GumboTag tag = (GumboTag) error->tag_stack.data[i]; print_message(parser, output, gumbo_normalized_tagname(tag)); } gumbo_string_buffer_append_codepoint(parser, '.', output); @@ -136,6 +136,13 @@ static const char* find_last_newline( const char* original_text, const char* error_location) { assert(error_location >= original_text); const char* c = error_location; + // If the error location itself is a newline then start searching for the + // preceding newline one character earlier, if possible. See: + // https://github.com/rubys/nokogumbo/commit/bd623555730cdd260f6cec6d7cf990ff297da63d + // https://github.com/google/gumbo-parser/pull/371 + if (*c == '\n' && c != original_text) { + c -= 1; + } for (; c != original_text && *c != '\n'; --c) { // There may be an error at EOF, which would be a nul byte. assert(*c || c == error_location); diff --git a/libs/litehtml/src/gumbo/include/gumbo.h b/libs/litehtml/src/gumbo/include/gumbo.h index 27e6c6c575..f8137cf061 100644 --- a/libs/litehtml/src/gumbo/include/gumbo.h +++ b/libs/litehtml/src/gumbo/include/gumbo.h @@ -43,13 +43,9 @@ #define GUMBO_GUMBO_H_ #ifdef _MSC_VER -#ifndef _CRT_SECURE_NO_WARNINGS #define _CRT_SECURE_NO_WARNINGS -#endif -#ifndef fileno #define fileno _fileno #endif -#endif #include <stdbool.h> #include <stddef.h> diff --git a/libs/litehtml/src/gumbo/include/gumbo/error.h b/libs/litehtml/src/gumbo/include/gumbo/error.h index 3aa54a6b27..afc998b1fc 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/error.h +++ b/libs/litehtml/src/gumbo/include/gumbo/error.h @@ -19,10 +19,8 @@ #ifndef GUMBO_ERROR_H_ #define GUMBO_ERROR_H_ #ifdef _MSC_VER -#ifndef _CRT_SECURE_NO_WARNINGS #define _CRT_SECURE_NO_WARNINGS #endif -#endif #include <stdint.h> #include "gumbo.h" @@ -201,7 +199,7 @@ void gumbo_error_destroy(struct GumboInternalParser* parser, GumboError* error); // Prints an error to a string. This fills an empty GumboStringBuffer with a // freshly-allocated buffer containing the error message text. The caller is // responsible for deleting the buffer. (Note that the buffer is allocated with -// the allocator specified in the GumboParser ~config and hence should be freed +// the allocator specified in the GumboParser config and hence should be freed // by gumbo_parser_deallocate().) void gumbo_error_to_string(struct GumboInternalParser* parser, const GumboError* error, GumboStringBuffer* output); @@ -209,7 +207,7 @@ void gumbo_error_to_string(struct GumboInternalParser* parser, // Prints a caret diagnostic to a string. This fills an empty GumboStringBuffer // with a freshly-allocated buffer containing the error message text. The // caller is responsible for deleting the buffer. (Note that the buffer is -// allocated with the allocator specified in the GumboParser ~config and hence +// allocated with the allocator specified in the GumboParser config and hence // should be freed by gumbo_parser_deallocate().) void gumbo_caret_diagnostic_to_string(struct GumboInternalParser* parser, const GumboError* error, const char* source_text, diff --git a/libs/litehtml/src/gumbo/include/gumbo/tag_enum.h b/libs/litehtml/src/gumbo/include/gumbo/tag_enum.h index 6d7aeb3d7d..7a33d1e114 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/tag_enum.h +++ b/libs/litehtml/src/gumbo/include/gumbo/tag_enum.h @@ -73,6 +73,7 @@ GUMBO_TAG_INS, GUMBO_TAG_DEL, GUMBO_TAG_IMAGE, GUMBO_TAG_IMG, +GUMBO_TAG_PICTURE, GUMBO_TAG_IFRAME, GUMBO_TAG_EMBED, GUMBO_TAG_OBJECT, @@ -149,5 +150,6 @@ GUMBO_TAG_MARQUEE, GUMBO_TAG_MULTICOL, GUMBO_TAG_NOBR, GUMBO_TAG_SPACER, +GUMBO_TAG_DIALOG, GUMBO_TAG_TT, GUMBO_TAG_RTC, diff --git a/libs/litehtml/src/gumbo/include/gumbo/tag_gperf.h b/libs/litehtml/src/gumbo/include/gumbo/tag_gperf.h index 378eaf958c..525453946f 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/tag_gperf.h +++ b/libs/litehtml/src/gumbo/include/gumbo/tag_gperf.h @@ -1,105 +1,321 @@ -static unsigned int tag_hash( - register const char *str, register unsigned int len) { - static unsigned short asso_values[] = {296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 6, 4, 3, 1, 1, 0, - 1, 0, 0, 296, 296, 296, 296, 296, 296, 296, 22, 73, 151, 4, 13, 59, 65, 2, - 69, 0, 134, 9, 16, 52, 55, 28, 101, 0, 1, 6, 63, 126, 104, 93, 124, 296, - 296, 296, 296, 296, 296, 296, 22, 73, 151, 4, 13, 59, 65, 2, 69, 0, 134, - 9, 16, 52, 55, 28, 101, 0, 1, 6, 63, 126, 104, 93, 124, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, - 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296, 296}; +static unsigned int tag_hash(register const char *str, register size_t len) +{ + static unsigned short asso_values[] = + { + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 30, + 27, 27, 9, 6, 3, 6, 6, 3, 274, 274, + 274, 274, 274, 274, 274, 78, 3, 171, 12, 30, + 39, 129, 12, 105, 24, 156, 9, 51, 60, 87, + 12, 96, 3, 6, 18, 75, 99, 96, 36, 123, + 274, 274, 274, 274, 274, 274, 274, 78, 3, 171, + 12, 30, 39, 129, 12, 105, 24, 156, 9, 51, + 60, 87, 12, 96, 3, 6, 18, 75, 99, 96, + 36, 123, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274, 274, + 274, 274, 274, 274, 274, 274, 274, 274, 274 + }; register unsigned int hval = len; - switch (hval) { - default: - hval += asso_values[(unsigned char) str[1] + 3]; - /*FALLTHROUGH*/ - case 1: - hval += asso_values[(unsigned char) str[0]]; - break; - } - return hval + asso_values[(unsigned char) str[len - 1]]; + switch (hval) + { + default: + hval += asso_values[(unsigned char)str[1]+3]; + /*FALLTHROUGH*/ + case 1: + hval += asso_values[(unsigned char)str[0]]; + break; + } + return hval + asso_values[(unsigned char)str[len - 1]]; } -static const unsigned char kGumboTagMap[] = {GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_S, GUMBO_TAG_H6, GUMBO_TAG_H5, GUMBO_TAG_H4, - GUMBO_TAG_H3, GUMBO_TAG_SPACER, GUMBO_TAG_H2, GUMBO_TAG_HEADER, - GUMBO_TAG_H1, GUMBO_TAG_HEAD, GUMBO_TAG_LAST, GUMBO_TAG_DETAILS, - GUMBO_TAG_SELECT, GUMBO_TAG_DIR, GUMBO_TAG_LAST, GUMBO_TAG_DEL, - GUMBO_TAG_LAST, GUMBO_TAG_SOURCE, GUMBO_TAG_LEGEND, GUMBO_TAG_DATALIST, - GUMBO_TAG_METER, GUMBO_TAG_MGLYPH, GUMBO_TAG_LAST, GUMBO_TAG_MATH, - GUMBO_TAG_LABEL, GUMBO_TAG_TABLE, GUMBO_TAG_TEMPLATE, GUMBO_TAG_LAST, - GUMBO_TAG_RP, GUMBO_TAG_TIME, GUMBO_TAG_TITLE, GUMBO_TAG_DATA, - GUMBO_TAG_APPLET, GUMBO_TAG_HGROUP, GUMBO_TAG_SAMP, GUMBO_TAG_TEXTAREA, - GUMBO_TAG_ABBR, GUMBO_TAG_MARQUEE, GUMBO_TAG_LAST, GUMBO_TAG_MENUITEM, - GUMBO_TAG_SMALL, GUMBO_TAG_META, GUMBO_TAG_A, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_EMBED, - GUMBO_TAG_MAP, GUMBO_TAG_LAST, GUMBO_TAG_PARAM, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_NOBR, GUMBO_TAG_P, GUMBO_TAG_SPAN, GUMBO_TAG_EM, - GUMBO_TAG_LAST, GUMBO_TAG_NOFRAMES, GUMBO_TAG_SECTION, GUMBO_TAG_NOEMBED, - GUMBO_TAG_NEXTID, GUMBO_TAG_FOOTER, GUMBO_TAG_NOSCRIPT, GUMBO_TAG_HR, - GUMBO_TAG_LAST, GUMBO_TAG_FONT, GUMBO_TAG_DL, GUMBO_TAG_TR, - GUMBO_TAG_SCRIPT, GUMBO_TAG_MO, GUMBO_TAG_LAST, GUMBO_TAG_DD, - GUMBO_TAG_MAIN, GUMBO_TAG_TD, GUMBO_TAG_FOREIGNOBJECT, GUMBO_TAG_FORM, - GUMBO_TAG_OBJECT, GUMBO_TAG_LAST, GUMBO_TAG_FIELDSET, GUMBO_TAG_LAST, - GUMBO_TAG_BGSOUND, GUMBO_TAG_MENU, GUMBO_TAG_TFOOT, GUMBO_TAG_FIGURE, - GUMBO_TAG_RB, GUMBO_TAG_LI, GUMBO_TAG_LISTING, GUMBO_TAG_BASEFONT, - GUMBO_TAG_OPTGROUP, GUMBO_TAG_LAST, GUMBO_TAG_BASE, GUMBO_TAG_ADDRESS, - GUMBO_TAG_MI, GUMBO_TAG_LAST, GUMBO_TAG_PLAINTEXT, GUMBO_TAG_LAST, - GUMBO_TAG_PROGRESS, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_ACRONYM, GUMBO_TAG_ARTICLE, GUMBO_TAG_LAST, GUMBO_TAG_PRE, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_AREA, - GUMBO_TAG_RT, GUMBO_TAG_LAST, GUMBO_TAG_OPTION, GUMBO_TAG_IMAGE, - GUMBO_TAG_DT, GUMBO_TAG_LAST, GUMBO_TAG_TT, GUMBO_TAG_HTML, GUMBO_TAG_WBR, - GUMBO_TAG_OL, GUMBO_TAG_LAST, GUMBO_TAG_STYLE, GUMBO_TAG_STRIKE, - GUMBO_TAG_SUP, GUMBO_TAG_MULTICOL, GUMBO_TAG_U, GUMBO_TAG_DFN, GUMBO_TAG_UL, - GUMBO_TAG_FIGCAPTION, GUMBO_TAG_MTEXT, GUMBO_TAG_LAST, GUMBO_TAG_VAR, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_FRAMESET, GUMBO_TAG_LAST, - GUMBO_TAG_BR, GUMBO_TAG_I, GUMBO_TAG_FRAME, GUMBO_TAG_LAST, GUMBO_TAG_DIV, - GUMBO_TAG_LAST, GUMBO_TAG_TH, GUMBO_TAG_MS, GUMBO_TAG_ANNOTATION_XML, - GUMBO_TAG_B, GUMBO_TAG_TBODY, GUMBO_TAG_THEAD, GUMBO_TAG_BIG, - GUMBO_TAG_BLOCKQUOTE, GUMBO_TAG_XMP, GUMBO_TAG_LAST, GUMBO_TAG_KBD, - GUMBO_TAG_LAST, GUMBO_TAG_LINK, GUMBO_TAG_IFRAME, GUMBO_TAG_MARK, - GUMBO_TAG_CENTER, GUMBO_TAG_OUTPUT, GUMBO_TAG_DESC, GUMBO_TAG_CANVAS, - GUMBO_TAG_COL, GUMBO_TAG_MALIGNMARK, GUMBO_TAG_IMG, GUMBO_TAG_ASIDE, - GUMBO_TAG_LAST, GUMBO_TAG_CODE, GUMBO_TAG_LAST, GUMBO_TAG_SUB, GUMBO_TAG_MN, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_INS, GUMBO_TAG_AUDIO, - GUMBO_TAG_STRONG, GUMBO_TAG_CITE, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_INPUT, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_NAV, GUMBO_TAG_LAST, GUMBO_TAG_COLGROUP, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_SVG, GUMBO_TAG_KEYGEN, GUMBO_TAG_VIDEO, - GUMBO_TAG_BDO, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_BODY, GUMBO_TAG_LAST, GUMBO_TAG_Q, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_TRACK, - GUMBO_TAG_LAST, GUMBO_TAG_BDI, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_CAPTION, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_RUBY, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_BUTTON, - GUMBO_TAG_SUMMARY, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_RTC, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_BLINK, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_LAST, - GUMBO_TAG_LAST, GUMBO_TAG_LAST, GUMBO_TAG_ISINDEX}; +static const unsigned char kGumboTagMap[] = { + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_B, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_S, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_H6, + GUMBO_TAG_SPACER, + GUMBO_TAG_LAST, + GUMBO_TAG_RP, + GUMBO_TAG_LAST, + GUMBO_TAG_P, + GUMBO_TAG_H5, + GUMBO_TAG_DIR, + GUMBO_TAG_LAST, + GUMBO_TAG_H4, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_SMALL, + GUMBO_TAG_HEADER, + GUMBO_TAG_SAMP, + GUMBO_TAG_LABEL, + GUMBO_TAG_DEL, + GUMBO_TAG_DETAILS, + GUMBO_TAG_RB, + GUMBO_TAG_LEGEND, + GUMBO_TAG_HEAD, + GUMBO_TAG_BASEFONT, + GUMBO_TAG_SELECT, + GUMBO_TAG_LAST, + GUMBO_TAG_H3, + GUMBO_TAG_SOURCE, + GUMBO_TAG_BGSOUND, + GUMBO_TAG_H2, + GUMBO_TAG_SUB, + GUMBO_TAG_BASE, + GUMBO_TAG_DATALIST, + GUMBO_TAG_FOOTER, + GUMBO_TAG_LAST, + GUMBO_TAG_H1, + GUMBO_TAG_HGROUP, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_SUP, + GUMBO_TAG_PICTURE, + GUMBO_TAG_EMBED, + GUMBO_TAG_LAST, + GUMBO_TAG_TIME, + GUMBO_TAG_TITLE, + GUMBO_TAG_XMP, + GUMBO_TAG_FONT, + GUMBO_TAG_TABLE, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_TEMPLATE, + GUMBO_TAG_SCRIPT, + GUMBO_TAG_NOBR, + GUMBO_TAG_METER, + GUMBO_TAG_LAST, + GUMBO_TAG_FOREIGNOBJECT, + GUMBO_TAG_FIELDSET, + GUMBO_TAG_LAST, + GUMBO_TAG_SPAN, + GUMBO_TAG_NOFRAMES, + GUMBO_TAG_MAP, + GUMBO_TAG_MATH, + GUMBO_TAG_PARAM, + GUMBO_TAG_LAST, + GUMBO_TAG_NOEMBED, + GUMBO_TAG_BR, + GUMBO_TAG_FIGURE, + GUMBO_TAG_SECTION, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_NOSCRIPT, + GUMBO_TAG_NEXTID, + GUMBO_TAG_LAST, + GUMBO_TAG_HR, + GUMBO_TAG_MGLYPH, + GUMBO_TAG_LAST, + GUMBO_TAG_EM, + GUMBO_TAG_LAST, + GUMBO_TAG_FORM, + GUMBO_TAG_TR, + GUMBO_TAG_LAST, + GUMBO_TAG_MARQUEE, + GUMBO_TAG_PROGRESS, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_MULTICOL, + GUMBO_TAG_BUTTON, + GUMBO_TAG_DATA, + GUMBO_TAG_LAST, + GUMBO_TAG_APPLET, + GUMBO_TAG_LAST, + GUMBO_TAG_DL, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_OPTGROUP, + GUMBO_TAG_LAST, + GUMBO_TAG_ABBR, + GUMBO_TAG_TEXTAREA, + GUMBO_TAG_VAR, + GUMBO_TAG_FIGCAPTION, + GUMBO_TAG_RT, + GUMBO_TAG_PRE, + GUMBO_TAG_HTML, + GUMBO_TAG_MENUITEM, + GUMBO_TAG_DIV, + GUMBO_TAG_LAST, + GUMBO_TAG_LI, + GUMBO_TAG_PLAINTEXT, + GUMBO_TAG_MAIN, + GUMBO_TAG_DT, + GUMBO_TAG_LAST, + GUMBO_TAG_BLOCKQUOTE, + GUMBO_TAG_LAST, + GUMBO_TAG_WBR, + GUMBO_TAG_BODY, + GUMBO_TAG_TT, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_STYLE, + GUMBO_TAG_STRIKE, + GUMBO_TAG_LAST, + GUMBO_TAG_FRAMESET, + GUMBO_TAG_OBJECT, + GUMBO_TAG_MENU, + GUMBO_TAG_MO, + GUMBO_TAG_BIG, + GUMBO_TAG_META, + GUMBO_TAG_TFOOT, + GUMBO_TAG_OUTPUT, + GUMBO_TAG_LAST, + GUMBO_TAG_FRAME, + GUMBO_TAG_LAST, + GUMBO_TAG_U, + GUMBO_TAG_IMAGE, + GUMBO_TAG_LAST, + GUMBO_TAG_LISTING, + GUMBO_TAG_DD, + GUMBO_TAG_DIALOG, + GUMBO_TAG_A, + GUMBO_TAG_MS, + GUMBO_TAG_OPTION, + GUMBO_TAG_LAST, + GUMBO_TAG_TD, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_RUBY, + GUMBO_TAG_MI, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_MTEXT, + GUMBO_TAG_LAST, + GUMBO_TAG_SUMMARY, + GUMBO_TAG_UL, + GUMBO_TAG_NAV, + GUMBO_TAG_ACRONYM, + GUMBO_TAG_TBODY, + GUMBO_TAG_LAST, + GUMBO_TAG_LINK, + GUMBO_TAG_LAST, + GUMBO_TAG_DFN, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_OL, + GUMBO_TAG_COL, + GUMBO_TAG_LAST, + GUMBO_TAG_TH, + GUMBO_TAG_LAST, + GUMBO_TAG_ARTICLE, + GUMBO_TAG_THEAD, + GUMBO_TAG_CENTER, + GUMBO_TAG_Q, + GUMBO_TAG_COLGROUP, + GUMBO_TAG_CANVAS, + GUMBO_TAG_LAST, + GUMBO_TAG_ANNOTATION_XML, + GUMBO_TAG_LAST, + GUMBO_TAG_DESC, + GUMBO_TAG_VIDEO, + GUMBO_TAG_KBD, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_AUDIO, + GUMBO_TAG_LAST, + GUMBO_TAG_CODE, + GUMBO_TAG_MN, + GUMBO_TAG_INS, + GUMBO_TAG_I, + GUMBO_TAG_ASIDE, + GUMBO_TAG_LAST, + GUMBO_TAG_CITE, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_ADDRESS, + GUMBO_TAG_LAST, + GUMBO_TAG_BDO, + GUMBO_TAG_MARK, + GUMBO_TAG_INPUT, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_MALIGNMARK, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_KEYGEN, + GUMBO_TAG_AREA, + GUMBO_TAG_LAST, + GUMBO_TAG_STRONG, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_BDI, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_IFRAME, + GUMBO_TAG_ISINDEX, + GUMBO_TAG_LAST, + GUMBO_TAG_IMG, + GUMBO_TAG_CAPTION, + GUMBO_TAG_BLINK, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_TRACK, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_SVG, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_LAST, + GUMBO_TAG_RTC +}; diff --git a/libs/litehtml/src/gumbo/include/gumbo/tag_sizes.h b/libs/litehtml/src/gumbo/include/gumbo/tag_sizes.h index 7c92de073b..5b93c22fa2 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/tag_sizes.h +++ b/libs/litehtml/src/gumbo/include/gumbo/tag_sizes.h @@ -1,4 +1,4 @@ // Generated via `gentags.py src/tag.in`. // Do not edit; edit src/tag.in instead. // clang-format off -4, 4, 5, 4, 4, 4, 5, 6, 8, 8, 4, 7, 7, 3, 5, 2, 2, 2, 2, 2, 2, 6, 6, 6, 7, 1, 2, 3, 10, 2, 2, 2, 2, 2, 2, 6, 10, 4, 3, 1, 2, 6, 5, 1, 4, 1, 3, 4, 4, 4, 4, 3, 4, 3, 3, 3, 1, 1, 1, 4, 4, 2, 2, 3, 3, 4, 2, 3, 3, 3, 5, 3, 6, 5, 6, 5, 5, 5, 6, 5, 6, 3, 4, 4, 2, 2, 2, 2, 5, 6, 10, 14, 3, 13, 4, 5, 7, 8, 3, 5, 5, 5, 2, 2, 2, 4, 8, 6, 5, 5, 6, 6, 8, 8, 6, 8, 6, 6, 8, 5, 7, 7, 4, 8, 6, 7, 7, 3, 5, 8, 8, 7, 7, 3, 6, 7, 9, 2, 6, 8, 3, 5, 6, 4, 7, 8, 4, 6, 2, 3,
\ No newline at end of file +4, 4, 5, 4, 4, 4, 5, 6, 8, 8, 4, 7, 7, 3, 5, 2, 2, 2, 2, 2, 2, 6, 6, 6, 7, 1, 2, 3, 10, 2, 2, 2, 2, 2, 2, 6, 10, 4, 3, 1, 2, 6, 5, 1, 4, 1, 3, 4, 4, 4, 4, 3, 4, 3, 3, 3, 1, 1, 1, 4, 4, 2, 2, 3, 3, 4, 2, 3, 3, 3, 5, 3, 7, 6, 5, 6, 5, 5, 5, 6, 5, 6, 3, 4, 4, 2, 2, 2, 2, 5, 6, 10, 14, 3, 13, 4, 5, 7, 8, 3, 5, 5, 5, 2, 2, 2, 4, 8, 6, 5, 5, 6, 6, 8, 8, 6, 8, 6, 6, 8, 5, 7, 7, 4, 8, 6, 7, 7, 3, 5, 8, 8, 7, 7, 3, 6, 7, 9, 2, 6, 8, 3, 5, 6, 4, 7, 8, 4, 6, 6, 2, 3,
\ No newline at end of file diff --git a/libs/litehtml/src/gumbo/include/gumbo/tag_strings.h b/libs/litehtml/src/gumbo/include/gumbo/tag_strings.h index 6540e2e6ba..03d793c05d 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/tag_strings.h +++ b/libs/litehtml/src/gumbo/include/gumbo/tag_strings.h @@ -73,6 +73,7 @@ "del", "image", "img", +"picture", "iframe", "embed", "object", @@ -149,5 +150,6 @@ "multicol", "nobr", "spacer", +"dialog", "tt", "rtc", diff --git a/libs/litehtml/src/gumbo/include/gumbo/utf8.h b/libs/litehtml/src/gumbo/include/gumbo/utf8.h index ee852abfba..bd31a781ee 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/utf8.h +++ b/libs/litehtml/src/gumbo/include/gumbo/utf8.h @@ -62,7 +62,7 @@ typedef struct GumboInternalUtf8Iterator { int _current; // The width in bytes of the current code point. - ptrdiff_t _width; + int _width; // The SourcePosition for the current location. GumboSourcePosition _pos; diff --git a/libs/litehtml/src/gumbo/include/gumbo/util.h b/libs/litehtml/src/gumbo/include/gumbo/util.h index 98a7d1c466..6ad65649e6 100644 --- a/libs/litehtml/src/gumbo/include/gumbo/util.h +++ b/libs/litehtml/src/gumbo/include/gumbo/util.h @@ -20,10 +20,8 @@ #ifndef GUMBO_UTIL_H_ #define GUMBO_UTIL_H_ #ifdef _MSC_VER -#ifndef _CRT_SECURE_NO_WARNINGS #define _CRT_SECURE_NO_WARNINGS #endif -#endif #include <stdbool.h> #include <stddef.h> @@ -43,12 +41,12 @@ struct GumboInternalParser; char* gumbo_copy_stringz(struct GumboInternalParser* parser, const char* str); // Allocate a chunk of memory, using the allocator specified in the Parser's -// ~config options. +// config options. void* gumbo_parser_allocate( struct GumboInternalParser* parser, size_t num_bytes); // Deallocate a chunk of memory, using the deallocator specified in the Parser's -// ~config options. +// config options. void gumbo_parser_deallocate(struct GumboInternalParser* parser, void* ptr); // Debug wrapper for printf, to make it easier to turn off debugging info when diff --git a/libs/litehtml/src/gumbo/parser.c b/libs/litehtml/src/gumbo/parser.c index 653fd85acc..968fcc0f41 100644 --- a/libs/litehtml/src/gumbo/parser.c +++ b/libs/litehtml/src/gumbo/parser.c @@ -45,7 +45,7 @@ typedef char gumbo_tagset[GUMBO_TAG_LAST]; #define TAG_MATHML(tag) [GUMBO_TAG_##tag] = (1 << GUMBO_NAMESPACE_MATHML) #define TAGSET_INCLUDES(tagset, namespace, tag) \ - (tag < GUMBO_TAG_LAST && tagset[(int) tag] == (1 << (int) namespace)) + (tag < GUMBO_TAG_LAST && tagset[(int) tag] & (1 << (int) namespace)) // selected forward declarations as it is getting hard to find // an appropriate order @@ -572,6 +572,10 @@ static GumboInsertionMode get_appropriate_insertion_mode( } assert(node->type == GUMBO_NODE_ELEMENT || node->type == GUMBO_NODE_TEMPLATE); + if (node->v.element.tag_namespace != GUMBO_NAMESPACE_HTML) + return is_last ? + GUMBO_INSERTION_MODE_IN_BODY : GUMBO_INSERTION_MODE_INITIAL; + switch (node->v.element.tag) { case GUMBO_TAG_SELECT: { if (is_last) { @@ -812,7 +816,7 @@ InsertionLocation get_appropriate_insertion_location( GumboNode* last_table = open_elements->data[last_table_index]; if (last_table->parent != NULL) { retval.target = last_table->parent; - retval.index = (int)last_table->index_within_parent; + retval.index = last_table->index_within_parent; return retval; } @@ -2512,8 +2516,8 @@ static bool handle_in_body(GumboParser* parser, GumboToken* token) { return success; } else if (tag_in(token, kStartTag, (gumbo_tagset){TAG(ADDRESS), TAG(ARTICLE), TAG(ASIDE), - TAG(BLOCKQUOTE), TAG(CENTER), TAG(DETAILS), TAG(DIR), - TAG(DIV), TAG(DL), TAG(FIELDSET), TAG(FIGCAPTION), + TAG(BLOCKQUOTE), TAG(CENTER), TAG(DETAILS), TAG(DIALOG), + TAG(DIR), TAG(DIV), TAG(DL), TAG(FIELDSET), TAG(FIGCAPTION), TAG(FIGURE), TAG(FOOTER), TAG(HEADER), TAG(HGROUP), TAG(MENU), TAG(MAIN), TAG(NAV), TAG(OL), TAG(P), TAG(SECTION), TAG(SUMMARY), TAG(UL)})) { @@ -2582,7 +2586,7 @@ static bool handle_in_body(GumboParser* parser, GumboToken* token) { } else if (tag_in(token, kEndTag, (gumbo_tagset){TAG(ADDRESS), TAG(ARTICLE), TAG(ASIDE), TAG(BLOCKQUOTE), TAG(BUTTON), TAG(CENTER), TAG(DETAILS), - TAG(DIR), TAG(DIV), TAG(DL), TAG(FIELDSET), + TAG(DIALOG), TAG(DIR), TAG(DIV), TAG(DL), TAG(FIELDSET), TAG(FIGCAPTION), TAG(FIGURE), TAG(FOOTER), TAG(HEADER), TAG(HGROUP), TAG(LISTING), TAG(MAIN), TAG(MENU), TAG(NAV), TAG(OL), TAG(PRE), TAG(SECTION), TAG(SUMMARY), TAG(UL)})) { @@ -2613,7 +2617,7 @@ static bool handle_in_body(GumboParser* parser, GumboToken* token) { return success; } else { bool result = true; - const GumboNode* node = state->_form_element; + GumboNode* node = state->_form_element; assert(!node || node->type == GUMBO_NODE_ELEMENT); state->_form_element = NULL; if (!node || !has_node_in_scope(parser, node)) { @@ -2622,10 +2626,15 @@ static bool handle_in_body(GumboParser* parser, GumboToken* token) { ignore_token(parser); return false; } + // Since we remove the form node without popping, we need to make sure + // that we flush any text nodes at the end of the form. + maybe_flush_text_node_buffer(parser); // This differs from implicitly_close_tags because we remove *only* the // <form> element; other nodes are left in scope. generate_implied_end_tags(parser, GUMBO_TAG_LAST); - if (get_current_node(parser) != node) { + if (get_current_node(parser) == node) { + record_end_of_element(token, &node->v.element); + } else { parser_add_parse_error(parser, token); result = false; } @@ -2845,7 +2854,7 @@ static bool handle_in_body(GumboParser* parser, GumboToken* token) { text_state->_start_position = token->position; text_state->_type = GUMBO_NODE_TEXT; if (prompt_attr) { - size_t prompt_attr_length = strlen(prompt_attr->value); + int prompt_attr_length = strlen(prompt_attr->value); gumbo_string_buffer_destroy(parser, &text_state->_buffer); text_state->_buffer.data = gumbo_copy_stringz(parser, prompt_attr->value); text_state->_buffer.length = prompt_attr_length; diff --git a/libs/litehtml/src/gumbo/tag.c b/libs/litehtml/src/gumbo/tag.c index a394c0a677..85d58d28f0 100644 --- a/libs/litehtml/src/gumbo/tag.c +++ b/libs/litehtml/src/gumbo/tag.c @@ -41,7 +41,6 @@ void gumbo_tag_from_original_text(GumboStringPiece* text) { if (text->data == NULL) { return; } - assert(text->length >= 2); assert(text->data[0] == '<'); assert(text->data[text->length - 1] == '>'); @@ -54,10 +53,10 @@ void gumbo_tag_from_original_text(GumboStringPiece* text) { // Start tag. text->data += 1; // Move past < text->length -= 2; - // strnchr is apparently not a standard C library function, so I loop // explicitly looking for whitespace or other illegal tag characters. + // see https://html.spec.whatwg.org/multipage/syntax.html#tag-name-state for (const char* c = text->data; c != text->data + text->length; ++c) { - if (isspace(*c) || *c == '/') { + if (*c == ' ' || *c == '\t' || *c == '\n' || *c == '\f' || *c == '/') { text->length = c - text->data; break; } @@ -91,5 +90,5 @@ GumboTag gumbo_tagn_enum(const char* tagname, unsigned int length) { } GumboTag gumbo_tag_enum(const char* tagname) { - return gumbo_tagn_enum(tagname, (unsigned)strlen(tagname)); + return gumbo_tagn_enum(tagname, strlen(tagname)); } diff --git a/libs/litehtml/src/gumbo/tag.in b/libs/litehtml/src/gumbo/tag.in index 4c25264857..5716462914 100644 --- a/libs/litehtml/src/gumbo/tag.in +++ b/libs/litehtml/src/gumbo/tag.in @@ -70,6 +70,7 @@ ins del image img +picture iframe embed object @@ -146,5 +147,6 @@ marquee multicol nobr spacer +dialog tt rtc diff --git a/libs/litehtml/src/gumbo/tokenizer.c b/libs/litehtml/src/gumbo/tokenizer.c index 0d0ea0f241..71b5d32ad4 100644 --- a/libs/litehtml/src/gumbo/tokenizer.c +++ b/libs/litehtml/src/gumbo/tokenizer.c @@ -377,7 +377,7 @@ static bool temporary_buffer_equals(GumboParser* parser, const char* text) { // TODO(jdtang): See if the extra strlen is a performance problem, and replace // it with an explicit sizeof(literal) if necessary. I don't think it will // be, as this is only used in a couple of rare states. - size_t text_len = strlen(text); + int text_len = strlen(text); return text_len == buffer->length && memcmp(buffer->data, text, text_len) == 0; } @@ -726,7 +726,8 @@ static void copy_over_original_tag_text(GumboParser* parser, original_text->data = tag_state->_original_text; original_text->length = utf8iterator_get_char_pointer(&tokenizer->_input) - tag_state->_original_text; - if (original_text->data[original_text->length - 1] == '\r') { + if (original_text->length > 0 + && original_text->data[original_text->length - 1] == '\r') { // Since \r is skipped by the UTF-8 iterator, it can sometimes end up // appended to the end of original text even when it's really the first part // of the next character. If we detect this situation, shrink the length of @@ -751,7 +752,7 @@ static void finish_tag_name(GumboParser* parser) { GumboTagState* tag_state = &tokenizer->_tag_state; tag_state->_tag = - gumbo_tagn_enum(tag_state->_buffer.data, (unsigned)tag_state->_buffer.length); + gumbo_tagn_enum(tag_state->_buffer.data, tag_state->_buffer.length); reinitialize_tag_buffer(parser); } @@ -839,7 +840,7 @@ static bool is_appropriate_end_tag(GumboParser* parser) { assert(!tag_state->_is_start_tag); return tag_state->_last_start_tag != GUMBO_TAG_LAST && tag_state->_last_start_tag == gumbo_tagn_enum(tag_state->_buffer.data, - (unsigned)tag_state->_buffer.length); + tag_state->_buffer.length); } void gumbo_tokenizer_state_init( diff --git a/libs/litehtml/src/gumbo/utf8.c b/libs/litehtml/src/gumbo/utf8.c index ad73cefa6a..fdd6f8377c 100644 --- a/libs/litehtml/src/gumbo/utf8.c +++ b/libs/litehtml/src/gumbo/utf8.c @@ -137,7 +137,7 @@ static void read_char(Utf8Iterator* iter) { for (const char* c = iter->_start; c < iter->_end; ++c) { decode(&state, &code_point, (uint32_t)(unsigned char) (*c)); if (state == UTF8_ACCEPT) { - iter->_width = (int)(c - iter->_start + 1); + iter->_width = c - iter->_start + 1; // This is the special handling for carriage returns that is mandated by // the HTML5 spec. Since we're looking for particular 7-bit literal // characters, we operate in terms of chars and only need a check for iter @@ -181,7 +181,7 @@ static void read_char(Utf8Iterator* iter) { } static void update_position(Utf8Iterator* iter) { - iter->_pos.offset += (int)iter->_width; + iter->_pos.offset += iter->_width; if (iter->_current == '\n') { ++iter->_pos.line; iter->_pos.column = 1; diff --git a/libs/litehtml/src/gumbo/vector.c b/libs/litehtml/src/gumbo/vector.c index b9aa474e67..51758dfe03 100644 --- a/libs/litehtml/src/gumbo/vector.c +++ b/libs/litehtml/src/gumbo/vector.c @@ -30,7 +30,7 @@ const GumboVector kGumboEmptyVector = {NULL, 0, 0}; void gumbo_vector_init(struct GumboInternalParser* parser, size_t initial_capacity, GumboVector* vector) { vector->length = 0; - vector->capacity = (unsigned)initial_capacity; + vector->capacity = initial_capacity; if (initial_capacity > 0) { vector->data = gumbo_parser_allocate(parser, sizeof(void*) * initial_capacity); |