From a5cf11dac52606141dd246f88d8c59688462e395 Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Sat, 6 Sep 2014 20:48:54 +0200 Subject: Entity declarations --- src/html/houdini_html_u.c | 99 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 99 insertions(+) create mode 100644 src/html/houdini_html_u.c (limited to 'src/html/houdini_html_u.c') diff --git a/src/html/houdini_html_u.c b/src/html/houdini_html_u.c new file mode 100644 index 0000000..762f980 --- /dev/null +++ b/src/html/houdini_html_u.c @@ -0,0 +1,99 @@ +#include +#include +#include + +#include "houdini.h" +#include "utf8.h" +#include "html_unescape.h" + +size_t +houdini_unescape_ent(strbuf *ob, const uint8_t *src, size_t size) +{ + size_t i = 0; + + if (size > 3 && src[0] == '#') { + int codepoint = 0; + + if (_isdigit(src[1])) { + for (i = 1; i < size && _isdigit(src[i]); ++i) + codepoint = (codepoint * 10) + (src[i] - '0'); + } + + else if (src[1] == 'x' || src[1] == 'X') { + for (i = 2; i < size && _isxdigit(src[i]); ++i) + codepoint = (codepoint * 16) + ((src[i] | 32) % 39 - 9); + } + + if (i < size && src[i] == ';') { + utf8proc_encode_char(codepoint, ob); + return i + 1; + } + } + + else { + if (size > MAX_WORD_LENGTH) + size = MAX_WORD_LENGTH; + + for (i = MIN_WORD_LENGTH; i < size; ++i) { + if (src[i] == ' ') + break; + + if (src[i] == ';') { + const struct html_ent *entity = find_entity((char *)src, i); + + if (entity != NULL) { + strbuf_put(ob, entity->utf8, entity->utf8_len); + return i + 1; + } + + break; + } + } + } + + return 0; +} + +int +houdini_unescape_html(strbuf *ob, const uint8_t *src, size_t size) +{ + size_t i = 0, org, ent; + + while (i < size) { + org = i; + while (i < size && src[i] != '&') + i++; + + if (likely(i > org)) { + if (unlikely(org == 0)) { + if (i >= size) + return 0; + + strbuf_grow(ob, HOUDINI_UNESCAPED_SIZE(size)); + } + + strbuf_put(ob, src + org, i - org); + } + + /* escaping */ + if (i >= size) + break; + + i++; + + ent = houdini_unescape_ent(ob, src + i, size - i); + i += ent; + + /* not really an entity */ + if (ent == 0) + strbuf_putc(ob, '&'); + } + + return 1; +} + +void houdini_unescape_html_f(strbuf *ob, const uint8_t *src, size_t size) +{ + if (!houdini_unescape_html(ob, src, size)) + strbuf_put(ob, src, size); +} -- cgit v1.2.3 From 9d86d2f32303ae0048f6a5daa552bacceb9b12ea Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Tue, 9 Sep 2014 04:00:36 +0200 Subject: Update the spec with better entity handling --- Makefile | 4 ++-- spec.txt | 22 ++++++++++++---------- src/html/houdini_html_u.c | 2 +- 3 files changed, 15 insertions(+), 13 deletions(-) (limited to 'src/html/houdini_html_u.c') diff --git a/Makefile b/Makefile index b5e487d..5d13272 100644 --- a/Makefile +++ b/Makefile @@ -1,5 +1,5 @@ -CFLAGS=-g -pg -O3 -Wall -Wextra -std=c99 -Isrc $(OPTFLAGS) -LDFLAGS=-g -pg -O3 -Wall -Werror +CFLAGS=-g -O3 -Wall -Wextra -std=c99 -Isrc $(OPTFLAGS) +LDFLAGS=-g -O3 -Wall -Werror SRCDIR=src DATADIR=data diff --git a/spec.txt b/spec.txt index ebd6d98..112dccc 100644 --- a/spec.txt +++ b/spec.txt @@ -3762,20 +3762,20 @@ as the "unknown codepoint" character (`0xFFFD`) [Hexadecimal entities](#hexadecimal-entities) consist of `&#` + either `X` or `x` + a string of 1-8 hexadecimal digits -+ `;`. ++ `;`. They will also be parsed and turned into their corresponding UTF8 values in the AST. . - " ആ ಫ +" ആ ಫ . -

 " ആ ಫ

+

" ആ ಫ

. Here are some nonentities: . -  &x; &#; &#x; � &ThisIsWayTooLongToBeAnEntityIsntIt; &hi?; +  &x; &#; &#x; &ThisIsWayTooLongToBeAnEntityIsntIt; &hi?; . -

&nbsp &x; &#; &#x; &#123456789; &ThisIsWayTooLongToBeAnEntityIsntIt; &hi?;

+

&nbsp &x; &#; &#x; &ThisIsWayTooLongToBeAnEntityIsntIt; &hi?;

. Although HTML5 does accept some entities without a trailing semicolon @@ -3808,7 +3808,7 @@ code blocks, including raw HTML, URLs, [link titles](#link-title), and . [foo](/föö "föö") . -

foo

+

foo

. . @@ -3816,7 +3816,7 @@ code blocks, including raw HTML, URLs, [link titles](#link-title), and [foo]: /föö "föö" . -

foo

+

foo

. . @@ -3824,7 +3824,7 @@ code blocks, including raw HTML, URLs, [link titles](#link-title), and foo ``` . -
foo
+
foo
 

 .
 
@@ -4817,12 +4817,14 @@ in Markdown:
 
link

 .
 
-URL-escaping and entities should be left alone inside the destination:
+URL-escaping and should be left alone inside the destination, as all URL-escaped characters
+are also valid URL characters. HTML entities in the destination will be parsed into their UTF8
+codepoints, as usual, and optionally URL-escaped when written as HTML.
 
 .
 [link](foo%20bä)
 .
-
link

+
link

 .
 
 Note that, because titles can often be parsed as destinations,
diff --git a/src/html/houdini_html_u.c b/src/html/houdini_html_u.c
index 762f980..b8e2d8d 100644
--- a/src/html/houdini_html_u.c
+++ b/src/html/houdini_html_u.c
@@ -24,7 +24,7 @@ houdini_unescape_ent(strbuf *ob, const uint8_t *src, size_t size)
 				codepoint = (codepoint * 16) + ((src[i] | 32) % 39 - 9);
 		}
 
-		if (i < size && src[i] == ';') {
+		if (i < size && src[i] == ';' && codepoint) {
 			utf8proc_encode_char(codepoint, ob);
 			return i + 1;
 		}
-- 
cgit v1.2.3


From 0ae7f4f53720e867c92ac9465062285293568856 Mon Sep 17 00:00:00 2001
From: Vicent Marti 
Date: Wed, 10 Sep 2014 20:02:01 +0200
Subject: Handle overflows in the codepoint parser

---
 src/html/houdini_html_u.c | 20 ++++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

(limited to 'src/html/houdini_html_u.c')

diff --git a/src/html/houdini_html_u.c b/src/html/houdini_html_u.c
index b8e2d8d..49b4956 100644
--- a/src/html/houdini_html_u.c
+++ b/src/html/houdini_html_u.c
@@ -15,13 +15,25 @@ houdini_unescape_ent(strbuf *ob, const uint8_t *src, size_t size)
 		int codepoint = 0;
 
 		if (_isdigit(src[1])) {
-			for (i = 1; i < size && _isdigit(src[i]); ++i)
-				codepoint = (codepoint * 10) + (src[i] - '0');
+			for (i = 1; i < size && _isdigit(src[i]); ++i) {
+				int cp = (codepoint * 10) + (src[i] - '0');
+
+				if (cp < codepoint)
+					return 0;
+
+				codepoint = cp;
+			}
 		}
 
 		else if (src[1] == 'x' || src[1] == 'X') {
-			for (i = 2; i < size && _isxdigit(src[i]); ++i)
-				codepoint = (codepoint * 16) + ((src[i] | 32) % 39 - 9);
+			for (i = 2; i < size && _isxdigit(src[i]); ++i) {
+				int cp = (codepoint * 16) + ((src[i] | 32) % 39 - 9);
+
+				if (cp < codepoint)
+					return 0;
+
+				codepoint = cp;
+			}
 		}
 
 		if (i < size && src[i] == ';' && codepoint) {
-- 
cgit v1.2.3