From 5f52f7be7fa1e119fc74cabe1a3a4fd44b6e4623 Mon Sep 17 00:00:00 2001 From: Nick Wellnhofer Date: Thu, 7 May 2015 17:26:36 +0200 Subject: Multiple issues with numeric entities This closes #33. --- src/houdini_html_u.c | 39 ++++++++++++++++++++++++++------------- 1 file changed, 26 insertions(+), 13 deletions(-) (limited to 'src/houdini_html_u.c') diff --git a/src/houdini_html_u.c b/src/houdini_html_u.c index 2cb14b4..eaf295e 100644 --- a/src/houdini_html_u.c +++ b/src/houdini_html_u.c @@ -12,32 +12,45 @@ houdini_unescape_ent(cmark_strbuf *ob, const uint8_t *src, size_t size) { size_t i = 0; - if (size > 3 && src[0] == '#') { - int codepoint = 0; + if (size >= 3 && src[0] == '#') { + int codepoint = 0; + int num_digits = 0; if (_isdigit(src[1])) { for (i = 1; i < size && _isdigit(src[i]); ++i) { - int cp = (codepoint * 10) + (src[i] - '0'); + codepoint = (codepoint * 10) + (src[i] - '0'); - if (cp < codepoint) - return 0; - - codepoint = cp; + if (codepoint >= 0x110000) { + // Keep counting digits but + // avoid integer overflow. + codepoint = 0x110000; + } } + + num_digits = i - 1; } else if (src[1] == 'x' || src[1] == 'X') { for (i = 2; i < size && _isxdigit(src[i]); ++i) { - int cp = (codepoint * 16) + ((src[i] | 32) % 39 - 9); + codepoint = (codepoint * 16) + ((src[i] | 32) % 39 - 9); - if (cp < codepoint) - return 0; - - codepoint = cp; + if (codepoint >= 0x110000) { + // Keep counting digits but + // avoid integer overflow. + codepoint = 0x110000; + } } + + num_digits = i - 2; } - if (i < size && src[i] == ';' && codepoint) { + if (num_digits >= 1 && num_digits <= 8 && + i < size && src[i] == ';') { + if (codepoint == 0 || + (codepoint >= 0xD800 && codepoint < 0xE000) || + codepoint >= 0x110000) { + codepoint = 0xFFFD; + } utf8proc_encode_char(codepoint, ob); return i + 1; } -- cgit v1.2.3