git.iko.soy/typeable/pandoc
on master help sign in
Bug fixes in readers:

+ LaTeX reader: skip anything after \end{document} + HTML reader: fixed bug skipping material after </html> -- previously, stuff at the end was skipped even if no </html> was present, which meant only part of the file would be parsed and no error issued + HTML reader: added new constant eitherBlockOrInline with elements that may count either as block-level or inline + Modified isInline and isBlock to take this into account + modified rawHtmlBlock to accept any tag (even an inline tag); this is innocuous, because rawHtmlBlock is tried only if a regular inline element can't be parsed. git-svn-id: https://pandoc.googlecode.com/svn/trunk@862 788f1e2b-df1e-0410-8736-df70ead52e1b

commit4399db4fd2d3f1f936e21cd0fe732142a7fccab2 authorfiddlosopher <fiddlosopher@788f1e2b-df1e-0410-8736-df70ead52e1b> dateSat 18 Aug 2007, 23:44 (19 years ago) parente48f046
2 files changed+20−7
compare with master
unifiedsplit
src/Text/Pandoc/Readers/HTML.hs+18−6 src/Text/Pandoc/Readers/LaTeX.hs+2−1
src/Text/Pandoc/Readers/HTML.hs+18−6
@@ -59,11 +59,21 @@ readHtml = readWith parseHtml
5959-- Constants
6060--
6161
62+eitherBlockOrInline = ["applet", "button", "del", "iframe", "ins",
63+ "map", "area", "object", "script"]
64+
6265inlineHtmlTags = ["a", "abbr", "acronym", "b", "basefont", "bdo", "big",
6366 "br", "cite", "code", "dfn", "em", "font", "i", "img",
6467 "input", "kbd", "label", "q", "s", "samp", "select",
6568 "small", "span", "strike", "strong", "sub", "sup",
66− "textarea", "tt", "u", "var"]
69+ "textarea", "tt", "u", "var"] ++ eitherBlockOrInline
70+
71+blockHtmlTags = ["address", "blockquote", "center", "dir", "div",
72+ "dl", "fieldset", "form", "h1", "h2", "h3", "h4",
73+ "h5", "h6", "hr", "isindex", "menu", "noframes",
74+ "noscript", "ol", "p", "pre", "table", "ul", "dd",
75+ "dt", "frameset", "li", "tbody", "td", "tfoot",
76+ "th", "thead", "tr"] ++ eitherBlockOrInline
6777
6878--
6979-- HTML utility functions
@@ -171,12 +181,15 @@ htmlEndTag tag = try $ do
171181 char '>'
172182 return $ "</" ++ tag ++ ">"
173183
174−-- | Returns @True@ if the tag is an inline tag.
184+-- | Returns @True@ if the tag is (or can be) an inline tag.
175185isInline tag = (extractTagType tag) `elem` inlineHtmlTags
176186
187+-- | Returns @True@ if the tag is (or can be) a block tag.
188+isBlock tag = (extractTagType tag) `elem` blockHtmlTags
189+
177190anyHtmlBlockTag = try $ do
178191 tag <- anyHtmlTag <|> anyHtmlEndTag
179− if isInline tag then fail "inline tag" else return tag
192+ if isBlock tag then return tag else fail "inline tag"
180193
181194anyHtmlInlineTag = try $ do
182195 tag <- anyHtmlTag <|> anyHtmlEndTag
@@ -193,7 +206,7 @@ htmlBlockElement = choice [ htmlScript, htmlComment, xmlDec, definition ]
193206
194207rawHtmlBlock = try $ do
195208 notFollowedBy' (htmlTag "/body" <|> htmlTag "/html")
196− body <- htmlBlockElement <|> anyHtmlBlockTag
209+ body <- htmlBlockElement <|> anyHtmlTag <|> anyHtmlEndTag
197210 sp <- many space
198211 state <- getState
199212 if stateParseRaw state then return (RawHtml (body ++ sp)) else return Null
@@ -260,8 +273,7 @@ parseHtml = do
260273 spaces
261274 optional (htmlEndTag "body")
262275 spaces
263− optional (htmlEndTag "html")
264− many anyChar -- ignore anything after </html>
276+ optional (htmlEndTag "html" >> many anyChar) -- ignore anything after </html>
265277 eof
266278 return $ Pandoc (Meta title authors date) blocks
267279
src/Text/Pandoc/Readers/LaTeX.hs+2−1
@@ -117,7 +117,8 @@ parseLaTeX = do
117117 spaces
118118 blocks <- parseBlocks
119119 spaces
120− optional $ try (string "\\end{document}") -- might not be present (fragment)
120+ optional $ try (string "\\end{document}" >> many anyChar)
121+ -- might not be present (fragment)
121122 spaces
122123 eof
123124 state <- getState