git.iko.soy/typeable/pandoc
on 7f65a4e help sign in
7f65a4eFixed issue #8: slow performance in parsing inline literals in RST reader. The problem was that ``#`` was seen by 'inline' as a potential link or image. Fix: insert 'notFollowedBy (char '`')' in link parsers.fiddlosopher 6e467c2Replaced "choice [(try (string ...), ...]" idiom with "oneOfStrings" in LaTeX reader.fiddlosopher f3437dd+ Added some needed "try"s before multicharacter parsers, especially in "option" contexts. + Removed the "try" from the "end" parser in "enclosed" (Text.Pandoc.Shared). Now "enclosed" behaves like "option", "manyTill", etc.fiddlosopher 73cbae7Added 'try' in front of 'string', where needed, or used a different parser, in RST reader. This fixes a bug where ````` would not be correctly parsed as a verbatim `.fiddlosopher ca93680Allow the URI in a RST hyperlink target to start on the line after the reference key.fiddlosopher df03ca2Added link to FreeBSD port in website.fiddlosopher f7fcbf5Use "gaps" in copyrightMessage string for cleaner code formatting.fiddlosopher 6c3d96aChanged 'encodeEntities' to 'escapeSGMLString'.fiddlosopher dc69255+ Simplified entity handling by removing stringToSGML from Entities.hs. It is no longer needed now that all entities are processed in the markdown and HTML readers. All calls to stringToSGML have been replaced by calls to encodeEntities. + Since inTag's attribute handling already encodes entities, calls to encodeEntities are no longer needed for attribute values, so they've been removed. + The HTML and Markdown readers now call decodeEntities on all raw strings (e.g. authors, dates, link titles), to ensure that no unprocessed entities are included in the native representation of the document. (In the HTML reader, most of this work is done by a change in extractAttributeName.) + The result is a small speed improvement (around 5% on my benchmark) and cleaner code.fiddlosopher 2148471Use encodeEntities rather than stringToSGML for contents of Str inline in Docbook and HTML writers, since now these strings should not contain literal entity references.fiddlosopher 8e0ad5aCleaned up handling of embedded quotes in link titles. Now these are stored as a '"' character, not as '&quot;'. The function escapeLinkTitle in the Markdown writer is unnecessary and was removed. Tests modified accordingly.fiddlosopher 141affdMore changes in entity handling: Instead of using entities for characters above 128 in HTML and Docbook output, we now just use unicode. After all, we're declaring UTF-8 content in the header. This makes the HTML and docbook files produced by pandoc much more readable and editable.fiddlosopher d064171Changes in entity handling: + Entities are parsed (and unicode characters returned) in both Markdown and HTML readers. + Parsers characterEntity, namedEntity, decimalEntity, hexEntity added to Entities.hs; these parse a string and return a unicode character. + Changed 'entity' parser in HTML reader to use the 'characterEntity' parser from Entities.hs. + Added new 'entity' parser to Markdown reader, and added '&' as a special character. Adjusted test suite accordingly since now we get 'Str "AT",Str "&",Str "T"' instead of 'Str "AT&T".. + stringToSGML moved to Entities.hs. escapeSGML removed as redundant, given encodeEntities. + stringToSGML, encodeEntities, and specialCharToEntity are given a boolean parameter that causes only numerical entities to be used. This is used in the docbook writer. The HTML writer uses named entities where possible, but not all docbook-consumers know about the named entities without special instructions, so it seems safer to use numerical entities there. + decodeEntities is rewritten in a way that avoids Text.Regex, using the new parsers. + charToEntity and charToNumericalEntity added to Entities.hs. + Moved specialCharToEntity from Shared.hs to Entities.hs. + Removed unneeded 'decodeEntities' from 'str' parser in HTML and Markdown readers. + Removed sgmlHexEntity, sgmlDecimalEntity, sgmlNamedEntity, and sgmlCharacterEntity from Shared.hs. + Modified Docbook writer so that it doesn't rely on Text.Regex for detecting "mailto" links.fiddlosopher f2de088Rewrote functions in Text/Pandoc/Shared so as not to use Text.Regex, which does not support unicode: - escapePreservingRegex removed - stringToSGML rewritten using Parsec parser - new parsers for SGML character entities - escapeSGML rewritten using specialCharToEntity - new function specialCharToEntityfiddlosopher 11456baChanged Markdown autoLink parsing to conform better to Markdown.pl's behavior. <google.com> is not treated as a link, but <http://google.com>, <ftp://google.com>, and <mailto:google@google.com> are.fiddlosopher c94daceFixed bug in 'extractTagType' in HTML reader: previous version was not skipping / in close tags.fiddlosopher 890fbe9Refactored markdown reader so that Text.Regex is not used. Replaced email regex test with a custom email autolink parser (autoLinkEmail). Also replaced 'selfClosingTag' with a custom function 'isSelfClosingTag'.fiddlosopher 8f0cfe9Fixed a bug in extractTagType in HTML Reader: the previous version extracted the attributes, too, which is not wanted.fiddlosopher c61f2b6Fixed bug in HTML attribute parser: now a space is required before an attribute. Previously, <a.b> would be parsed as an HTML tag with an attribute!fiddlosopher ca6cb23Modified Markdown writer to use autolinks when possible. So, instead of [site.com](site.com) we get <site.com>. Changed test suite accordingly.fiddlosopher 0646eefRewrote 'extractTagType' in HTML reader so that it doesn't use regexs.fiddlosopher 96919a6More smart quote bug fixes: + LaTeX writer now handles consecutive quotes properly: for example, ``\,`hello'\,'' + LaTeX reader now parses '\,' as empty Str + normalizeSpaces function in Shared now removes empty Str elements + Modified tests accordinglyfiddlosopher e6cc2aaFixed bug in smart quoting: recognize ' in contractions like "don't" as not beginning single quoted contexts.fiddlosopher 1121e87Removed 'gsub' entirely and replaced its uses with 'substitute'.fiddlosopher 8f07505+ Added a 'substitute' function to Shared.hs. This is a generic list function that can be used to substitute one substring for another in a string, like 'gsub' except without regular expressions. + Use 'substitute' instead of 'gsub' in the LaTeX writer. This avoids what appears to be a bug in Text.Regex, whereby "\\^" matches "\350". There seems to be a slight speed improvement as well. (Note: If this works, it would be good to replace other uses of gsub that don't employ regexs with 'substitute'.)fiddlosopher a7839a1Small bug fix to last change, and count "'S" as well as "'s" as possessive when followed by non-alphanumeric.fiddlosopher 2c64e79More tweaks to smart quote parsing: a ' is not a single quote start if followed by 's' and then a non-alphanumeric. (Yes, this is English-centric, I'm afraid. But it does help, and I can't think of a language in which 's' by itself is a word.)fiddlosopher 5a48839Minor tweaks to smart quoting code.fiddlosopher 8c25738Fixed bug in smart quote recognition: ' before ) or certain other punctuation must not be an open quote.fiddlosopher f3cf1ccFixed haddock documentation errors.fiddlosopher b4f3eb5Fix debian/changelog.roktas 60989d0Added support for tables in markdown reader and in LaTeX, DocBook, and HTML writers. The syntax is documented in README. Tests have been added to the test suite.fiddlosopher 4224d91Changed website to link to Google's download details page (with SHA1 checksum) rather than directly to the files.fiddlosopher 9a2c653More website tweaks. Added demo of extra xsl configuration and CSS in chunked xhtml produced from docbook.fiddlosopher 3528c72Minor changes to Makefile required by changes to website build system.fiddlosopher e7dabc9More Changes to website target. Moved to a templating system for the examples page.fiddlosopher 8d4a1edMore website changes. Include demo of docbook postprocessed by xmlto.fiddlosopher c80f181Reorganized Makefile target - now uses a subsidiary Makefile that can be run from the website directory for small changes.fiddlosopher 5503109Need to export TMPDIR in tempdir.sh.fiddlosopher f55b549On Cygwin, set TMPDIR to . before using mktemp. Otherwise one gets an error creating the output file in the /tmp directory. I haven't tracked this one down, but this should serve as a workaround.fiddlosopher 4658014Reverted r471. My alternative to --strip-trailing-cr didn't work. This only affects the test target on systems without GNU diff (rare), so I'm not too worried about it.fiddlosopher edfac2bSmall tweak on last demo in website.fiddlosopher 5bcfa22Added DocBook to description of package in Pandoc.cabal.in.fiddlosopher b37f8daSmall change in web page for "Pandoc features."fiddlosopher b25706eCleaned up markdown2pdf.in. Note that bibtex does not return an error condition when it gives warnings, so instead we grep for warnings or error messages to see if we need to print the log.fiddlosopher 8adb142Minor changes to markdown2pdf: removed an unnecessary '|| exit $?', and made sure error output goes to stderr.fiddlosopher 458bb40Fixed docbook writer test -- removed named entities.fiddlosopher 19123a5Don't use named entities in docbook writer. Instead, use numerical entities, for portability across stylesheets.fiddlosopher acb4dabAdded comment relevant to last revision.fiddlosopher 35a06b9Replaced diff --strip-trailing-cr with something more portable in runtests.pl. (This is a GNU option.)fiddlosopher