summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2014-12-31Updated dingus - incorporated changes made in CommonMark-site.John MacFarlane
2014-12-31Updated spec changelog.John MacFarlane
2014-12-31Omit footer on template.John MacFarlane
2014-12-31Bump spec version to 0.14.John MacFarlane
2014-12-31Added CC-BY-SA license to documentation.John MacFarlane
Closes #55.
2014-12-31Merge pull request #256 from nwellnhof/windows_fixesJohn MacFarlane
Windows fixes
2014-12-31Recreate scanners.c only on demandNick Wellnhofer
2014-12-31Write to stdout in binary mode on WindowsNick Wellnhofer
This fixes the output of newlines.
2014-12-31Fix spec tests on WindowsNick Wellnhofer
2014-12-31Fix C++ API testNick Wellnhofer
2014-12-31Fix API test on WindowsNick Wellnhofer
set_tests_properties must not be called before adding the test.
2014-12-31Rename test builddir to 'build/testdir'Nick Wellnhofer
The directory 'build/test' clashed with the 'test' Makefile target. Since nmake doesn't support phony targets, the tests wouldn't be run on Windows.
2014-12-31Remove useless void* castNick Wellnhofer
2014-12-31Include guards and C linkage for cmark_ctype.hNick Wellnhofer
2014-12-31Feature test for va_copyNick Wellnhofer
MSVC doesn't support va_copy.
2014-12-31Add missing va_endNick Wellnhofer
2014-12-30Revert "Remove unneeded va_copy"John MacFarlane
This reverts commit 485ef21b95e257e9d9cbcaa804c3c164f1f49a80. Apparently the va_copy IS needed, because without this code we get segfaults in some cases. Closes #253. @nwellnhof, can you have a look at this issue and comment? I understand that this code was removed for portability reasons. Is there an alternative solution?
2014-12-30Revert "man: use a variable to store the char * from node_get_url."John MacFarlane
This reverts commit 46b67b710788be7924b5a412ab68eea3cac0cd96. I was mistaken that this helped. Which is good, because I would not have understood why it helped.
2014-12-30man: use a variable to store the char * from node_get_url.John MacFarlane
We get segfaults on some platforms when we do cmark_strbuf_printf(man, " (%s)", cmark_node_get_url(node)); but they go away with: url = cmark_node_get_url(node); cmark_strbuf_printf(man, " (%s)", url); I don't understand why. Closes #253.
2014-12-30Require cmake >= 2.8.9, use a ppa for travis.John MacFarlane
`package_find(PythonInterp)` doesn't find python 3.2 with the cmake travis provides.
2014-12-29Attempted optimization of cmark_ctype.John MacFarlane
Use a single lookup table for all character types. I'm not sure this actually helps so much.
2014-12-29Renamed a couple variables.John MacFarlane
2014-12-29.travis.yml: install python3 for build.John MacFarlane
2014-12-29Added cmark_ prefix to functions in cmark_ctype.John MacFarlane
2014-12-29Fixed leakcheck for python3.John MacFarlane
2014-12-29Use python3 for bench in Makefile.John MacFarlane
2014-12-29Added cmark_ctype.h with locale-independent isspace, ispunct, etc.John MacFarlane
Otherwise cmark's behavior varies unpredictably with the locale. `is_punctuation` in utf8.h has also been adjusted so that everything that counts all ASCII symbol characters count as punctuation, even though some are not in P* character classes.
2014-12-29spec: Define ascii punctuation character separately.John MacFarlane
That's because some ascii punctuation characters (like `<`) are not in a "P" character class (they are mathematical symbols). They need to be escapable so we'll treat them as "punctuation" here.
2014-12-29Added wrapper3.py (python3 sample wrapper).John MacFarlane
2014-12-29Updated tests to use python3.John MacFarlane
2014-12-29Fixed md2man for python3.John MacFarlane
Partially addresses #254. This change makes make_man_page.py work with either python3 or python2. The test suite is still dependent on python2, and this should be changed, but for now this at least makes a build possible.
2014-12-29Avoid warning about unused options parameter in man writer.John MacFarlane
2014-12-29Make `--sourcepos` affect xml writer too.John MacFarlane
2014-12-29Added options parameter to renderers.John MacFarlane
To keep the API simple and avoid API changes when new options are added, this is just a long integer. Set it by disjoining options that are defined as powers of 2: e.g. `CMARK_HTML_SOURCEPOS | CMARK_HTML_HARDREAKS`. Test options using `&`: `if (options & CMARK_HTML_SOURCEPOS)`. Added `--hardbreaks` and `--sourcepos` command-line options.
2014-12-28Added cmark_node_set_list_delim, cmark_node_get_list_delim.John MacFarlane
Even though this doesn't make a difference in default HTML output, it's worth keeping track; some output formats may allow you to distinguish lists with `1)` and with `1.` delimiters.
2014-12-28Added CMARK_NO_DELIM to cmark_delim_typeJohn MacFarlane
2014-12-28Removed old 'ast' format, now that we have 'xml'.John MacFarlane
The xml representation of the AST is not quite as pretty, but it contains the same information and is not in an ad hoc format. See #53.
2014-12-28Adjust benchmarks for cmark.John MacFarlane
2014-12-28Improved end column/end line calculations in finalize.John MacFarlane
2014-12-28Added end_column to cmark_node struct.John MacFarlane
API exports cmark_node_get_column. XML writer indicates start and end line and column for block-level nodes.
2014-12-28blocks.c - removed unneeded start_line parameter from make_block.John MacFarlane
2014-12-28blocks.c: removed redundant line_number param in finalize.John MacFarlane
Also break_out_of_lists.
2014-12-28Rename CMARK_NODE_LIST_ITEM -> CMARK_NODE_ITEM.John MacFarlane
2014-12-28Revamped xml writer and dtd to use lowercase element names.John MacFarlane
2014-12-28Rename cmark_node_type_string -> cmark_node_get_type_string.John MacFarlane
And export in cmark.h public header. Also, use lowercase names, not uppercase.
2014-12-28Added xml writer, to dump the AST in XML.John MacFarlane
This is a work-in-progress. CommonMark.dtd gives the DTD for the generated XML. Closes #53.
2014-12-28Expose cmark_node_type_string in node.h (private header).John MacFarlane
Previously was static function S_type_string.
2014-12-27pathological_tests.py: limit to 65000 reps.John MacFarlane
This is because some versions of python have a limit of N = 65535 in regex `{N}`.
2014-12-27Spec: added example with two blank lines in fenced code in sublist.John MacFarlane
Actually I just tacked this on to the existing Example 170. Closes #180.
2014-12-27Clarify precedence of empty list item over setext header line.John MacFarlane
CLoses #95.
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-05 15:34:13 +0000