summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2014-11-18Added python version of test runner.John MacFarlane
This tests the dynamic library when run without a second argument. The code makes use of an amended version of the normalization method from karlcow/markdown-testsuite. Closes #161.
2014-11-18make test: Use cmake to run tests. So we get api tests too.John MacFarlane
2014-11-18Add api_test directory to archive.John MacFarlane
2014-11-18Merge branch 'api_tests' of https://github.com/nwellnhof/CommonMark into ↵John MacFarlane
nwellnhof-api_tests
2014-11-18Removed remarkable.js benchmark.John MacFarlane
2014-11-18Fix api_test warningsNick Wellnhofer
2014-11-18Set compiler flags for api_testNick Wellnhofer
2014-11-18Add header file to CMake api_test sourcesNick Wellnhofer
2014-11-18Try to support older CMake versionsNick Wellnhofer
2014-11-18Boldface commonmark.js too in benchmarks.John MacFarlane
2014-11-18Added a few more benchmarks.John MacFarlane
2014-11-18Start with tests for the C APINick Wellnhofer
The C API tests can be run individually via build/api_test/api_test Or together with the spec tests via cmake --build build --target test
2014-11-18Add node constructor and accessors to the public APINick Wellnhofer
The approach I'm taking is to copy inline literals internally to NULL-terminated C strings if requested by an accessor. This allows to return a 'const char *' that doesn't have to be freed by the caller.
2014-11-18Make render_html support nodes with no childrenNick Wellnhofer
For empty inline nodes like EMPH, the parser always creates a child containing an empty string. Using the tree manipulation API, nodes with no children can be created. Adjust render_html to cope.
2014-11-18Added remarkable.js benchmark.John MacFarlane
2014-11-18Initialize all fields in node when creating inlines.John MacFarlane
2014-11-18Fix to #205 that doesn't require raw HTML.John MacFarlane
2014-11-18Updated benchmarks.John MacFarlane
2014-11-18html: Removed union from RenderStack.John MacFarlane
It doesn't make sense to have a union here, and this simplifies the code.
2014-11-18Merge pull request #206 from nwellnhof/improve_node_checkJohn MacFarlane
Improve output of tree integrity check
2014-11-18Improve output of tree integrity checkNick Wellnhofer
2014-11-17Make sure the emph inline added in process_emphasis has a parent.John MacFarlane
2014-11-17Set CMARK_DEBUG_NODES for debug builds.John MacFarlane
2014-11-17Make parse_inlines add directly to parent.John MacFarlane
Previously parse_inlines returned a list of parsed inlines. This had to be added to the parent, and fix_parents had to be called to manually add the 'parent' links to the children, and the 'last_child' link to the parent. Now parse_inlines takes the parent block as a parameter, and uses cmark_node_append_child to add the children, so that the pointers should be properly managed. This avoids the need for the fix_parents pass.
2014-11-17Set last_child and parent in make_linkNick Wellnhofer
2014-11-17Store link labels as children in tree structureNick Wellnhofer
2014-11-17Set prev, parent and last_child for inlinesNick Wellnhofer
2014-11-17Move cmark_free_nodes to node.cNick Wellnhofer
2014-11-17Rename ast.h to parser.hNick Wellnhofer
2014-11-17Move reference type definitions to reference.hNick Wellnhofer
2014-11-17Move inline-related stuff to inlines.cNick Wellnhofer
2014-11-17Move public stuff from node.h to cmark.hNick Wellnhofer
2014-11-17Remove old node_block and node_inlNick Wellnhofer
2014-11-17Switch cmark_node_inl over to cmark_nodeNick Wellnhofer
2014-11-17Switch cmark_node_block over to cmark_nodeNick Wellnhofer
2014-11-17Start with unified nodesNick Wellnhofer
2014-11-17Merge pull request #205 from jesstelford/patch-1John MacFarlane
Fix display of triple-backtick to be cross-renderer
2014-11-18Fix display of triple-backtick to be cross-rendererJess Telford
GitHub Flavored Markdown (where this README.md is rendered) was showing the triple tick as <code> </code>\`<code> </code>. This fixes it to correctly display 3 backtick characters when rendered. Tested in reference implementation, Github, and dillinger.io
2014-11-17Make benchmarks.md table properly aligned.John MacFarlane
2014-11-17Put benchmarks in separate benchmarks.md file.John MacFarlane
2014-11-17More 'make bench' improvements.John MacFarlane
2014-11-17More 'make bench' improvements.John MacFarlane
2014-11-17'make bench': Use larger benchmark file, fewer runs.John MacFarlane
2014-11-17More 'make bench' improvements.John MacFarlane
2014-11-17Improved 'make bench'.John MacFarlane
2014-11-17Better benchmark.John MacFarlane
'make bench' runs the program 20 times against a file composed of 10 copies of Pro Git (about 5MB in all). It then does statistics on the results. We run it with high priority to get more consistent timings.
2014-11-17Merge pull request #203 from gcrico/Issue-202John MacFarlane
Fix Issue 202 - Catch RangeError thrown by native String.fromCodePoint (js)
2014-11-17Fix Issue 202 - Catch RangeError thrown by native String.fromCodePoint (js ↵Guillaume Crico
implementation) When using a JS engine that provides a native String.fromCodePoint ES6 implementation (e.g. SpiderMonkey), a RangeError is thrown if the codepoint is invalid. When adding the from-code-point.js polyfill, the js implementation was modified in order to handle invalid code point by returning the 0xFFFD placeholder glyph. So this is not a real "polyfill", but an specific implementation (adapted to the parser needs). So, if a native String.fromCodePoint implementation is availbale, the fromCodePoint function should catch the RangeError and return the 0xFFFD placeholder glyph.
2014-11-16Free the string returned by html renderer in main program.John MacFarlane
2014-11-16Moved refmap from the top-level block node to the doc_parser.John MacFarlane