| Age | Commit message (Collapse) | Author |
|---|
|
until whitespace characters are reached. Improve tests accordingly.
|
|
|
|
|
|
|
|
|
|
|
|
somewhat harder to fix.
|
|
|
|
Idea: Only start unrolling anything on the parser stack
if an element that matches the given close token is found.
This requires the endToken method in DocumentChildHandler
to search for the given descriptor that might be ended.
While performing this search, only a specified number of
"explicit" structures/fields opened by the Stack class may
be skipped (those with implicit default fields).
Added an integration test ("python_code") which requires
this new (hopefully sane) behaviour.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
successfully serialized.
|
|
explicit fields.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
seemed semantically more fitting.
|
|
|
|
folder.
|
|
|
|
|
|
|
|
|
|
* Move test math environment to integration folder
* Add math_limits integration test
|
|
poems. Move \end{poem} to previous line to make sure closing special tokens are treated correctly.
|
|
bug involved here
|
|
* "poem" contains a nice use case for user defined syntax that still needs to be implemented (indent and dedent as open/close tag)
* "simple_short_token" triggers a non-deterministic bug that is not caught by valgrind
* "wikilike" contains various user defined syntax tests as well as the not yet implemented per-field user defined syntax
|
|
different form)
|
|
|
|
|
|
|
|
|
|
typesystems exist anymore.
|
|
DocumentChildHandler::startAnnotation
|
|
|
|
|
|
|
|
|
|
or data have been given beforehand. Added unit tests.
|
|
|
|
|
|
|
