Mathieu Duponchelle

I have addressed all the review comments, either correcting the issues that were pointed out or answering the comments. This branch now installs a new public header "cmark_extension_api.h", separate from...

@jgm , ack, can you prioritize reviewing and merging this if you're happy with the design over merging (significant) patches in cmark? Keeping this branch cleanly rebased on master does...

@nwellnhof to be honest it feels like you're mostly paying attention to the form of this proposal, and not at all to its actual design. I can break up the...

By the way just clearing things out: > I also don't like that two additional pointers are added to struct cmark_node This is not the case, only one pointer is...

> For example, a table node could specify that it can appear wherever a paragraph can, > and a table cell could say that it can contain whatever a document...

By the way a better API for user data would be equivalent to https://developer.gnome.org/gobject/unstable/gobject-The-Base-Object-Type.html#g-object-set-data and https://developer.gnome.org/gobject/unstable/gobject-The-Base-Object-Type.html#g-object-set-data-full , with nodes maintaining a reference map, this would make it actually robust, that's...

I'll be in holidays for the next two weeks, going to https://wiki.gnome.org/Hackfests/GstSpringHackfest2016 , I'll check on that issue from time to time but I can't promise immediate reactivity :)

@jgm , ping?

Hello @jgm, I just rebased this branch on master, and this time it was a very painful experience. I'd really appreciate it if I could get some feedback :)

@jgm, the idea with this branch is that I would like cmark, given an AST, to always produce a correct output, except if "custom nodes" are used. The first solution...