GNOME Bugzilla – Bug 693040
Preparations for major g-ir-doctool work
Last modified: 2015-02-07 16:48:06 UTC
This is stuff I've done at the DX hackfest.
Created attachment 235036 [details] [review] girparser: Remove unused method pylint was complaining about the non-existent instance member, self._filename
Created attachment 235037 [details] [review] Update .gitignore
Created attachment 235038 [details] [review] Remove the insane pre-commit hook I'm sick and tired of this.
Created attachment 235039 [details] [review] tests: %NULL-terminated is wrong This translates into "None-terminated" in Python, which is bad. We should eventually strip this phrase entirely for certain languages with a prefilter.
Created attachment 235040 [details] [review] ast: Track enum members by symbol
Created attachment 235041 [details] [review] ast: Track enum/bitfield member parents
Created attachment 235042 [details] [review] giscanner: Apply standard annotations to constant values
Created attachment 235043 [details] [review] girparser: Move <doc> handling to passthrough
Created attachment 235044 [details] [review] girparser: Clean up passthrough handling Instead of remembering to have to set unknown_depth, make a wrapper function that does it for us.
Created attachment 235045 [details] [review] girparser: Serialize and read back the instance_parameter g-ir-doc-tool wants to use the instance parameter to read docs and the parameter name, so it needs to be shuttled through the GIR.
Created attachment 235046 [details] [review] giscanner: Save function macros In order to replace gtk-doc, we need to ensure we add documentation for function macros. Put these in the GIR file as well, which is another thing that won't be in the typelib. We don't save the expansion, as it's hard to do from the parser as far as I can tell.
Review of attachment 235036 [details] [review]: Git log isn't very helpful for why this exists, so looks OK.
Review of attachment 235037 [details] [review]: Sure.
Review of attachment 235038 [details] [review]: Yeah...this would be an easier story to tell if we had the tests automated. Though maybe the Ubuntu Jenkins builder is running through g-i make check? Yep, it is: https://jenkins.qa.ubuntu.com/view/Raring/view/JHBuild%20Gnome/job/jhbuild-amd64-gobject-introspection/lastBuild/artifact/gobject-introspection.log
Review of attachment 235039 [details] [review]: Right.
Review of attachment 235040 [details] [review]: This is OK.
Review of attachment 235041 [details] [review]: Ok.
Review of attachment 235042 [details] [review]: Yep, looks right.
Review of attachment 235043 [details] [review]: Yep, nicer.
Review of attachment 235044 [details] [review]: ::: girepository/girparser.c @@ +345,3 @@ +static void +passthrough (ParseContext *ctx) Maybe set_passthrough_mode() or something? But not a big deal.
Review of attachment 235045 [details] [review]: Looks right.
Review of attachment 235046 [details] [review]: So gtk-doc has bits to ignore the standard FOO_IS_BAR macros that we should probably lift before we land this: http://git.gnome.org/browse/gtk-doc/tree/gtkdoc-scan.in#n847
Attachment 235036 [details] pushed as 7cfc4ce - girparser: Remove unused method Attachment 235037 [details] pushed as 4076684 - Update .gitignore Attachment 235038 [details] pushed as de84ea0 - Remove the insane pre-commit hook Attachment 235039 [details] pushed as 4f8c0f4 - tests: %NULL-terminated is wrong Attachment 235040 [details] pushed as 5581554 - ast: Track enum members by symbol Attachment 235041 [details] pushed as f027b68 - ast: Track enum/bitfield member parents Attachment 235042 [details] pushed as 0f5f94e - giscanner: Apply standard annotations to constant values Attachment 235043 [details] pushed as 5a6f6a7 - girparser: Move <doc> handling to passthrough Attachment 235044 [details] pushed as 93cd5b2 - girparser: Clean up passthrough handling Attachment 235045 [details] pushed as 9c0bd59 - girparser: Serialize and read back the instance_parameter I put macros in another branch, and we'll revisit them later.
*** Bug 656444 has been marked as a duplicate of this bug. ***
[Mass-moving gobject-introspection tickets to its own Bugzilla product - see bug 708029. Mass-filter your bugmail for this message: introspection20150207 ]