TYPE_METHODS on record types

[philberty]

I think your repo on github is slightly out of date, since i have some of your commits in this. But i am behind a proxy not sure what way to use git send-mail to send the patch. Or what the best way is.

[davidmalcolm]

Yeah, sorry; the primary repo for the project is the one on fedorahosted.org:
http://git.fedorahosted.org/git/?p=gcc-python-plugin.git;a=summary
whereas this github repo is just a copy, and I haven't automated keeping the latter up-to-date with the former.

I've pushed 08ce1c4 from the Fedora repo to this github repo; not sure what github will make of that...

[philberty]

Hey David i think i got it updated now pprint is really nice you can see the type_methods really easily now 👍

[davidmalcolm]

Thanks.

Sorry, I meant the stdlib's pprint.pprint (on the methods), rather than gccutils.pprint. The latter does gives you a lot of info, but it tends to make for fragile test cases, since it's so verbose. If anything changes (e.g. something gain a new attribute), the stdout in the test case will change, and we'll get a failure.

Indeed, the Travis CI build:
https://travis-ci.org/davidmalcolm/gcc-python-plugin/builds/25233608
failed for f5dc99e due to an extra:

    .const = False

on the gcc.RecordType (the Travis build is running on gcc 4.6.3; presumably you're running on a different gcc, giving a diffferent stdout).

[I spent a few hours yesterday getting the build to pass cleanly on Travis, so I'm keen to keep it that way :) ]

[philberty]

Ah awesome :D i quite like github's with travis and the pull requests its very web'y and i am behind a fairly insane firewall at work so its easier to work with github for me. Will fix that now and reup

[philberty]

Whoop it passed travis, i am trying to resurrect cython pxd generator using this again i think i know what i am doing better so it does work better but kind of need that to compile classes in C++.

[davidmalcolm]

Thanks!

[philberty]

Thanks :) its such a good project!