Thanks for the contribution! I'll take a look tonight, but looks like the tests are failing?

Thanks for the contribution! I'll take a look tonight, but looks like the tests are failing?

That is kinda weird to me considering I can run these tests on my machine with no problems.

It seems like they're primarily failing on Python 2.7? My initial impression is this looks fine to me, but it doesn't properly handle types with templates in them (eg, Foo<X, Y, Z> x, y, *z;).

I have a local branch that I was working on earlier this week which is rewriting the way we parse function definitions to properly interpret things like std::function<void(int)> which will include code for consuming a template properly. I was hoping to work on it last night, but it looks like it'll be tonight or tomorrow night.

I pushed my branch as PR #53 ... if you cherry-pick 286001c that will give you the function to properly consume templates.

Taking a harder look at my problems tonight, I remembered the direction I wanted to go with parsing, so I'll be updating my function rework branch with those changes. I think it'll make the most sense to fold the function/property detection together, since it's ambiguous which you're dealing with until a ( or , is encountered (except as part of a template).

As part of that work, I'll fold in your unit tests but probably discard the rest.

Taking a harder look at my problems tonight, I remembered the direction I wanted to go with parsing, so I'll be updating my function rework branch with those changes. I think it'll make the most sense to fold the function/property detection together, since it's ambiguous which you're dealing with until a ( or , is encountered (except as part of a template).

As part of that work, I'll fold in your unit tests but probably discard the rest.

Alright, sounds like a good plan.
Thanks for the quick and active responses, even though I still find it weird that the tests are failing and can't reproduce it on my machine.

... so I ended up rewriting the entire library?

As a result, cppheaderparser will be abandoned soon in favor of https://github.com/robotpy/cxxheaderparser (which doesn't have this bug and should be a lot more robust). Please check it out, would love any feedback. It's not a drop-in replacement, but it should be much more future proof.

I'm not inclined to spend time fixing this PR for Python 2.7, nor am I inclined to merge this until the tests pass. If you can make the tests pass I'm happy to merge this and push a final release to pypi.