GNOME Bugzilla – Bug 615299
add more fine-grained tests to Everything
Last modified: 2015-02-07 16:44:58 UTC
Bindings that want to have separate unit tests for marshalling of input, out and return values are creating their own tests, which overlap with Everything, for example pygi/tests/libtestgi.[hc] This code should be shared, and the best place is probably gobject-introspection.
I'm going to propose a patch that adds a new typelib called BasicTypes with functions for testing marshalling.
Created attachment 158307 [details] [review] Add testing typelib Marshalling-1.0.typelib
If people would prefer to merge these functions to Everything we should think about having methods that overlap and probably deprecating them. But it may make some sense having a typelib focused on marshalling for more systematic testing and another (Everything) where to add functions as we find corner cases and fix bugs.
Review of attachment 158307 [details] [review]: Was this code autogenerated? If so, where does the script live out of curiosity? I'm still unhappy with adding these things to the introspection core...exposing a new public API is not a trivial thing. I won't block adding it since we already have Everything and I understand the desire for separating the two. The namespacing is kind of poor here too, can we call the .gir like "GIMarshallingTests" or something? Thanks for the test submission though, I think having a lot of tests like this will be valuable.
(In reply to comment #4) > Review of attachment 158307 [details] [review]: > > Was this code autogenerated? If so, where does the script live out of > curiosity? No, I think they were extended from the tests I wrote for PyGI last fall (right, Tomeu)?
(In reply to comment #5) > (In reply to comment #4) > > Review of attachment 158307 [details] [review] [details]: > > > > Was this code autogenerated? If so, where does the script live out of > > curiosity? > > No, I think they were extended from the tests I wrote for PyGI last fall > (right, Tomeu)? This is the typelib you wrote as it is in PyGI now.
(In reply to comment #4) > Review of attachment 158307 [details] [review]: > > Was this code autogenerated? If so, where does the script live out of > curiosity? > > I'm still unhappy with adding these things to the introspection core...exposing > a new public API is not a trivial thing. I won't block adding it since we > already have Everything and I understand the desire for separating the two. Yeah, I'm afraid we cannot get this properly discussed in a reasonable timeframe and including all bindings authors, so maybe we'll have to bite the bullet and add it now if we think this beings clear value to bindings. > The namespacing is kind of poor here too, can we call the .gir like > "GIMarshallingTests" or something? Sure, that sounds way better. > Thanks for the test submission though, I think having a lot of tests like this > will be valuable. Btw, this is the hard work of Simon, I just moved the files over to g-i. Thanks for reviewing.
Pushed http://git.gnome.org/browse/gobject-introspection/commit/?id=d9e09987e2e7e
*** Bug 609436 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 ]