Re: guile-gnome stack for Fedora Core
steve tell <[email protected]>
| Newsgroups | gmane.lisp.guile.gtk |
|---|---|
| Message-ID | <[email protected]> |
On Tue, 12 Jun 2007, Andy Wingo wrote: >> Every time I try to do somthing interesting, I find big dark corners >> of >> gtk+ that just aren't working yet in guile-gnome-platform/gtk. >> I can't help but thing that there's a long way to go before "stable", >> but >> keep those great releases and improvements comming and we'll slowly >> get >> there! > > You might be right; gtk generally speaking has the 'normal' parts and > the 'legacy' parts. The normal parts map naturally onto GObject and the > standard bindings, and the others, well, you're seeing it ;) Depends on > what you're doing. > > OTOH I don't think I've ever written an app using anything from > guile-gnome without hacking guile-gnome itself. Not that I've written > all that many, but still... Glad to hear that I'm not alone :-) I wonder if worthy goals might be: - encouraging contributions to greatly expanded the test/demo suite - some kind of functional coverage check: if a method or function is called by a demo script that doesn't error out, there's a good chance that it isn't completely broken. On the latter, idea, could some module-system magic be used to take every exported function or method in a module, wrap it in code that would record when that routine was called, and the re-export the wrapper, so that the resulting instrumented module act just like the original to calling code? Of course the harder part would be automating the GUI actions needed to fully run the tests. Steve