GNOME Bugzilla – Bug 760309
updated to build against clutter-1.25.3
Last modified: 2016-08-20 23:11:23 UTC
Created attachment 318452 [details] [review] updates configure.ac to use c++11 and reverts version bump This series of patches are the minimum to allow building cluttermm against clutter-1.25.3
Created attachment 318453 [details] [review] updates methods.defs
Created attachment 318454 [details] [review] this patches the methods where gmmproc misinterprets the parameter types
Created attachment 318455 [details] [review] regenerated enums.defs file
Created attachment 318456 [details] [review] regnerated signals.defs file
Created attachment 318457 [details] [review] removes deprecated API blocking the build from actor.hg
Created attachment 318458 [details] [review] Corrects the usage of BinAlignment enum in BinLayout
Created attachment 318459 [details] [review] Build blocks due to CairoContext being used as a parameter instead of cairo_t; manually altered.
Created attachment 318460 [details] [review] adds group.hg to filelist.am
Created attachment 318461 [details] [review] actor.h now required as #include for InputDevice.
Created attachment 318462 [details] [review] Removes reference to (deprecated) Clutter::Alpha
Created attachment 318463 [details] [review] Path parameters are now int, not float.
Created attachment 318464 [details] [review] Removes reference to (deprecated) Clutter::Content.
Is clutter-2.0 not worth worrying about at the moment then?
I wonder if it would be easier to patch the cluttermm-1.24 branch, which can then become cluttermm-1.26 (or master again).
My understanding (which may be poor) is that he clutter devs (especially Emmanuele Bassi) are working on GSK - GTK+ Scene-graph Kit, which will integrate a Clutter API into a GTK DrawingArea instead of having 2 different scene-graph APIs. However, they've been working on it for about 4 years now, and it still doesn't appear much closer than it was when I was working on cluttermm a couple of years back. Information on progress is sparse; there's an excellent LWN report outlining the reasons for moving past clutter from GUADEC 2014 at http://old.lwn.net/Articles/608501/ Having said that, the indications are that there won't be a 2.0 clutter release unless GSK proves too difficult; there's a pretty compelling argument for making sure it isn't too difficult, so I believe we'd be best to forget 2.0 until it actually happens. Patching to a 1.24 branch makes sense.
Hi, Anything I can do to get this accepted? I'm currently working my way through the tutorial updating it, but I'd at least like to have a buildable and working cluttermm branch in git. I'm also thinking it makes sense to merge the tutorial into the documentation built with cluttermm; it ties the documentation to a version if nothing else. It shouldn't be too difficult given other projects have managed it( the clutter cookbook is build as part of clutter). Ian.
(In reply to Ian Martin from comment #16) > Hi, > Anything I can do to get this accepted? Sorry, I had understood that you would instead make a patch against the cluttermm-1.24 branch, which presumably should need far less work. We might want to make make that master again, leaving any cluttermm-2.0 work in a branch, in case it's ever useful. But that's a separate issue.
OK, sorry, my bad. I was working off the gnome repo which doesn't contain that branch. I'll change to github.
What github repo. GNOME's git is mirrored on github, but we don't use it for anything. The cluttermm-1.24 branch is here: https://git.gnome.org/browse/cluttermm/log/?h=cluttermm-1-24
I got it now. Apologies, that was a 2am response. I've figured out my problem since - jhbuild modulesets point to the 1.18 branch.
I've opened two new bugs. https://bugzilla.gnome.org/show_bug.cgi?id=761847 : updates to cluttermm. https://bugzilla.gnome.org/show_bug.cgi?id=761971 : tutorial updates.
cluttermm-1.0 work is now back on git master, as we now know that clutter-2.0 won't happen. None of these bugs seem to make sense now. But please open a new bug report if you have any changes. It would be nice to really remove use of deprecated Clutter API and add the necessary replacement API.