"Whise" mailed in pywebkitgtk mailing list that set_transparent doesn't seem to work. I tried using webkit_web_view_set_transparent to true in GtkLauncher but it doesn't seem to work as well. "doesn't work" meaning we're not getting any transparency at all.
im planing to port dashboard to linux , and without this working feature it is dificult
webkit_web_view_set_transparent() did work a few months ago; I still have an old GTK webkit build (r30763, svn head somewhere around March 2008) I can build against on which the mc.c hack Alp Toker presented to FOSDEM 2008 has working transparency:
At least for me, transparency does not work anymore with mc.c built against current trunk (r35658) -- it has not since around May; unfortunately, I don't have a narrower interval.
I confirm this bug is still widely open. I'm trying to use it today in an applet of Cairo-Dock, and it has never been working for me.
Here is the part of the source code that is using this functionality, if it may be of some help:
I wonder if http://trac.webkit.org/changeset/67137 helped here?
Created attachment 67312 [details]
Alp Toker's transparency demo presented to FOSDEM 2008
Created attachment 67313 [details]
Web page with semi-transparent body
Created attachment 67314 [details]
Screenshot of test.html using Alp Toker's Demo
(In reply to comment #4)
> I wonder if http://trac.webkit.org/changeset/67137 helped here?
Thanks for the heads-up.
Enclosed is a screenshot of Alp Toker's mc.c demo (no longer available from his blog) compiled against Webkit/GTK+ revision 67308: as you can see, transparency is back: that's great news!
Out of curiosity, I ran the same demo on a Webkit build from April the 15th 2010 (I don't know the exact revision -- it's called version 1.1.90 in pkgconfig); to my surprise, it also worked: I therefore ignore when and how the webkit_web_view_set_transparent() functionality has been restored...
Regardless, it seems to me this bug can be closed. I should probably add that my latest build (svn r67308) is unfortunately very unstable (it segfaults on about any page, from google.com to webkit.org), but I really haven't investigated why, and have no reason to believe this is specifically related to your r67137 commit.