On 27 January 2014 15:06, Jacob Nevins
> Marko Lindqvist writes:
>> Windows gtk3-client might be badly broken. I've got a report that one
>> crosser-based build I made just a couple of days ago doesn't show any
>> graphics at all. [...] we cannot release anything before we've
>> confirmed that it works.
> Since we're not yet shipping the Windows Gtk3 client, I assume you're
> treating this as a blocker out of concern for (a) non-Windows Gtk3
> function, (b) possible non-Gtk3 brokenness, until we understand the root
Those do count of course, but I think even c) we support windows
gtk3-client in source code, though we don't ship official binaries of
it. I don't want reggressions compared to 2.4.1.
> (If we end up using new Windows Gtk3 libraries, I think it's worth cproc
> doing another test build, to see what state we're in.)
Maybe it's my imagination, or fluctuations in other load on my
computer, but when I tested the latest pixbuf unref patch, I noticed
(surprising) speed difference with and without the patch in linux (I
don't understand how it could affect speed, but maybe there's some
internal lists in libraries used that get slower when there's more
pixbufs in memory?).
I just tested crosser-based S2_4 build on Win7 and gfx seemed to
work. But so it worked when I tested the build reported to be broken
in XP. Also, client (both gt2k and gtk3) halted every time I tried to
spawn server - I had to run server separately instead. That's despite
me testing the "-A none" commandline -patch (the obvious suspect) on
same setup before it was committed.
Freeciv-dev mailing list