| Mon | Tue | Wed | Thu | Fri | Sat | Sun |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 8 | 9 | 10 | 11 | 12 | 13 | 14 |
| 15 | 16 | 17 | 18 | 19 | 20 | 21 |
| 22 | 23 | 24 | 25 | 26 | 27 | 28 |
| 29 | 30 | 31 |
This section:
Blog postings by Andrew Cowie about Open Source and Software Development
The syndication links at top left will give you a feed for the blog as a whole.
If you'd like a feed specific to this sub-category, see bottom of page.
blogs > andrew > software > java-gnome > gtk26-to-gtk28-part1
Thu, 25 Aug 2005
Testing the GTK 2.6 to 2.8 upgrade path - Part 1
As a part of getting java-gnome ready for the next release, I’ve been testing the upgrade path.
Quite often in the past we’ve had situations where people have had a newer GTK+ but older libgtk-java, so the first test was is to see if we were ok in that scenario.
On a machine with GNOME 2.10 on board, I upgraded gtk+ to the newly released 2.8. At present, in Gentoo land, it’s still, quite wisely, hard masked, so I had to add appropriate temporary entries to /etc/portage/package.unmask and ~x86 keyworded them in /etc/portage/package.keywords.
>=dev-libs/glib-2.8
>=x11-libs/pango-1.10
>=x11-libs/cairo-0.9.2-r1
>=x11-libs/gtk+-2.8
With that, off emerge went.
Things were quite successful.
One bug I did hit was damage to the background as windows moved. I gather that problem was something to do with cairo not recognizing the vendor string in Gentoo’s X server. Had a good chat with vektor while I was doing this, and he pointed out the bug — which the Gentoo dev had already patched in cairo-0.9.2-r1 — but it turns out I was still hitting it because I was still running xfree on that box, not xorg. Anyway, he noted it in freedesktop bugzilla. And I finally got off my ass and upgraded my X server. :)
vektor also suggested I try Cairo 1.0.0 which is fresh out today. I subsequently do so with a manual ebuild bump. All good. [I note that Paul has reported a problem due to issues down in librsvg, but I haven’t run into that so far]
Next, I started running various apps to see that things are still working. All my daily use apps render fine.
Finally, and the point of the exercise, I fired off the various applications I have that use java-gnome. Those apps are running on top of the stable branch of libgtk-java 2.6 (ie CVS branch gtk-java-2-6, not HEAD) Everything worked fine. If anything, they’re faster.
In fact, the net result of these upgrades has been that desktop performance is dramatically improved. Granted, it’s impossible to tell how much of that is xfree to xorg, gtk 2.6 to 2.8, introducing cairo, or cairo 0.9.2 to cairo 1.0.0, but collectively it represents a user visible responsiveness improvement. Awesome work!
So as a field test from a stable, current, non-jhbuild, non-GARNOME environment, I can report that the GTK 2.8 upgrade is looking in good shape. And a 2.6 language binding, in this case Java, works fine against it.
AfC
Category Specific Feeds.
Use these links for an RSS or ATOM feed limited to this category and its descendants.
Technorati Profile
