To sleep or not to sleep

It’s over 1AM again, and I’m still not finished with Gnome 2.28 review. I’ve spent quite some time this weekend and tonight looking at what was wrong with gnote, gnome-system-monitor, gparted and a few other c++ apps suddendly starting to crash after I updated glib on Friday. Turns out something in the mm stack is doing something wrong so I filled Gnome bug #598209.

Updated to epiphany-2.28 since I got sick of epiphany-2.26 crashing when I wanted to make it remember a new password. Turns out it’s not as nice as I would have thought a nearly two years efforts would be. Lots of problems where loading of a page would stop in the middle of the process. I had to install firefox to fill bug reports and access the pages that fails. That’s quite a regression but upstream is now aware of it through Gnome bug #598115. Hopefully it’ll be fixed for Gnome 2.28.1.

I also spent some time cleaning up unneeded revisions in tree since I had to occupy myself when building all those c++ bindings. So where are we now, a bit less than 41 packages to go for review and about 80% of completeness on my gnome 2.28 status page.

5 thoughts on “To sleep or not to sleep”

  1. Wow! Thanks for all of your hard work. Enjoying Gnome 2.26.3 and looking forward to 2.28.x. BTW, I use the gnome-light meta-ebuild so I don’t use epiphany or evolution.

  2. On epiphany side, I also tested some time ago epiphany-2.27* from gnome-overlay for checking webkit backend. I found a lot of bugs, but most of them webkit bugs :-/

    I went back to 2.26 that, even with some known bugs, works better for me. But I think that would be interesting to try to put epiphany-2.28.x as soon as possible in main tree as hardmasked for letting people to test new backend (and webkit in general) and report bugs, hoping epiphany-2.30 get better

    It’s only a suggestion of course, feel free to know what you prefer ๐Ÿ™‚

    Thanks a lot

  3. I have seen crashes of many gnome applications with new glib compiled with gcc-4.4 and -O3. Recompiling glib with -O2 fixes it.

  4. @J, yes, this is a know problem with gcc 4.4, it breaks lots of software because the optimizer has droped quite a few safeguards which makes it more useful as it bubbles up problems to the surface. Hence you should report these bugs to toolchain herd which would elevate relevant information to gcc people hopefully.

Comments are closed.