Hi! This time we will take a look at a rabbit hole involving debugging segmentation faults caused at the init(8) boot stage.
I first noticed something was wrong when seeing this in my boot messages:
* Caching service dependencies
segmentation fault
segmentation fault
Continue reading
Hi! With me and listout having plowed through a lot of build time issues with musl libc we have started to face some larger issues and noticed some patterns.
One of the issues that constantly pop up is error.h not existing.
fatal error: error.h: No such file or directory
Hi! This is my weekly report for week 4 of Summer of Code. This week I’ve done some more ebuild development and fixed some more build-time bugs for various packages. We’ve also discussed adding *-standalone compatibility libraries for musl.
Now the GNOME desktop can be emerged on musl profile. The installation of the full GNOME desktop on musl was quite
uneventful. I didn’t come across anything that would require patching. However,
one thing I’ve noticed is that GNOME on musl doesn’t seem to work very well with
-wayland flag turned on globally. So if you do plan on running GNOME on musl,
make sure to have +wayland. Other than that I don’t think there’s much to add.
Other than that one can now install the MATE desktop, which comes very close to
GNOME 2, on musl too. It was a part of my proposal, and with a couple of spare
hours at hand, I decided to try to emerge the mate meta-package. I did face two
issues which are patched [1][2] and one of the patches has been submitted
upstream[3]. The two packages that needed patching for getting MATE to install
were mate-power-manager and mate-sensors-applet.
The package mate-power-manager needed two patches, the header file execinfo.h is
not available in musl and hence need to be surrounded in include guards so that
it’s only included in ulibc and glibc systems, secondly there were some debug
backtrace functions that too needed to be run on only ulibc and glibc system
hence similar to the execinfo patch, I surrounded those instances of the
function in include guards. Funnily, the mate-sensors-applet package needed some
include guards to be removed. There were include guards around stdio.h and
local.h, for some reason the `ifdef` condition around local.h was getting
evaluated as false on musl, resulting in local.h not being included which in
turn resulted in the build failing. A similar issue was reported earlier[4] but
was unfortunately never patched.
However, I didn’t get much time to test MATE desktop so some run time bugs can
be expected.
For the next week, I’ll begin testing MATE and maybe set up GRS so that other
people and mentors can reproduce my environment easily (thanks to blueness).
Hope this is a concise report for Week 3. I’ll update the blog as soon as the
blog site comes up again, it’s down for now. Till then, see ya.
[1]: https://github.com/gentoo/gentoo/pull/26198
[2]: https://github.com/gentoo/gentoo/pull/26197
[3]: https://github.com/mate-desktop/mate-sensors-applet/pull/128
[4]: https://github.com/mate-desktop/mate-sensors-applet/issues/123
I finally have some time to catch up on my blog backlogs.
The second week of the coding period for this year’s Google Summer of Code has come to an end. Nothing interesting happened this week, just the continuation of the previous week’s work, so I’ll keep it short and simple.
As per my previous report, I worked further on debugging the librsvg crash.
Eventually, I was successful in getting gnome-shell to work with latest librsvg
from ::gentoo, unfortunately, I still don’t know what the actual issue is. What
worked was increasing either gnome-shell, gjs or pixman’s stack size, and with
advice from my mentor sam_ I have submitted a patch to increase gjs’s stack size
on musl[1].
I’ve started with using Gentoo’s package testing tools, initially, I went with
pkg-testing-tools, as it was quite easy to set up and get going, unfortunately
I didn’t work quite well on musl due to multilib-strict not working on musl,
hence I had to move with tatt. So, now I’m learning to use tatt.
My patch for samba has been merged and the patch for gnome-terminal is awaiting
review. With that merged, it’ll be possible to using GNOME desktop (gnome-light)
on musl.
I had few extra hours at hand and used to try and compile firefox on musl. As of
now (date), I’m unable to build firefox:rapid::gentoo with “+clang” or “+pgo”.
The bug with clang with firefox has already been reported[2] and is being worked on
however I don’t know why “+pgo” is not working. In brief, to build firefox (as
of date) on musl one must use rust from ::musl, and turn off clang and lto flag,
probably this will be fixed on newer releases.
With that, I’ll conclude my report for the second week on Musl support expansion to
support the GNOME desktop. I’ll start with gnome extras from Monday and for the next
two weeks, I’ll be working on that. Hope to bring in new exciting findings next
week and I’ll try to update the blog too.
[1]: https://github.com/gentoo/gentoo/pull/26056
[2]: https://bugs.gentoo.org/829033
This is my report for the third week of Gentoo GSoC.
In my proposal it says this week will be dedicated to getting Plasma desktop and core functionality working, but it already seems to run just fine by fixing the dependencies, which was done in week 1 & 2. The topics I’ve learnt about this week has instead been more ebuild development, Rust on musl, low level DBUS C api, GLib, setting up LDAP, Linux namespaces, and some other things.
Generally speaking the first two weeks was moslty me spamming out pretty easy to fix build time patches while this week’s been moslty learning and taking my time to resolve harder ones.
