webkitgtk build fails do to suspected race

  • Done
  • quality assurance status badge
Details
2 participants
  • Jack Hill
  • Ludovic Courtès
Owner
unassigned
Submitted by
Jack Hill
Severity
normal
J
J
Jack Hill wrote on 28 Jun 2019 16:53
webkitgtk build fails duo to suspected race
(address . bug-guix@gnu.org)
alpine.DEB.2.20.1906281043570.17508@marsh.hcoop.net
Hi Guix,

It appears that the attatched derivation to build webkitgtk using the
following guix

```
Generation 2 Jun 26 2019 20:21:09 (current)
guix ba91247
branch: master
commit: ba912479e8481d69b699ce348b35d5d70f0c9367
```

fails to build in a way that looks like a race. The suspicious lines in
the attatched log are:

```
Traceback (most recent call last):
File "/tmp/guix-build-webkitgtk-2.24.2.drv-0/webkitgtk-2.24.2/Source/WebKit/Scripts/generate-message-receiver.py", line 28, in <module>
import webkit.messages
EOFError: EOF read where object expected
```

I have been able to sucessfully build webkitgtk on multiple hosts by using
`-c 2`

Best,
Jack
J
J
Jack Hill wrote on 28 Jun 2019 16:59
(no subject)
(address . control@debbugs.gnu.org)
alpine.DEB.2.20.1906281057150.17508@marsh.hcoop.net
retitle 36419 webkitgtk build fails do to suspected race

Thanks
J
J
Jack Hill wrote on 28 Jun 2019 17:01
(address . 36419@debbugs.gnu.org)
alpine.DEB.2.20.1906281059510.17508@marsh.hcoop.net
The previously attatched log was generated with -c 12. I am currently
running a build with -c 1 to verify that the problem occurs there as well.

(I remember it happening with -c 1 in the past, but my memory could be
faulty).
J
J
J
Jack Hill wrote on 23 Sep 2019 22:54
webkitgtk fixed in later release?
(address . 36419@debbugs.gnu.org)
alpine.DEB.2.20.1909231650220.15057@marsh.hcoop.net
Today I was able to build webkitgkt 2.24.4 with Guix commit
4343eebfc9f7d44ec9a665f3374d1397301b542e. It looks like it is also
available on ci.guix.gnu.org:


This is a newer release of webkitgtk than the one that failed (2.24.2).
Do we think this means that webkitgtk build problem has been fixed, or did
we just get lucky with this release?

My feeling is that we should close this bug and can open a new one if it
re-occurs.

Thoughts?
Jack
L
L
Ludovic Courtès wrote on 24 Sep 2019 10:19
(name . Jack Hill)(address . jackhill@jackhill.us)(address . 36419-done@debbugs.gnu.org)
87v9tikswp.fsf@gnu.org
Hello,

Jack Hill <jackhill@jackhill.us> skribis:

Toggle quote (13 lines)
> Today I was able to build webkitgkt 2.24.4 with Guix commit
> 4343eebfc9f7d44ec9a665f3374d1397301b542e. It looks like it is also
> available on ci.guix.gnu.org:
>
> https://ci.guix.gnu.org/25hz999kgcjh4alm6liwm12jm3blf5zz.narinfo
>
> This is a newer release of webkitgtk than the one that failed
> (2.24.2). Do we think this means that webkitgtk build problem has been
> fixed, or did we just get lucky with this release?
>
> My feeling is that we should close this bug and can open a new one if
> it re-occurs.

I agree. Closed!

Thanks for following up on this,
Ludo’.
Closed
?