I committed the patch and the revert of arm64 hack on Debian git
https://salsa.debian.org/sdl-team/libsdl2/-/commit/de427c7c9501c9083749de0293712c6418764b30
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bug
upstream fixed it in https://github.com/libsdl-org/SDL/commit/345efdcb
** Changed in: libsdl2 (Ubuntu)
Status: Triaged => Fix Committed
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1976288
Ti
Reported upstream now on https://github.com/libsdl-org/SDL/issues/5748
** Bug watch added: github.com/libsdl-org/SDL/issues #5748
https://github.com/libsdl-org/SDL/issues/5748
** Changed in: libsdl2 (Ubuntu)
Status: New => Triaged
--
You received this bug notification because you are
Ok, after a few retries it failed once in the ppa (without change so
it's flaky behaviour, there were tries where the call under gdb worked
and the testsuite next failed)
https://launchpadlibrarian.net/605051571/buildlog_ubuntu-kinetic-
arm64.libsdl2_2.0.22+dfsg-4ubuntu3.5_BUILDING.txt.gz
Thread
So there is no error when building on a canonistack cloud instance nor
on a virtual ppa, it still fails on the real builders. I need to check
with the launchpad team the difference and if it's possible to get
access to that env but they are away for a long U.K weekend.
I tried to get some debug fr
A workaround is present in 2.0.22+dfsg-4, but the fact that `testatomic`
crashes seems like a bug somewhere (SDL? Ubuntu's toolchain? Ubuntu's
buildds? ...) so I'm reopening this.
** Changed in: libsdl2 (Ubuntu)
Status: Fix Released => New
--
You received this bug notification because you
This bug was fixed in the package libsdl2 - 2.0.22+dfsg-4
---
libsdl2 (2.0.22+dfsg-4) unstable; urgency=medium
[ Gianfranco Costamagna ]
* Ignore tests if the system is Ubuntu and architecture arm64,
requiring some real new hw to complete successfully
(workaround for LP: #
Interestingly it doensn't fail on qemu/arm64 chroot with kinetic
baseline, this is why I still think its some hw issue...
INFO: Mode: LockFree
INFO: Starting 4 readers
INFO: Starting 4 writers
INFO: Finished in 3.493000 sec
INFO:
INFO: Writer 0 wrote 100 events, had 2702 waits
INFO: Writer 1
As I had hoped, libsdl2_2.0.22+dfsg-4 in Debian is now running all the
tests (successfully), while libsdl2_2.0.22+dfsg-4~build1 in Ubuntu is
skipping the one that previously crashed on Ubuntu and running the rest.
--
You received this bug notification because you are a member of Ubuntu
Bugs, whic
** Changed in: libsdl2 (Ubuntu)
Assignee: (unassigned) => Gianfranco Costamagna (costamagnagianfranco)
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1976288
Title:
testatomic segfaults on Ubunt
@Simon, thanks for the report. I don't think there is a know problem
with the Ubuntu builders, that delta seems weird and would deserve at
least some explanations. I've pinged Gianfranco about it now
** Changed in: libsdl2 (Ubuntu)
Importance: Undecided => High
--
You received this bug notifi
libsdl2_2.0.22+dfsg-4 in Debian hopefully works around this crash, while
still having at least minimal test coverage on Ubuntu arm64.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1976288
Title:
tes
I would prefer it if any tests that need to be skipped conditionally are
accompanied by a reference to a bug report (on the basis that a failing
test is technical debt, and technical debt is a bug), either in
Launchpad for workarounds for Ubuntu-specific issues, in the Debian BTS
for workarounds fo
13 matches
Mail list logo