<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Hi Jeremy Bícha,</p>
<p>Thanks for your patient.<br>
</p>
<p>On Thu, 5 Sep 2024 08:04:06 -0400 =?UTF-8?Q?Jeremy_B=C3=ADcha?= <jeremy.bicha@canonical.com>
wrote:</jeremy.bicha@canonical.com></p>
<p><jeremy.bicha@canonical.com></jeremy.bicha@canonical.com></p>
<jeremy.bicha@canonical.com>> On Mon, Sep 2, 2024 at 5:12 AM
zhangdandan <zhangdandan@loongson.cn> wrote:<br>
> > Now, there are still 2 failing tests.<br>
> > ```<br>
> > 182/185 mutter:core+mutter/x11 / x11 FAIL 10.27s (exit
status 133 or signal 5 SIGTRAP)<br>
> > 183/185 mutter:core+mutter/x11 / x11-sync FAIL 10.27s
(exit status 133 or signal 5 SIGTRAP)<br>
> ><br>
> > Ok: 178<br>
> > Expected Fail: 5<br>
> > Fail: 2<br>
> > Unexpected Pass: 0<br>
> > Skipped: 0<br>
> > Timeout: 0<br>
> ><br>
> > ```<br>
> ><br>
> > Could you add loong64 to the override_dh_auto_test
lists in d/rules?<br>
> <br>
> Well that is progress. The Debian GNOME team requests that
we have<br>
> some proof that mutter actually works on a non-release
architecture<br>
> before disabling test failures. The goal here from our
perspective is<br>
> not trying to get loong64 to build 99% of Debian packages,
but whether<br>
> those built packages actually are usable.<br>
> <br>
> The strongest evidence is if you can show that GNOME Shell
works on<br>
> that architecture.<br>
> <br>
> Alternatively, gnome-settings-daemon has build tests that
use mutter.<br>
> Those currently aren't run on several architectures because
mutter is<br>
> so broken that the tests fail. If you can remove the
!loong64<br>
> overrides from its debian/control so that the mutter tests
run and all<br>
> pass, I think that could be enough evidence too.<br>
> <br>
> Thank you,<br>
> Jeremy Bícha<br>
> <br>
> <br>
</zhangdandan@loongson.cn></jeremy.bicha@canonical.com>
<p>Please consider add loong64 in mutter source package's d/rules.<br>
```<br>
# Please keep this list in sync with debian/tests/installed-tests<br>
override_dh_auto_test:<br>
-ifeq (,$(filter mips64el riscv64 s390x,$(DEB_HOST_ARCH)))<br>
+ifeq (,$(filter mips64el riscv64 s390x loong64,$(DEB_HOST_ARCH)))<br>
$(TEST_COMMAND)<br>
-$(TEST_COMMAND_FLAKY) --logbase flaky-tests<br>
<br>
```<br>
There are also 2 tests fail with independent graphics card.<br>
<br>
But based on the above patch, I built mutter 46.4-1 in local ENV
(Due to mutter blocks many gnome-* packages).<br>
Please check
<a class="moz-txt-link-freetext" href="https://buildd.debian.org/status/package.php?p=gnome-shell&suite=sid">https://buildd.debian.org/status/package.php?p=gnome-shell&suite=sid</a>.<br>
Please check
<a class="moz-txt-link-freetext" href="https://buildd.debian.org/status/package.php?p=gnome-remote-desktop&suite=sid">https://buildd.debian.org/status/package.php?p=gnome-remote-desktop&suite=sid</a>.<br>
<br>
I built gnome-shell and gnome-remote-desktop in local ENV with
independent graphics card(Radeon HD 8570). <br>
gnome-shell and gnome-remote-desktop's test cases are all passed.
<br>
<br>
I installed gnome-shell(depend gnome-remote-desktop) and run
command "gnome-shell --version" and "gnome-shell --replace", <br>
the gnome-shell window is displayed normally. <br>
Since bugs can't display images, I'm purposely emailing you the
interface(3 images) of gnome-shell after running it.<br>
So sorry to bother you on IRC, thanks for your advise.<br>
<br>
BTW, without independent graphics card, build gnome-shell and
gnome-remote-desktop have test cases failure.<br>
After the maintainer merged mutter's patch, maybe I need to add
patch(add loong64 in d/rules) for gnome-shell and
gnome-remote-desktop.<br>
<br>
Thanks,<br>
Dandan Zhang<br>
</p>
</body>
</html>