Bug#1114754: uqm: flaky autopkgtest on riscv64: startup takes too long?

[Simon McVittie]

Source: uqm
Version: 0.8.0+dfsg-3.2
Severity: serious
Justification: https://release.debian.org/testing/rc_policy.txt §6a
X-Debbugs-Cc: debian-riscv at lists.debian.org
User: debian-riscv at lists.debian.org
Usertags: riscv64
User: debian-ci at lists.debian.org
Usertags: flaky
Control: affects -1 + src:libsdl2

Similar to #1051783, the autopkgtest 'savegames' intermittently fails on 
ci.debian.net, especially on riscv64 (which has much slower hardware 
than the other architectures tested):

https://ci.debian.net/packages/u/uqm/testing/riscv64/

but also occasionally on other architectures:

https://ci.debian.net/packages/u/uqm/testing/armhf/
https://ci.debian.net/packages/u/uqm/testing/ppc64el/

The testing migration infrastructure assumes this is a regression in a 
package that was upgraded, preventing packages like libsdl2 from 
migrating to testing until the uqm autopkgtest is retried.

If this test is expected to be unreliable, please mark it with 
"Restrictions: flaky" so that it isn't used to gate migrations from 
unstable to testing:

    Tests: savegames
    Restrictions: flaky
    Depends: ...

Or, if it's only expected to pass on "fast" architectures, it could be 
limited to those architectures:

    Tests: savegames
    Architecture: amd64
    Depends: ...

Or it could be marked with the "skippable" restriction:

    Tests: savegames
    Restrictions: skippable
    Depends: ...

and do its own logic to classify its results into three buckets:

* success: exit 0
* skipped, expected/ignorable failure or other neutral status: exit 77
* failure: any other exit status

Thanks,
    smcv