Re: [Tails-dev] [review'n'merge:1.1] #6559, #7062, #6275: te…

Delete this message

Reply to this message
Autor: intrigeri
Data:  
A: The Tails public development discussion list
Assumpte: Re: [Tails-dev] [review'n'merge:1.1] #6559, #7062, #6275: test/6559-adapt-test-suite-for-Wheezy
anonym wrote (09 May 2014 15:31:06 GMT) :
> Could you please elaborate on all this just so we don't misunderstand
> each other, possibly by attaching screenshots of both cases to the ticket?


I'll do that later, as we really should fix this properly at some
point. But for now, being able to merge this branch seems
higher-priority.

>> presumably since the move to QXL.


> For the record, I get the exact same results with QXL and cirrus (I took
> screenshots and compared them to be really sure).


> Could you test if the temporary change


>     sed -i 's/qxl/cirrus/' features/domains/default.xml


> fixes the issue (or, equivalently, reverting 353fcde)?


This fixes the issue for me.

>> Needless to say, it prevents from running 99% of the test suite
>> without fixing it myself => reassigning to anonym.


> By the way, are you also affected by #7171 (Test suite fails to start
> Xorg with cirrus graphics device on x86_64 arch)?


I am (at least on this branch - 353fcde).

> Even if you are, you could use the cirrus + <cpu> workaround from #7171
> to be able to run the full test suite and at least test the other things
> in the branch, which may be more convenient for you depending on you
> availability. For that, just apply the attached patch.


Can't do that:

      Call to virDomainCreateWithFlags failed: unsupported
      configuration: guest and host CPU are not compatible: Host CPU
      does not provide required features: aes (Libvirt::Error)


I think it raises the bar way to high in terms of hardware needed to
run the test suite. I also tried:

* Westmere: my CPU isn't good enough
* Nehalem: fails to start X
* <cpu mode="host-model"> fails to start X
* kvm64: seems to work, but likely suboptimal

=> I've started the full test suite run with kvm64, and will report
back later.

Was the "fails to start X with libvirt from wheezy-backports" bug
reported upstream? If you do report the bug, I'll try to reproduce it
on current sid. Piling up workarounds is fun, but if this could be
fixed properly where it should, we would be very happy in the end :)

Cheers!
--
intrigeri
| GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc
| OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc