Tails

#6 Updated by Anonymous 2015-10-02 17:18:03

Could not reproduce the bug for now. Tested only using the “normal” installation way.

#10 Updated by intrigeri 2015-10-31 08:09:22

> Target version changed from Tails_1.7 to Tails_1.8

> I am postponing this.

Formally speaking, this is blocking Feature #8557 that is scheduled for 1.7 (and should really happen ASAP in practice), so I’m afraid that there’s some planning inconsistency here. But perhaps this should not be blocking Feature #8557.

> Until now I have not seen those errors again.

My understanding of the current situation is:

• we know there was a bug back in July, you’ve seen it a few times (but apparently we didn’t capture the debugging info anywhere back then, did we?);
• the chances that we’ve unconsciously fixed that bug already are extremely low, so most likely the bug will happen in the real world; OTOH I doubt we’ve introduced that bug in feature/jessie, so it’s likely has been in Tails forever, and in this case this should not be blocking Feature #8557 (and is not really a problem or I formally committed to work on);
• we know the bug only happens in some failure mode situations; we have a quite good understanding of what kind of situations they are (see the ticket description);
• you didn’t manage to set up the needed environment for reproducing it yet.

Is this a fair assessment?

If yes, then I think we need to re-evaluate whether this is a blocker for uploading to Debian. I can perhaps help do this by trying to set up what’s needed to reproduce the bug (it seemed to me, at some point, that I had a clearer understanding than you of how exactly one could trigger the bug). Once I’ve done this we can evaluate how grave the problem is and reconsider its blocking status.

What do you think?

#11 Updated by Anonymous 2015-11-02 11:43:13

intrigeri wrote:
> > Target version changed from Tails_1.7 to Tails_1.8
>
> > I am postponing this.
>
> Formally speaking, this is blocking Feature #8557 that is scheduled for 1.7 (and should really happen ASAP in practice), so I’m afraid that there’s some planning inconsistency here. But perhaps this should not be blocking Feature #8557.

thanks for the clarification which i was not aware of.

> > Until now I have not seen those errors again.
>
> My understanding of the current situation is:
>
> * we know there was a bug back in July, you’ve seen it a few times (but apparently we didn’t capture the debugging info anywhere back then, did we?);
> * the chances that we’ve unconsciously fixed that bug already are extremely low, so most likely the bug will happen in the real world; OTOH I doubt we’ve introduced that bug in feature/jessie, so it’s likely has been in Tails forever, and in this case this should not be blocking Feature #8557 (and is not really a problem or I formally committed to work on);
> * we know the bug only happens in some failure mode situations; we have a quite good understanding of what kind of situations they are (see the ticket description);
> * you didn’t manage to set up the needed environment for reproducing it yet.
>
> Is this a fair assessment?
>
> If yes, then I think we need to re-evaluate whether this is a blocker for uploading to Debian. I can perhaps help do this by trying to set up what’s needed to reproduce the bug (it seemed to me, at some point, that I had a clearer understanding than you of how exactly one could trigger the bug). Once I’ve done this we can evaluate how grave the problem is and reconsider its blocking status.
>
> What do you think?

Ack. Yes, I’d be grateful if you could help set up what’s needed to reproduce the bug.
As indeed, I only remember that it was related to non-english locales, but that’s about it. Not when it was triggered.

#14 Updated by intrigeri 2015-11-10 06:52:24

Note to myself: look for all remaining instances of subprocess.Popen.

#15 Updated by intrigeri 2015-11-10 07:19:38

Note to myself: try reproducing with self.popen too.