Bug#1012318: diffoscope 214 produced no output and was killed after running into timeout after 150m

[Chris Lamb]

Vagrant Cascadian wrote:

>> Indeed, the --timeout option was implemented precisely to avoid having
>> no output whatsoever on tests.r-b.org.
>
> Nevertheless, it could get stuck on some arbitrary file, no? Or does it
> actually try to abort processing a file if it is taking too long?

You are right — diffoscope could get stuck processing some arbitrary
file and --timeout does not cause diffoscope abort halfway through
processing a file (or, more likely, a call to /usr/bin/readelf!) if
it's taking too long.

I have tried to make this pretty clear in the `--help` output as well
as clarifying that in any Salsa issue involving the --timeout flag
but it still seems to be causing some confusion, alas. :)  If it's
okay with you, I'll save an explanation of why this would be difficult
to implement within diffoscope for another venue, and not in this
(slightly) unrelated Debian bug report.

Either way, getting stuck on a particular file does not appear to be
the problem in this case. Assuming that the issue is not specific to
Jenkins' build of halide, you appear to be able to run it successfully
to the end...


Best wishes,

-- 
      o
    ⬋   ⬊      Chris Lamb
   o     o     reproducible-builds.org 💠
    ⬊   ⬋
      o