severity 861523 important thanks Chris Lamb wrote...
> If you can't reproduce, feel feel to close...
In spite of many efforts, nobody here at the Zürich BSP was able to
reproduce the issue. Attempts included several different build systems
(sbuild, pbuilder, docker, manual chroots etc.), fiddling time zone
information in several ways, poking docker configuration and many
things more.
Therefore I'm lowering the bug severity since that package does not
deserve to get kicked out of stretch for an issue that obviously occurs
under rare circumstances only. I will however leave it open, since
still I'd like to know what triggered the build failure on Chris'
system.
Some bits if anyone wishes to investigate further:
The expected output is '1977-09-05-125600' which is created from
a unix time_t in
vfs.py:553
| l = time.localtime(date)
| ls = time.strftime('%Y-%m-%d-%H%M%S', l)
The timestamp is created in
test-ls.sh:32
| WVPASS bup save -n src -d 242312160 src
The fact Chris observes '1977-09-05-1*3*5600' suggests the outside time
zone information (he is in +0100) is used in the localtime() call
although the bup test itself sets TZ to UTC. This looks quite weird. So
it might as well be a symptom of an, er, surprising behaviour somewhere
else.
Christoph
signature.asc
Description: Digital signature
