Please reply to buildd mails
Hi,
there are now 125 packages in state Build-Attempted.
Dear co-builddadmins, if you get a mail saying log for _attempted_
build (not “for failed build”, these get giveback immediately by
the buildd software), do:
① Look through the build log (gzip compressed attachment, or on
http://buildd.debian-ports.org/status/package.php?p=$pkgname ).
② Scan for a failure reason. If none can be seen, reply to the
mail, to the buildd, with a mail whose body is just the one
word “fail” on a line of its own, then goto 5. Otherwise,
$failure be the reason.
③ Look whether it’s worth to reportbug against the package. If
so, mention $reason in the bug, wait until the bug is recorded
in the BTS, and set $reason = "#123456" where 123456 is the bug
number in debbugs
④ Reply to the mail, to the buildd, with a message body looking
like "fail\n\n$reason\n". Optionally, if it’s not been reported
to the BTS and Julien Cristeau is not the maintainer (or any
other person complaining about “yay, more debian-ports spam”),
consider putting the package maintainer on Cc (may want to put
the buildd on Bcc then)
⑤ Now, and only now, you’re allowed to delete that mail ;-)
⑥ If the package’s reason to FTBFS is that it needs porting, and
such porting is unlikely to be done in the next upload or other‐
wise RSN, ssh to leda and set the package on the Not-For-Us list:
$ wanna-build -A m68k -d unstable -n ${Source}_${Source:Version}
*BEFORE* doing so, make sure that the “fail” mail from your buildd
is processed, by running:
$ wanna-build -A m68k -d unstable --info ${Source}
And inspecting the output; good looks like this (relevant fields):
python-event:
State : Failed
Previous-State : Build-Attempted
Failed : #632763
And bad looks like this:
mingw32:
State : Build-Attempted
Previous-State : Building
That’s bad because it’s missing a reason, *and* your buildd is
still keeping state. For the record, after -n it looks like:
mrpt:
State : Not-For-Us
Previous-State : Failed
Old-Failed : -------------------- 1:1.0.0-1 --------------------
virtual memory exhausted: Cannot allocate memory
apparently 768 MiB RAM + 3.6 GiB swap aren=E2=80=99t enough.
We cannot provide more on m68k really.
That’s why. (Though p.d-p.o doesn’t seem to display these reasons
for N-F-U packages, only for Failed ones, which we might want to fix.)
Thanks,
//mirabilos
--
Sorry, I’m annoyed today and you came by as an Arch user. These are the
perfect victims for any crime against humanity, like systemd, feminism
or social democracy.
-- Christoph Lohmann on dev@suckless.org
Reply to: