[11:20] <mitya57> Hi! https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/4316/+build/20190461 (riscv64) keeps failing with the following error:
[11:20] <mitya57> Error reading configuration: MAILPROG binary '/usr/sbin/sendmail' does not exist or is not executable at /usr/share/perl5/Sbuild/ConfBase.pm line 59.
[11:21] <mitya57> Is this a problem on Launchpad side, or hirsute riscv64 chroot is broken?
[11:21] <tomwardill> mitya57: I think that's known and should work on a retry.. paging wgrant for an update.
[11:21] <mitya57> Retried two times already, but ok, I will try again a bit later.
[11:27] <wgrant> mitya57: It's a weird glitch, the same as https://bugs.launchpad.net/ubuntu/+source/sbuild/+bug/1859010. I believe it only affects the first build on each boot, or something like that, so a retry should work. I retried the affected primary archive builds, but hadn't gone through the full buildd history yet.
[11:27] <wgrant> (I upgraded the riscv64 buildds earlier, so they all got rebooted a few times. and then the host crashed, so they rebooted again, which was super helpful.)
[11:28] <mitya57> wgrant: ok, thanks!
[11:49] <cjwatson> wgrant: Any reason we shouldn't just drop $mailto from our sbuildrc?
[12:05] <cjwatson> wgrant: The issue is that the validation for MAILPROG depends on the current value of MAILTO, but validation happens when config values are set, which AFAICS happens in hash iteration order so can't be relied on for anything.  This is very confusing and I'm not sure how it doesn't result in more problems than just this, but maybe I'm missing something.  But not setting $mailto would avoid this ...
[12:05] <cjwatson> ... particular manifestation of it at least.
[12:07] <cjwatson> Ah, $mailto used to be required, but not since xenial I think.
[13:32] <wgrant> cjwatson: Ah, I see! And yeah, mailto used to be mandatory.
[13:44]  * wgrant rolls out an image with mailto unset.
[17:54] <arunpyasi> Hi there, What takes it so much time to publish a binary package after its built?
[18:06] <arunpyasi> Just curious :D
[18:22] <arunpyasi> pappacena, cjwatson ^^
[19:18] <cjwatson> arunpyasi: There's a single publisher process that can sometimes have a long queue of work to do.  In the current architecture it's difficult to do very much about this; we have some plans for after our next internal cloud comes online
[19:19] <arunpyasi> cjwatson, sounds nice! :D Do we have an ETA for that? :D
[19:21] <cjwatson> arunpyasi: Not one I'm willing to commit to externally
[19:21] <cjwatson> The plans are in quite general terms at present and aren't in our roadmap for the 21.04 cycle
[19:22] <arunpyasi> cjwatson, Oh ok :) Thank you very much for the information.
[19:22] <cjwatson> The underlying cloud infrastructure work will hopefully happen this cycle, but most of that stuff is out of my hands so definitely no promises
[21:14] <arunpyasi> cjwatson, I see :)