[00:48] ScottK: this was me playing around with the buildd network firewall and possibly affecting NTP, which the buildd doesn't deal with very well [00:48] so manual to make sure nothing goes through the check while the daemon is resyncing [00:48] wgrant: testing your stuff ? that's next [01:00] wgrant: got a favorite build to play with on staging? [01:08] lamont: staging works now? [01:08] But no, I don't play favourites when building on shitty buildds. [01:08] I once chose hello, thinking it would take just a couple of minutes. [01:08] It naturally took 2.5 hours. [01:09] You probably have a better idea of what's going to build in a reasonable amount of time. [01:09] heh. yeah - I can test this with bind9, so I'm just doing that with my prod ppa and some quick footwork [01:10] dear publisher. hurry up. kthx [01:10] Ah, if it's a prod builder that should be fine... [01:10] Why do you need a publisher? [01:12] well... got a ppa with a convenient test package? [01:12] I uploaded a package, you see [01:12] But building and publishing are independent. [01:13] (unless the PPA is private) [01:13] well, ok. I wanted upload processing then [01:13] Ah, right. [01:15] and test build is building now (berkelium) [01:15] * wgrant watches. [01:15] * wgrant watches buildd-manager fail. [01:15] I did not expect to ever be thankful that all the other buildds were busy with really long builds [01:15] ? [01:15] Well, it's slow. [01:15] So we won't see much log for a while, probably. [01:16] unpacking chroot [01:16] Hm, I guess since everything else is busy, it's being speedy. [01:16] one of the other fixes in this is to flush stdout better in the early stages [01:16] fewer builders atm, too [01:17] A lot of the problem is buildd-manager latency, though. [01:17] checking build system type... x86_64-unknown-linux-gnu [01:17] checking host system type... x86_64-unknown-linux-gnu [01:17] checking whether make sets $(MAKE)... yes [01:18] wgrant: dear sir, please to propose your branch for merging, so that I can +1 it. that is all. kthx. [01:19] lamont: Is it easy enough to set it in LP to i386, and get a build dispatched? [01:19] Just to check that that still actually works too. [01:19] Since that is sorta the whole point of the change :) [01:21] you're killing me. :-p [01:21] * lamont applies the big hammer again [01:22] Building i386 build of bind9 1:9.7.1.dfsg.P2-2 in ubuntu lucid RELEASE [lamont/ppa] [01:22] Now, let's hope it works. [01:23] world order restored, I'll let this build finish [01:23] Thanks. [01:25] and not just because I was only totally lucky on the timing that let me see those lines in the output on the amd64 run [01:26] Heh. [01:26] -I./x86_32/include <-- but I'm hopeful [01:28] Yeah. [01:31] wgrant: lp:~lamont/launchpad/lp-buildd-70 rev 11407 is what I plan to deploy on monday [01:31] just pushed now, with changelog for rev 70 [01:32] Looks good to me. [01:32] i386 build finishing up now. [01:32] And it works. Yay. [01:32] Thanks. [01:33] checking build system type... i686-pc-linux-gnu [01:33] checking host system type... i686-pc-linux-gnu [01:33] \o/ [01:33] so... got any bzr-builder knowledge? [01:33] A little. You want to test that change? [01:34] not really so much, no. [01:35] I'll just plan on making them what wrote the change test it with me on monday before the rollout [01:35] Ah. Well, it's a pretty binary works-or-not thing that's easy to test. [01:35] james_w speculated last night that it might not. [01:35] he had a patch for me this morning [01:35] which is also in 70 [01:36] test cases are: (1) does it work on maverick (2) does it still work elsewhere [01:36] That can easily be tested now, if you want. [01:37] I figured I'd let the build queue get smaller , while I have a weekend [01:38] Ah, fair point. === bjf[afk] is now known as bjf [02:06] robbiew, skat: are both of you reading the ubuntu-release ml? Just had some (yet entirely uncoordinated) thoughts about growing the release-team, that I tried to formulate in a mail sent there === bjf is now known as bjf[afk] === doko__ is now known as doko