[00:11] wgrant: It does indeed fix that bug, anyway. [00:44] cjwatson: Yay. [13:45] 2015-10-13 18:15:37.484 10525 WARNING Too long db name: launchpad_ftest_template_10065_e95fa37c_71a7_11e5_961a_00163e9546d5 [13:45] hmm [13:45] Maybe a UUID was a bit ambitious [13:48] Using uuid.hex would be enough to get us under the limit though [15:23] cjwatson: really going for scalability there, eh? [15:34] Totally! [15:34] (PIDs were clashing in practice sometimes, esp with LXC ...) [15:36] 16 bits is not a lot... [15:36] Then again, I've worked with long, randomised pids and those are no fun either. :) [15:36] (Even if they are probably safer) [15:37] But security was more important than convenience... For a few weeks I had no independent access to food, bathrooms, the outside, or the internet while doing my work. [15:38] You really get to hate those long PIDs through a slow, fixed-80-column Windows NT terminal emulator with acute line-break bugs after a while... [15:38] I mean really _hate_. [22:44] wgrant: what happened to bug linking [22:44] wgrant: was it behind a feature flag, or never delivered? [22:44] lifeless: Never implemented. [22:45] The team was vaporised days after you left :) [22:45] ouch, thanks [22:46] lifeless: We're vaguely working on arbitrary object cross-references, though. [22:46] oh, that would be good [22:46] if meta [22:46] lifeless: question/bug/cve/spec links use generic infrastructure now, though the UI is not genericised yet. [22:46] Just changed over the last couple of weeks. [22:46] nice [22:47] The plan is to genericise the UI, and then basic bug linking becomes easy, and we can work from there. [22:49] lifeless: You probably recall the whole weirdness around multi-task bugs and bugwatches, though. [22:49] Some cases of buglinking are handled natively by tasks. [22:49] And external bugs can be linked to in a limited way. [22:50] But linking to other internal bugs is not currently possible. [22:50] And adding a third way is Fun™. [22:52] yes [22:52] federated 3 eva