[03:51] Should questions about staging.launchpad.net be asked here or somewhere else? === maclin1 is now known as maclin [04:21] luk3yx: Here is fine. [04:21] How long does it take for it to clone a git repo? [04:22] luk3yx: staging.launchpad.net doesn't have git support today. What are you doing with it? [04:23] I was trying to clone a git repository to test the snap building feature. [04:23] staging can only upload to the staging snap store, which you can't install from unless you build a custom snapd. You probably want to use production for that instead. [04:24] staging's better if you're testing scripts to manipulate lots of bugs, or do other things that might disrupt other users. [04:24] Oh, okay. [04:24] Thank you. [04:25] What's https://qastaging.launchpad.net/? [04:27] luk3yx: It's like staging... except different. It runs a different branch of the code, and is mostly used by us for QA purposes. But it does have git support. [04:27] Oh, okay. [04:27] How often are the staging ones reset? [04:29] staging is reset weekly. qastaging is reset when we feel like it. [04:29] Okay. [04:30] Thank you for your help. [04:30] np [04:31] How long does qastaging take to clone a git repo? === maclin1 is now known as maclin === chihchun_afk is now known as chihchun === cpaelzer_ is now known as cpaelzer === dgadomski_ is now known as dgadomski [09:41] luk3yx: That's a bit of a "how long is a piece of string" question. What repository in particular? Do you mean that you're trying to set up a code import? [09:46] Hi, I tried copying to packages from one PPA to another, and I got some launchpad error mails. Now it's been stuck at either building or pending to be published, for 1 hour or so. Should I just wait, or should I report the issue? [09:46] https://launchpad.net/~ts.sch.gr/+archive/ubuntu/ppa/+packages?field.name_filter=java8&field.status_filter=published&field.series_filter= [09:46] *s/to/two [09:55] alkisg: It's backed up due to a particularly large rebuild in one PPA, but it'll get there eventually. [09:56] cjwatson: so no action needed on my part, thank you [09:56] Indeed. === MmikeM is now known as Mmike === chihchun is now known as chihchun_afk [14:54] I want to submit a bug to Launchpad, but I can't because I cannot validate my Ubuntu One account. Can someone do it for me? [14:56] What goes wrong when you try to validate it? [14:57] I need to complete a CAPTCHA, but it doesn't appear when JavaScript is disabled [14:57] I surf the web with JS disabled because of privacy. [14:57] The bug should be sent here: https://bugs.launchpad.net/midori [14:58] Here it is the bug report, you just have to copy and paste: https://paste.debian.net/955781 [14:58] I would be very helpful if someone could fill it for me :) [14:58] I have a separate browser profile for stuff like this. Its set to clear everything on exit - basically private browsing mode. I don't believe there's any realistic loss in privacy by jumping in there to do the odd task like this. [14:59] We aren't likely to drop the captcha requirement; it's an important spam defence. [15:01] I'm not asking you that, I just asked for someone to take 30 secs to help me fill a bug report. Though it would be an improvement if the CAPTCHA would work with JS disabled. [15:02] I'm afraid I don't file bugs on other people's behalf because that breaks the ability for developers to contact the bug reporter. [15:05] I thought midori just uses webkit === nacc_ is now known as nacc [15:07] From https://developers.google.com/recaptcha/docs/faq it looks like we can only support non-JS signups by weakening security, and our spam situation is already bad enough, sorry :( [15:09] There is an IRC channel called #midori, I'm also trying there. Thanks for your responses. === chihchun_afk is now known as chihchun [16:39] cjwatson: I've also been encountering some more LP OOPSes recently e.g. https://oops.canonical.com/?oopsid=OOPS-fbdcf5eec6283761189e84d5d66da7ba [16:39] https://oops.canonical.com/?oopsid=OOPS-fbdcf5eec6283761189e84d5d66da7ba === chihchun is now known as chihchun_afk [16:58] bdmurray: That looks worth filing a bug on: we appear to be spending at least a second on lots of per-bug queries that could almost certainly be done in bulk instead. [17:00] cjwatson: okay === JanC_ is now known as JanC [18:44] hi [18:45] how do I workaround "virtual memory exhausted: Operation not permitted" in launchpad i386 builds? [18:45] trying building mame there... [18:45] The usual approach is to reduce your build's parallelism level [18:45] Also worth looking into various memory-saving linker options [18:45] dh $@ --parallel [18:45] hummm [18:46] removing that will take a lot of time to build heh [18:46] there's a --max-parallel option. see debhelper(7) [18:46] let's see [18:46] so you could e.g. try --max-parallel=2 [18:47] what's the maximum used in launchpad farm? [18:47] the builder VMs have 4 VCPUs, so --parallel will default to that [18:47] ok, let's see if it works [20:45] didn't work, let's see without --parallel [21:46] didn't work either cjwatson [21:47] sergio-br2: 19:45 Also worth looking into various memory-saving linker options [21:47] sergio-br2: but I'm sorry, I can't really debug your build for you, only make general suggestions. [21:48] sergio-br2: https://launchpad.net/ubuntu/+source/mame appears to build fine on i386 so you should look at what it does; I see at least one memory-saving tweak in debian/rules there. [22:01] ahh, true [22:01] #export DEB_LDFLAGS_MAINT_APPEND = -Wl,--no-keep-memory [22:01] let's see there [22:05] Doesn't look like it's your linker step that's failing, so you might need to look into other flags there. I didn't look very hard. [22:46] cjwatson, does launchpad farm use swap? [23:08] sergio-br2: 4GiB swap IIRC [23:08] wut [23:08] man, mame became a huge monster heh [23:08] sergio-br2: Not that that's relevant to "virtual memory exhausted". [23:09] isn't it using swap in the build? [23:10] virtual memory = RAM + swap ? [23:12] RAM + swap would be more like physical memory, sort of [23:12] sergio-br2: You get that error when you try to allocate more than will fit in your process' address space. [23:12] sergio-br2: virtual memory is more about the addressable memory (so exceeding 2^32) [23:13] 2^32 = 4 GB right [23:13] Possibly even 3GiB on i386, I forget. At any rate it's not relevant how much RAM+swap is available beyond that, in this case. [23:13] 3 GB in PC arch [23:14] cjwatson: good point, it's not quite 4GB [23:14] Note that the amd64 builders are exactly the same VMs. [23:14] yeah, I forgot this detail, too much used to amd64