=== zyga_ is now known as zya === zya is now known as zyga [07:44] bryce, hey, did you see my php-xmlrpc question on ubuntu-devel@lists ? [07:52] I still don't understand how the i386 whitelist works exactly :/ [07:52] Voodoo and other dark magic. [07:52] why did libvirt-glib start building on i386 in impish? it's not listed on https://people.canonical.com/~ubuntu-archive/packagesets/impish/i386-whitelist [07:52] also https://launchpad.net/ubuntu/+source/webkit2gtk/2.32.1-1ubuntu1/+build/21562277 [07:53] webkit now depends on libmanette so that one should start building on i386, what's the magic to make it do so? [08:01] seb128: I also don't see why and think it is odd; the one remotely related change that I see is that it got a Arch:all package added, but I thought that shouldn't be enough to make it pass without whitelist [08:02] cpaelzer, right, the main reason I'm asking is because i386 autopkgtests are failing which is blocking the update [08:03] it also isn't in https://people.canonical.com/~ubuntu-archive/germinate-output/i386.impish/i386+build-depends [08:03] unsure what's the right way to resolve [08:03] I could hint the i386 tests but still I would like to understand how the i386 is working exactly, is that documented somewhere? [08:03] I guess it comes down to ask vor_lon and document whatever you learn at https://wiki.ubuntu.com/i386 afterwards ? [08:04] right, that page points to https://people.canonical.com/~ubuntu-archive/germinate-output/i386.impish/i386+build-depends [08:04] but libvirt-glib isn't listed in there [08:05] yes, I said that above [08:05] vorlon, ^ could you help me understanding what trigger libvirt-glib to start building on i386? [08:05] triggered [08:05] cpaelzer, thanks for the replies! [08:10] seb128: it's not that it started building, you can see that it didn't by looking at https://launchpad.net/ubuntu/+source/libvirt-glib/4.0.0-1 [08:10] it got an arch:all package [08:11] needs to be hinted, let me add that [08:15] oh it is only tests [08:15] that at least makes some sense then [08:19] Laney, oh thanks, I don't understand the logic that a arch all data binary makes it start being tested on an arch it isn't built for though? [08:21] arch:all binaries are architecture independent, it makes some sense if you think of it like that [08:22] hum [08:22] but that is translations only, how could that ever work? [08:22] proposed-migration doesn't know what it is [08:22] you probably want an actually binary or lib for a test to work [08:22] or the contents of the package [08:22] if we had the test-depends availble we could check instalalbility or something, but we don't [08:23] and so my hint that an arch:all was added turned out to really be the root cause - thanks for the explanation Laney [08:23] I see, so basically any package not built on i386 but with an arch all binary needs to be hinted [08:23] check the hints, we have looooooooooooooads of these already! [08:24] alright [08:24] another reason which makes me think the partial i386 archive thing is more work than it should be [08:24] but I guess that's another topic and we are stucked dealing with those issues for now :/ [08:24] Laney, thanks! [08:24] cpaelzer, thanks as well ;-) [08:44] seb128: for the webkit one, I think you need to add it to the script (hardcoding) temporarily, then once it's built it can be dropped as germinate will keep it then [08:53] Laney, thanks, let me try if I can figure that out [08:57] see previous commits, there's lots of examples of it [09:06] bah, update-i386-whitelist wants to add a stack of nvidia drivers, I'm not going to be the one to ack those changes [09:06] but at least it's commited for whoever is doing a refresh [09:09] hack out that code block imho [09:09] or commit it 🤠 [09:09] * Laney invokes apw [09:09] ? [09:10] seb128, is that wanting to add like all the numbers up to 1000 ? [09:10] seb128, those get intersected with reality of what exists so you are in theory adding none. [09:11] seb128, i do wish it did the intersection before telling you about it. [09:11] seb128, so i am happy that all and any nvidia drivers that exist should end up on the whitelist, and ack that in that sense. [09:11] apw, yes, seems that's what it is doing [09:12] seb128, it is ... challenging to look at for sure. but if you do run it, then dump the resulting packageset it only contains drivers which are in the archive [09:12] I've no idea how that intersection works though and no confidence it's really not going to screw up things from the output it's giving me [09:12] apw, could you please do it and check it is doing what you expect? [09:12] the intersection works by adding them, and the archive not adding them i believe. [09:13] seb128, for which series. [09:13] I'm just interested to get libmanette to build to unblock webkitgtk [09:13] apw, impish [09:16] one thing is that packagesets can contain sources which aren't in the actual ubuntu archive [09:16] like if any PPA has had the source uploaded, you can add it to a packageset [09:16] right, they have to be known to launchpad, so in any PPA, but yes [09:16] so it's a bit tricky output wise [09:16] it _really_ should work out which ones can be actually added and emit just those. [09:17] I dunno how the script can know, is there an API call which can tell you that? [09:17] * apw has a look at it [09:17] you can say "is this in ubuntu?" but that's not the same thing [09:17] for the nvidia thing even doing that would help === pitti is now known as Guest59598 [09:18] as we actually do want to add them when they are in a PPA, so when they hit the archive the first time [09:18] they get built, and we don't have to do the copy dance to fix it [09:19] yeah, so that limits the filtering you can do [09:19] for nvidia specfially known to launchpad is really what i want [09:19] well really I guess this script should be using an allowlist [09:20] ubuntu archive OR these trusted PPAs [09:20] likely that too, yes [09:20] then you could do the intersection thing, and avoid printing nonsense [09:20] well its normally not possible, because it uses the archive miss list [09:20] so it has to be in the archive in the automated case; it is these manual ones whcih are poop [09:21] and literally only this one, is a problem [09:23] yeah I guess it could be hardcoded couldn't it, in that loop [09:23] only add to nv_list if it's in the ppa [09:24] apw: what's the ppa? I can probably do this quickly [09:24] it makes more sense to just lookup the published versions in the "appropriate" places. [09:25] Laney, already doing it [09:25] ah YAY [09:25] even better! [09:42] ok implemented, waiting on confirmation of the 'pre-proposed' PPA for these before i propose for merge [09:42] seb128, whitelist updated with libmanette [09:56] apw, thanks! [14:58] seb128, yes I did, sorry yesterday was interrupt heavy, I'll investigate today [15:05] bryce, hey, thank you === didrocks999 is now known as didrocks === Foxtrot is now known as foxtrot [17:11] Remember Bitcoin in 2008??? Pi is a new digital currency developed by Stanford PhDs, To claim your piece of pi goto https://minepi.com and use "ilkde" as your invitation code. Get your piece of the pi now ! [17:11] ew spam ads [17:25] these uninstrusive ads help support the network [17:34] Remember Bitcoin in 2008??? Pi is a new digital currency developed by Stanford PhDs, To claim your piece of pi goto https://minepi.com and use "ilkde" as your invitation code. Get your piece of the pi now ! [17:42] Remember Bitcoin in 2008??? Pi is a new digital currency developed by Stanford PhDs, To claim your piece of pi goto https://minepi.com and use "ilkde" as your invitation code. Get your piece of the pi now ! [22:32] just in case you didn't want to leave anyway