2021/04/20/#launchpad.txt

[07:21] <krytarik> Hey, I just noticed that on the "FrontPage" of the wiki at <https://help.launchpad.net/>, the "Privacy... in Launchpad" item has faulty markup which leads to everything that follows it getting made bold or non-bold loosely at random - but even after having logged in the page is still immutable to me, so I'm just gonna mention here what exactly needs changed. ...
[07:21] <krytarik> ... ">~+'''[[PrivacyConfidentialityAndDisclosure | Privacy, confidentiality and disclosure]] in Launchpad" just needs the closing "+~" markup appended right after this.  I even managed to find the revision where this snuck in.. in 2012 XD - https://help.launchpad.net/FrontPage?action=diff&rev1=106&rev2=107
[07:57] <cjwatson> krytarik: Thanks!  Now if I can get that wiki to let me log in ...
[08:06] <cjwatson> krytarik: Fixed, thanks
[08:08] <krytarik> Yay, thanks!
=== cpaelzer__ is now known as cpaelzer
[09:58] <tomwardill> cjwatson: did mattermost just go down?
[09:58] <cjwatson> tomwardill: working for me
[09:58] <tomwardill> oh no, it's back
[09:58] <tomwardill> sigh.
[12:12] <coreycb> hi, we're consistently getting "No space left on device" for ceph builds in the cloud archive - https://launchpad.net/~ubuntu-cloud-archive/+archive/ubuntu/wallaby-updates/+packages?field.name_filter=ceph&field.status_filter=published&field.series_filter=
[12:13] <coreycb> surprisingly armhf passed! but is there anything we might be able to do if this is a builder issue?
[12:40] <cjwatson> All builders have the same disk provisioning.  What you can often do is adjust the build to remove intermediate build results as it goes along
[12:40] <cjwatson> armhf is probably just creating smaller object files because 32-bit
[12:41] <cjwatson> I see you already have a "running out of disk space on riscv64" tweak along those lines; you could just extend that to all arches?
[13:39] <coreycb> cjwatson: I'll take a look at that, thanks for the suggestion
=== dilfridge is now known as Brother_Maynard
=== Brother_Maynard is now known as dilfridge
[19:19] <krytarik> So my discovery of a markup issue on the wiki yesterday was in context of creating a dark theme for every corner of Launchpad (for CGit I got a dedicated general one though), and while I can't be too sure yet there aren't some remaining bits it doesn't cover, I believe it'd be pretty much ready for deployment and wonder how the interest would be to add dark theme support to Launchpad proper?   ...
[19:19] <krytarik> ... It's currently in the form of a web browser userstyle and I've just pasted it to: http://paste.openstack.org/show/VjUumeFOOGBCd5h1wHja/
[19:21] <sarnold> oo cool
[19:25] <cjwatson> krytarik: Not opposed but I wonder how this would interact with the mid-term goal of converting Launchpad to Vanilla
[19:25] <cjwatson> I'd be a bit concerned about potentially putting more obstacles in the path of that, since it's already likely to be pretty complex - does Vanilla have dark theme support?
[19:26] <krytarik> Since I'm seeing ubuntu.com in all its usual bright glory at least, it doesn't seem so no. >_<
[19:29] <cjwatson> I think I'd at least want to know that there would be a forward path there, so it would be great if you or somebody could discuss this with Canonical's web team
[19:30] <cjwatson> We haven't started on a port there yet, and it's obviously a very large job, but it seems like part of the most likely path towards things like better mobile-friendlines
[19:30] <cjwatson> s
[19:30] <krytarik> And I wonder if the base theming gets changed to that, if that makes the dark theme CSS actually less complex to cover everything than how it is currently.. XD
[19:37] <krytarik> Got it, but adding dark theme support wouldn't really be a thing that can't be just trivially undone at any point, since one would just add an '@media (prefers-color-scheme: dark) {...}' rule for it and put the whole dark theme CSS in there.
[19:38] <krytarik> Whether directly or via includes or such, the latter of which we've done for a CGit instance.
[19:53] <cjwatson> Right, that's true, but I don't want it to be another potential regression that'd need to be overcome
[19:54] <krytarik> Right.
[19:54] <cjwatson> cgit may be easier since I doubt it would be vanillaized directly
[19:55] <krytarik> But then it'd look even less integrated than currently.. :3
[19:57] <cjwatson> It's true, but its integration is slightly a lost cause :)
[19:57] <cjwatson> We always at least half-planned to have an integrated code browser instead, but that's a decent-sized project and cgit was/is an adequate stopgap
[19:57] <cjwatson> But integration wouldn't start from cgit, I don't think
[20:00] <krytarik> Hahaha, I guess that's true indeed, and tbh I think CGit proper is the better solution for browsing the whole repo..
[20:01] <krytarik> https://loki.unit193.net/cgit/users/unit193/cgit-assets.git/ - is what we've got currently.
[20:03] <cjwatson> cgit's not awful in itself, certainly, but ultimately we'd like something that can be inline with the normal webapp
[20:03] <cjwatson> more github-like, if you will
[20:04] <cjwatson> Still, nothing concrete is planned there
[20:04] <krytarik> Yeah, I understand.
[20:05] <krytarik> Speaking of which, any plans for updating it to a somewhat current version?
[20:05] <cjwatson> cgit?  We just run what's in bionic
[20:06] <cjwatson> We should upgrade the git backend to focal at some point
[20:06] <cjwatson> That would come with an upgrade to a focal version of cgit along the way
[20:06] <krytarik> Yeah, and syntax highlighting is the thing I always miss the most on it.
[20:06] <cjwatson> I'm not sure whether any of my colleagues have looked into this as yet; I haven't
[20:07] <cjwatson> That doesn't need a newer version, it needs us to sort out apparmor confinement or something similar to mitigate risks of the highlighter getting out of hand
[20:07] <krytarik> Oh, ok.
[20:08] <cjwatson> And/or seccomp
[20:08] <cjwatson> Anyway, kids' bedtime
[20:08] <krytarik> Ok, see you!
[20:09] <cjwatson> very happy to help mentor somebody in things like adding confinement so we can enable syntax highlighting
[20:09] <cjwatson> just not quite right now :)