[00:01] <ItzSwirlz> Hey guys, is anyone able to assist with Bug #1887312
[00:01] <ubottu> bug 1887312 in cinnamon-session (Ubuntu) "[SRU] Cinnamon session doesn't shutdown until all apps close regardless if they act to Cinnamon's logout request" [Low,In progress] https://launchpad.net/bugs/1887312
[00:36] <ItzSwirlz> I guess everyone's sleeping right now, it's 8:36 PM for me. Goodnight /o
[11:27] <rbasak> teward, AsciiWolf: could teward maybe triage the bugs a little for torbrowser-launcher please - get rid of the "master bug", etc?
[11:27] <rbasak> "Do not create a meta-bug with a title like "Please SRU this" rather than using existing bug reports. It is redundant and is opaque to original bug reporters, whose feedback is valuable for verification, so these bugs will be invalidated by the SRU team and corresponding uploads will be rejected from the queue." from https://wiki.ubuntu.com/StableReleaseUpdates#Procedure
[11:28] <rbasak> eg. dupe things around - so there's just one bug per root cause left
[11:28] <rbasak> Feel free to adjust bug titles to match their root causes
[11:29] <rbasak> And then as Thomas said they're missing SRU information
[11:29] <rbasak> That's step 3 at https://wiki.ubuntu.com/StableReleaseUpdates#Procedure
[12:58] <AsciiWolf> rbasak, sorry, I was offline on IRC, it was just my bouncer online :)
[13:01] <teward> rbasak: easy to nix the master bug.  I can just close it.
[13:01] <rbasak> Sure
[13:01] <teward> the other bugs need the SRU template
[13:02] <AsciiWolf> rbasak, sure, teward, feel free to edit the original bug reports if there is anything wrong in them :) and also feel free to triage them
[13:02] <teward> then they can all be lumped together
[13:02] <AsciiWolf> I have already tested the new torbrowser-launcher build and everything seems to work fine now :)
[13:02] <teward> AsciiWolf: templates.  thats where me uploading this to proposrd and providing a diff for rbasak to look at is blocked
[13:02] <teward> asciiwolf: yep i saw your comment.
[13:04] <AsciiWolf> teward, if the SRU templates for other bugs are needed, could you please add them? I am quite busy today and will have to go away in ~15 minutes :/
[13:10] <AsciiWolf> teward, btw. there's a small typo in the changelog
[13:10] <AsciiWolf> "found in the Tor Browser launcehr package."
[13:11] <AsciiWolf> >launcehr
[13:20] <teward> @AsciiWolf: i can add the templates but will need to analyze regression risk
[13:20] <teward> i'll spellcheck before a final upload lol
[14:12] <teward> AsciiWolf: original bug retired, the remaining 4 got the template, it seems the issue is alos on Xenial per some bugs.
[14:12] <teward> s/Xenial/Bionic/
[14:12] <teward> ERR:NeedCoffee
[14:12] <teward> rbasak: ^ for awareness
[15:32] <AsciiWolf> teward, I have added a comment here regsrding not fixing the missing gnupg dependency: https://bugs.launchpad.net/ubuntu/+source/torbrowser-launcher/+bug/1897306/comments/2
[15:32] <ubottu> Launchpad bug 1897306 in torbrowser-launcher (Ubuntu) "[SRU] torbrowser-launcher has missing gnupg dependency" [Undecided,Fix released]
[15:32] <teward> rbasak: ^
[15:32] <teward> rbasak is higher than me on the decision tree ;)
[15:32] <teward> also
[15:33] <teward> >This issue happens on all custom (for example installed by netinst with custom package set) Ubuntu systems.< 'custom' sets of packages are sliiiiightly beyond the scope of standard SRU coverage
[15:33] <teward> so unless rbasak changes their mind in this case...
[15:34] <AsciiWolf> true, but this is still a torbrowser-launcher packaging bug
[15:34] <AsciiWolf> missing dependency on a core component that torbrowser-launcher uses and won't work at all without this component installed
[15:36] <AsciiWolf> What about the alternative solution that I suggested (not adding the patch, just adding "gnupg2" to debian/control? :)
[15:38] <teward> `gnupg2` doesn't exist as a package anymore - it's `gnupg` now
[15:38] <teward> it's a dummy transitional package though
[15:38] <teward> so... that could go away at any time
[15:39] <teward> thereby going back to the original fix of using the gnupg binary.  BUT again, if this is a 'nonstandard' installation edge fix, SRU might not want to include such an edge-fix
[15:39] <teward> (this is why I poked rbasak)
[15:41] <AsciiWolf> well, if gnupg2 will go away at some point in the 20.04 life cycle, torbrowser-launcher will stop working without the patch anyway
[15:42] <AsciiWolf> also, I wonder whether the gnupg2 package is even preinstalled on Ubuntu 20.04.1... I don't have any Ubuntu system/vm here at the moment, so can't check it
[15:43] <rbasak> "This issue happens on all custom (for example installed by netinst with custom package set) Ubuntu systems."
[15:43] <rbasak> I agree it's a bug.
[15:43] <rbasak> We fix bugs in the development release.
[15:43] <rbasak> For stable release updates, we try not to disrupt things.
[15:44] <rbasak> So they need further justification, such as more concrete use cases actually impacted
[15:44] <rbasak> Rather than theoretical impacts
[15:44] <rbasak> People using custom installation methods can just add the package they want.
[15:45] <rbasak> The workaround for them is trivial as they have a package list anyway
[15:45] <rbasak> AsciiWolf: ^ you're welcome to expand on what use cases are broken to change my mind
[15:45] <rbasak> But I need actual justifications please, rather than theoretical ones
[15:46] <rbasak> For example: find me someone who is actually impacted, so I can understand why the SRU is necessary for them
[15:46] <rbasak> (and how what they're doing is already broken)
[15:46] <rbasak> (and how what they're doing *isn't* already broken)
[15:47] <rbasak> I can phrase that better
[15:47] <rbasak> How they're *method* isn't already broken
[15:47] <rbasak> How their *method* isn't already broken
[15:47]  * rbasak isn't with it right now apparently. Sorry.
[15:53] <AsciiWolf> rbasak, are you sure that all of the official Ubuntu flavors have gnupg2 (that is, as teward said, "a dummy transitional package that can go away at any time") in their base installation?
[15:53] <AsciiWolf> and if you want actual example, I can give you one of my own use cases :)
[15:53] <teward> here's a question that's valid
[15:54] <teward> AsciiWolf: are you using Ubuntu or Mint?
[15:54] <AsciiWolf> both :)
[15:54] <AsciiWolf> neither of them on this machine, but I am using various distros
[15:54] <AsciiWolf> usually test things on a clean Ubuntu 20.04 VMs
[15:55] <AsciiWolf> but I am currently not at home and have an ancient mint desktop here and a Fedora laptop
[15:56] <AsciiWolf> anyway, to give you the "actual example":
[15:58] <AsciiWolf> I have Ubuntu Server installed on one of my machines at home and have xrdp installed there... it is a headless server, but I also use it as a "terminal server" for GUI apps
[15:59] <AsciiWolf> if I install torbrowser-launcher on this machine, it won't work
[16:00] <AsciiWolf> if I am not mistaken (but cannot check it now), there is no gnupg2 dummy package preinstalled on Ubuntu Server (at least on the minimal installation that I have) out-of-box
[16:02] <AsciiWolf> by the way, I am now downloading Xubuntu 20.04.1 to check whether the gnupg2 package is included out-of-box there :) will let you know...
[16:02] <teward> > ancient Mint
[16:02] <teward> there's problem #1 because that's ancient Ubuntu base ;)
[16:04] <AsciiWolf> no, the Mint installation is latest :)
[16:04] <AsciiWolf> I meant ancient PC, not ancient Mint ;)
[16:04] <AsciiWolf> Linux Mint 20 that I have there is based on Ubuntu 20.04.1
[16:18] <rbasak> AsciiWolf: in your response to your question, I don't know, but that's the way round we drive SRUs.
[16:19] <rbasak> We make a decision to apply an SRU (and take the consequent risk of user regression) on the basis of concrete problems reported that are impacting users that need fixing.
[16:20] <rbasak> If nobody tells us about a problem, then we assume it's safer not to risk regressing existing happy users and only fix it for the next release.
[16:20] <AsciiWolf> rbasak, I am downloading Xubuntu to test whether gnupg2 is preinstalled there and will also test some other official Ubuntu flavors today
[16:20] <rbasak> The reverse applies when considering regression risk of course
[16:21] <AsciiWolf> *Xubuntu 20.04.1
[16:21] <rbasak> AsciiWolf: sure. If you find that torbrowser-launcher is broken when using an Ubuntu flavour in its default configuration for example, that's SRU justification enough to fix it. If non-default then we'd have to consider the individual case.
[16:22] <rbasak> Correction: in your response to your question, I don't know, but that's *not* the way round we drive SRUs. I don't know what's with me today, sorry.
[16:22] <rbasak> My fingers don't seem to type what I'm thinking today
[16:23] <AsciiWolf> :)
[16:23] <rbasak> Note that a transitional package will stay for lifetime of a release.
[16:24] <AsciiWolf> rbasak, that's a good news, thanks
[16:26] <AsciiWolf> this means that the easiest (and still safe) way to fix this in Focal would be adding "gnupg2" to torbrowser-launcher's debian/control ;)
[16:27] <rbasak> Have you considered the regression risk associated with that?
[16:27] <rbasak> For example, does anything in the archive currently conflict with gnupg2?
[16:28] <rbasak> If not, what about all the things like that I haven't thought of right now?
[16:28] <rbasak> My point is just that it's not possible to think of everything :)
[16:29] <rbasak> https://xkcd.com/1172/ is relevant here :)
[16:30] <AsciiWolf> yeah, true
[16:52] <teward> rbasak: coffee.
[16:52] <teward> you need some :P
[16:53]  * rbasak can't stand the stuff
[17:01] <AsciiWolf> rbasak, teward, so, I have just installed Ubuntu 20.04.1 in vm...
[17:01] <AsciiWolf> from latest ubuntu-20.04.1-desktop-amd64.iso
[17:01] <teward> rbasak: would you prefer injectable caffeine?  ;)
[17:01] <teward> AsciiWolf: and?
[17:01] <AsciiWolf> and there's no gnupg2 package preinstalled as well
[17:01] <AsciiWolf> "Command 'gpg2' not found, but can be installed with:"
[17:02] <teward> so we need to just make it use standard gpg instead of gpg2
[17:02] <teward> that's trivial (and still the patch)
[17:02] <teward> rbasak: ^
[17:02] <teward> dep on gnupg: no
[17:02] <AsciiWolf> teward, ok, sounds fine :)
[17:02] <teward> patch to make it use /usr/bin/gpg instead of /usr/bin/gpg2
[17:02] <teward> yes
[17:02] <teward> IF that's the case
[17:02] <AsciiWolf> thanks!
[17:02] <teward> rbasak: ^ my analysis on that
[17:02] <teward> AsciiWolf: add this test documentation to the bug
[17:03] <teward> and make sure you reference Xubuntu as the test
[17:03] <teward> not Mint, etc.
[17:03] <AsciiWolf> ok :)
[17:03] <AsciiWolf> wait, I will also try Xubuntu 20.04.1 and Ubuntu Server
[17:04] <AsciiWolf> to be sure there's gpg (not gpg2) preinstalled there
[17:04] <AsciiWolf> but I think it will, I think apt depends on it
[17:05] <teward> ubuntu server doens't have a GUI
[17:05] <teward> AsciiWolf: yes I'm pretty sure gpg is there because apt, etc. need it
[17:07] <teward> i know server has gpg there because I deploy that heavily ;:)
[17:07] <teward> ;) *
[17:08] <AsciiWolf> :) ok
[17:19] <teward> AsciiWolf: you're not updating the bugs ;)
[17:19] <teward> so i did for you
[17:22] <AsciiWolf> oops, I was just writing it... did not notice that you already sent the update comment :D
[17:22] <AsciiWolf> sorry, I am a little bit slow... too much other things I work on at the moment
[17:24] <AsciiWolf> teward, should I edit the original bug description to reflect the changes?
[17:25] <teward> nah just add a comment
[17:25] <teward> that it affects xubuntu, etc.
[17:25] <teward> Mint doesn't matter in this case xD
[17:26] <AsciiWolf> well, it seems that it affects *all* flavors including main Ubuntu Desktop :)
[17:26] <teward> well except for an upgrade from 18.04 to 20.04 :P
[17:26] <teward> which has gpg2 on my end
[17:26] <AsciiWolf> yup
[17:30] <AsciiWolf> teward, I have added a new comment mentioning this
[17:35] <teward> i saw
[20:05] <teward> AsciiWolf: rbasak'll review when he can, he EOD'd already so it's on Robie's plate for now
[20:05] <teward> since he and I and you have already had in-depth discussions on that package