[09:38] <alan_g[m]> Another pile of USN notifications. Is it my turn yet?
[09:38] <Saviq> Nah, all(most) done
[10:21] <Saviq> alan_g FYI, I may need to stop the CVE rebuilds pending some explanation from the security team - most of the CVEs we got notice of are only fixed in [ESM](https://ubuntu.com/security/esm), so we're in a rebuild loop on those.
[12:10] <alan_g[m]> I'm looking at the naming of displays within the display configuration and also in the wl-output v4 proposal and think we're starting from the wrong place...
[12:10] <alan_g[m]> Our configuration names outputs by numbering within type on each card independently, but that means we can have, for example, "HDMI-A-2" on two different cards.
[12:10] <alan_g[m]> The wl-output proposal currently creates unique names by ignoring the card and type and using a global sequence number.
[12:10] <alan_g[m]> None of these names are stable as cards may be detected in different orders and and while I've not seen the outputs on a card appear in different orders I don't think there is any guarantee
[12:12] <Saviq> That's what I asked RAOF (he/they) about recently and IIUC at least with KMS they should be deterministic?
[12:12] <Saviq> Ultimately, at least for display configuration, we're more likely to want to match EDID data (particularly, serial#) than outputs, don't you think?
[12:16] <alan_g[m]> I don't think that was the full answer (I think he's seen cards come up in different orders). And remember physical outputs can be moved between cards and physical cards can be moved between slots
[12:16] <Saviq> > physical outputs can be moved between cards
[12:16] <Saviq> ¿?
[12:17] <alan_g[m]> But yeah, we need a more stable identifier for config
[12:18] <alan_g[m]> I mean we can unplug a monitor cable and stick it in another socket
[12:18] <Saviq> > I think he's seen cards come up in different orders
[12:18] <Saviq> He conceded that was probably differet USB-C ports
[12:19] <Saviq> > I mean we can unplug a monitor cable and stick it in another socket
[12:19] <Saviq> Right, which matching by EDID data would help with indeed.
[12:26] <alan_g[m]> So, yes, a new configuration key would make sense. But that likely breaks the current support. And doesn't give good names for wl-output v4 (which are admittedly not supposed to be taken seriously)
[12:29] <Saviq> I don't think it necessarily breaks the current support.
[12:29] <Saviq> We have `cards:`, and that could continue in the way it exists, possibly with better output names (which could break, but we could also keep backwards compat), and matching by more properties.
[12:29] <Saviq> We could have `screens` next to `cards`, which could match by EDID, whichever card/output it was connected to.
[12:31] <alan_g[m]> Yeah, it could go alongside and we just have to figure which is the truth. (I feel a utility to validate the config file coming on)
[12:32] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq closed [issue #123](https://github.com/MirServer/ubuntu-frame/issues/123): Hiding cursor doesn't work?
[12:37] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq closed [issue #75](https://github.com/MirServer/ubuntu-frame/issues/75): Is there any way to run chromium on Ubuntu Core?
[12:38] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq edited [pull request #122](https://github.com/MirServer/ubuntu-frame/pull/122): Automate reloading display configuration
[14:19] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths opened [pull request #2814](https://github.com/MirServer/mir/pull/2814): [miral] refactor YamlFileDisplayConfig::apply_to()... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/eabbe2b2a3007d45902b67c2727fa24715cb798d>)
[16:02] <Saviq> alan_g on `snap set wrong-key=`…, do you recall if `snapctl get` will work ;)
[16:02] <alan_g[m]> You have to specify a key
[16:03] <alan_g[m]> There's no analog of snap get snapname
[16:03] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** bors[bot] edited [pull request #122](https://github.com/MirServer/ubuntu-frame/pull/122): [Merged by Bors] - Automate reloading display configuration
[16:03] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** bors[bot] closed [pull request #122](https://github.com/MirServer/ubuntu-frame/pull/122): [Merged by Bors] - Automate reloading display configuration
[16:03] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** bors[bot] closed [issue #20](https://github.com/MirServer/ubuntu-frame/issues/20): Difficult to bootstrap the `display` config option
[16:03] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2814](https://github.com/MirServer/mir/pull/2814): [miral] refactor YamlFileDisplayConfig::apply_to()
[16:40] <Saviq> The ";)" 👆️ was muscle memory… It was supposed to be a question mark 🤦‍♂️
[16:45] <alan_g[m]> Don't worry. That was my reading
[16:47] -GitHub[m]:#mir-server- **[MirServer/mir]** graysonguarino opened [issue #2815](https://github.com/MirServer/mir/issues/2815): Need a better display naming solution
[16:47] -GitHub[m]:#mir-server-  
[16:47] -GitHub[m]:#mir-server- > Looking into #2809 brought `src/common/output_type_names.cpp` to my attention. We need a better solution for reporting display names in wl_output. There is a TODO from 6 years ago talking about the need to introduce type safety.
[16:58] <Saviq> Good weekend all o/
[16:59] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths edited [issue #2815](https://github.com/MirServer/mir/issues/2815): Need a better display output naming solution
[17:00] <alan_g[m]> o/
[17:13] -GitHub[m]:#mir-server- **[MirServer/wlcs]** AlanGriffiths opened [issue #266](https://github.com/MirServer/wlcs/issues/266): Test wl_output features up to version 4
[17:13] -GitHub[m]:#mir-server-  
[17:13] -GitHub[m]:#mir-server- > We've updated Mir, we should have tests in wlcs
[18:12] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2809](https://github.com/MirServer/mir/pull/2809): Add wl_output version 4
[18:12] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] closed [issue #2792](https://github.com/MirServer/mir/issues/2792): [23.04] swaybg requires a newer version of wl-output than we support