Saviq | Morn' | 07:05 |
---|---|---|
alan_g[m] | <insert time of your day>! | 07:52 |
Saviq | ? | 07:53 |
alan_g[m] | s/</\</, s/>!/\>!/ | 07:53 |
-GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths opened [pull request #2546](https://github.com/MirServer/mir/pull/2546): [miral] Tidy up event filtering API... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/c6a75662377ac97a1112e1e926c5f526c414bbd6) | 09:33 | |
alan_g[m] | Saviq want to review and merge a suggestion? IMO we can then land the v8 stuff: https://github.com/MirServer/mir/pull/2508#discussion_r943319792 | 10:24 |
Saviq | Sure | 10:26 |
alan_g[m] | Oh, and this ought to be easy: https://github.com/MirServer/mir/pull/2545 (failure on Debian SID is "not us") | 10:43 |
Saviq | alan_g why is it legacy? mirclient, or got fixed in the mean time? | 10:53 |
Saviq | And if it wasn't fixed, we'd see failures in CI? | 10:53 |
alan_g[m] | Was workaround for a race using mirclient | 10:54 |
-GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2545](https://github.com/MirServer/mir/pull/2545): Delete obsolete workaround | 11:23 | |
-GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2508](https://github.com/MirServer/mir/pull/2508): Bump wl_seat to v8 and implement hi-res scrolling | 11:58 | |
-GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] closed [issue #2499](https://github.com/MirServer/mir/issues/2499): Upgrade to wl_seat v7 | 11:58 | |
alan_g[m] | I got MultiThreadedCompositor to crash locally, which gives a chance of finding a cause for its flakiness. \o/ | 13:52 |
alan_g[m] | Why does it need to be so hard to find out why I don't have a core file. I remember when it was just there waiting to be looked at. Now I havwe to track down a log entry: `ERROR: apport (pid 1189810) Thu Aug 11 15:05:00 2022: host pid 1189741 crashed in a separate mount namespace, ignoring` | 14:15 |
alan_g[m] | (Naturally, it works perfectly under a debugger) | 14:15 |
alan_g[m] | This is what makes software such fun! It crashes reasonably often when run from the CLion snap. But then apport eats my core. Run from the terminal it isn't crashing. | 14:27 |
alan_g[m] | Who decided I wanted apport anyway... | 14:28 |
Saviq | Does it not drop it in `/var/crash`? | 14:35 |
alan_g[m] | It did not. I finally got it to crash outside a namespace and it landed in `/var/lib/apport/coredump/` | 14:36 |
alan_g[m] | What I really want is "Please, before I cleanup the process would you like me to attach gdb and download symbols for any dependencies"? | 14:38 |
Saviq | Something going that direction was recently presented at one of the sprints | 14:39 |
alan_g[m] | Something like that was in Fedora when I was testing Mir there a few years ago | 14:40 |
Saviq | Will try and find a reference | 14:40 |
alan_g[m] | I don't want a reference (I saw it too), I want it an hour ago | 14:41 |
Saviq | https://wiki.debian.org/Debuginfod anyway | 14:41 |
Saviq | sophie I rebuilt Mir in the armhf environment with ccache, so it should behave better now. Using `-j4` is likely to help as well | 15:01 |
sophie-w | Nice, I have some suspicions as to what's going wrong, but it's really weird that main also seems to be broken in that container. | 15:03 |
sophie-w | Update: changing the name of our mmap wrapper (so real mmap is used instead) fixes the failing tests. It's definitely something to do with that. | 16:12 |
alan_g[m] | In the past, things like that have turned out to signatures being different on different targets | 16:14 |
Saviq | Good evening all o/ | 16:18 |
sophie-w | This looks suspiciously similar, however when I print `dlInfo.dli_fname` I get `/lib/arm-linux-gnueabihf/libc.so.6` (which seems correct), so maybe that's not the problem. https://optumsoft.com/dangers-of-using-dlsym-with-rtld_next/ | 16:28 |
-GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths opened [issue #2547](https://github.com/MirServer/mir/issues/2547): MultiThreadedCompositor.when_compositor_thread_fails_start_reports_error is flaky... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/bd8bfb7c511f0171021bb34aea49a50dc9ca5d30) | 16:35 | |
-GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths assigned AlanGriffiths to [issue #2547](https://github.com/MirServer/mir/issues/2547): MultiThreadedCompositor.when_compositor_thread_fails_start_reports_error is flaky | 16:35 | |
-GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** graysonguarino marked [pull request #80](https://github.com/MirServer/ubuntu-frame/pull/80): Add crash reporter as ready for review | 16:49 | |
-GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths drafted [pull request #2548](https://github.com/MirServer/mir/pull/2548): Don't destroy CompositingFunctors that are not ready | 17:01 | |
-GitHub[m]:#mir-server- | 17:01 | |
-GitHub[m]:#mir-server- > Fixes: #2547 | 17:01 | |
grayson-g[m] | <GitHub[m]> "**[MirServer/ubuntu-frame..." <- Regarding this, I'm not *totally* sure if `draw()`ing in `CrashReporter::Self::draw_on_diag_update()` is wise. I am writing to the buffer from two different threads, but it `draw()`s under a lock, so that isn't a problem, right? | 17:16 |
grayson-g[m] | > <@github:maunium.net> **[MirServer/ubuntu-frame]** graysonguarino marked [pull request #80](https://github.com/MirServer/ubuntu-frame/pull/80): Add crash reporter as ready for review | 17:17 |
grayson-g[m] | * Regarding this, I'm not _totally_ sure if `draw()`ing in `CrashReporter::Self::draw_on_diag_update()` is wise. I am writing to the buffer from two different threads, but it `draw()`s under a lock, so that isn't a problem, right? It's worked properly in all my testing, but I'm not sure if this could become an issue. | 17:17 |
grayson-g[m] | Moving that back to a draft because I wasn't thinking about the snap when I tested and I got a snap-related bug (plus I screwed up the merge). | 18:12 |
-GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** graysonguarino marked [pull request #80](https://github.com/MirServer/ubuntu-frame/pull/80): Add crash reporter as ready for review | 20:51 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!