| jbicha | sil2100: I'm doing a few more Desktop tests. Thanks for pinging | 00:03 |
|---|---|---|
| sil2100 | jbicha: thanks! | 00:05 |
| didrocks | good morning | 06:26 |
| lissyx | FTR gnome-shell seems to be leaking memory over time on 22.04 | 06:37 |
| lissyx | 3.7% reported in htop, where it was much less at start of the session | 06:37 |
| lissyx | a few week ago it was hittinng 5.5%, making gnome-shell unusable (high latencies) | 06:38 |
| lissyx | 3.7% ~= 1.2GiB | 06:38 |
| lissyx | 1.1% of memory at session start | 06:41 |
| duflu | Hi didrocks and lissyx | 06:50 |
| duflu | lissyx, if your leak is not already covered by one of these: https://bugs.launchpad.net/ubuntu/+bugs?field.tag=gnome-shell-leak then please open a new bug (but not before eliminating all thirdparty extensions) | 06:51 |
| lissyx | I only have "Lock keys" enabled | 06:51 |
| duflu | Please delete that, then verify the leak happens, then log a new bug | 06:52 |
| lissyx | https://bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/1983161 | 06:53 |
| ubottu | Launchpad bug 1983161 in gnome-shell (Ubuntu) "gnome-shell memory leak" [Undecided, Incomplete] | 06:53 |
| lissyx | I have the same appindicator stuff | 06:53 |
| lissyx | on solaar though | 06:53 |
| duflu | Ah please don't use any Incomplete bugs -- they're almost certainly all bogus | 06:54 |
| duflu | (caused by extensions usually) | 06:55 |
| lissyx | well the others dont really match ... | 06:56 |
| duflu | No problem. Please delete all extensions, log in again and then when the leak happens log a new bug (ubuntu-bug gnome-shell) | 06:57 |
| duflu | lissyx, I guess my first question should be: are you using Intel 4th gen or older? | 06:58 |
| seb128 | goood morning desktopers | 07:04 |
| duflu | Hi seb128 | 07:04 |
| seb128 | hey duflu, how are you? | 07:10 |
| duflu | seb128, another weather change another sinus headache. Not just me but other people too... How are you? | 07:11 |
| seb128 | duflu, I'm alright thanks! | 07:18 |
| seb128 | :-( for the weather changes and sinus headache | 07:19 |
| oSoMoN | good morning desktoppers | 07:20 |
| duflu | Hi oSoMoN | 07:22 |
| oSoMoN | hey duflu | 07:23 |
| lissyx | duflu, no | 07:24 |
| lissyx | duflu, I disabled all extensions except Lock Keys weeks ago, and restarted the session three times since | 07:24 |
| duflu | lissyx, the next question is: are you using the screenshot feature? It still leaks memory | 07:25 |
| lissyx | AMD Ryzen 7 PRO 5850U is my GPU | 07:25 |
| lissyx | I did take a few screenshots | 07:25 |
| lissyx | but nothing that would explain growing from 1.5% to 4% of memory? | 07:25 |
| duflu | Yeah that leaks reliably, although much less since the latest update | 07:25 |
| lissyx | 4% is 1.28GiB | 07:25 |
| duflu | Yeah repeated screenshots can reach that without too much effort | 07:26 |
| duflu | Still, that may not be the main cause here | 07:26 |
| lissyx | honestly, with out without extensions, I have always suffered some leaks on gnome-shell for months ... if not years | 07:27 |
| lissyx | commute time | 07:28 |
| duflu | lissyx, the appindicators theory hasn't been verified yet. It would be helpful if you could disable that extension and see if the leaks stop | 07:31 |
| * lool o/ | 08:56 | |
| duflu | Hi lool | 09:00 |
| lool | Hey duflu! | 09:00 |
| lool | jbicha, seb128: Heya! You remember these GStreamer SRUs we worked on in Frankfurt? I think I just got a message about one of them – "[focal/gstreamer1.0] Package Phasing Stopped" | 09:01 |
| lool | I hadn't ever gotten this notice, it's pretty cool! but looking at the specifics I couldn't quite figure where was the bump in errors | 09:02 |
| lool | Looking at https://people.canonical.com/~ubuntu-archive/phased-updates.html it seems that gstreamer1.0 is halted | 09:03 |
| lool | I wonder if this package is going to collect all crashes with weird/out of Ubuntu plugins that might trigger some gst-inspect crashes | 09:03 |
| lool | At least, plugin scanning crashes seem to be the top crashers in https://errors.ubuntu.com/?release=Ubuntu%2020.04&package=gstreamer1.0&period=week&version=1.16.3-0ubuntu1 | 09:04 |
| seb128 | lool, hey. Those metrics are not really reliable, we are getting low number of reports so it's probably a false positive or in the noise margin. We could contact the SRU asking to ignore it or just wait, those cases often autoresolve, | 09:06 |
| lool | So give it a few days? I can do that | 09:06 |
| seb128 | yes | 09:09 |
| lool | leaving the frying pan soak in the kitchen sink? CAN DO! | 09:38 |
| jbicha | lissyx: there is supposed to be a small memleak fix in gjs 1.72.1 but it hasn't made it into jammy yet | 10:56 |
| jbicha | good morning | 10:56 |
| nteodosio | good morning jbicha | 11:00 |
| oSoMoN | good morning jbicha & nteodosio | 11:28 |
| lissyx | jbicha, is there some tooling that could be used to track that? | 11:53 |
| jbicha | lissyx: I don't understand your question | 11:59 |
| lissyx | jbicha, gjs is built from v8 right, is it possible to connect it to a debugger/profiler and maybe get hints on what could be leaking ? | 12:05 |
| jbicha | gjs uses the SpiderMonkey JavaScript engine from Firefox ESR (packaged as mozjs*) | 12:06 |
| lissyx | oh | 12:06 |
| lissyx | even better ... | 12:07 |
| jbicha | yes, there is some way to debug & profile but that's outside my expertise | 12:07 |
| lissyx | fortunately, I have paris-based colleagues working explicitely on the js engine :) | 12:07 |
| jbicha | you could also talk to gnome-shell upstream to ask what they use | 12:08 |
| lissyx | I guess my past experience makes me a bit relunctant to interact with gnome-shell upstream | 12:09 |
| lissyx | oSoMoN, \o/ https://github.com/canonical/firefox-snap/runs/7663824243?check_suite_focus=true#step:5:33937 | 13:19 |
| oSoMoN | lissyx, yes, that was successful for two nightly builds already :) | 13:21 |
| lissyx | and some beta | 13:21 |
| lissyx | https://launchpad.net/~mozilla-snaps/firefox/+snap/firefox-snap-beta | 13:21 |
| oSoMoN | and I can view the symbols uploaded with my token, the web interface is neat | 13:21 |
| lissyx | but those dont produce debug symbols? | 13:21 |
| oSoMoN | not yet | 13:22 |
| lissyx | I guess you still need to merge the changes to beta/stable ? | 13:22 |
| oSoMoN | yes | 13:22 |
| oSoMoN | not merged yet | 13:22 |
| lissyx | have you issued some gnome-sdk rebuild ? | 13:22 |
| lissyx | it's the last bit where we might have a bug on our side to pull from launchpad | 13:22 |
| lissyx | (based on how builds are discovered) | 13:22 |
| lissyx | jbicha, I'm being told perf is the best way | 13:28 |
| lissyx | oh | 13:34 |
| lissyx | https://launchpad.net/~desktop-snappers/+snap/gnome-3-38-2004-sdk/+build/1839135 | 13:34 |
| lissyx | https://treeherder.mozilla.org/jobs?repo=mozilla-central&searchStr=snap&selectedTaskRun=Wkv8-QcXQVuOnl9bqxjnSQ.0 | 13:37 |
| lissyx | https://crash-stats.mozilla.org/report/index/a04aead9-c47d-4191-9d8a-14e7d0220803 | 13:49 |
| lissyx | https://crash-stats.mozilla.org/report/index/f649e87d-788c-463f-8bc2-bd2b50220804#tab-details | 13:49 |
| lissyx | crashes from snap with symbols :) | 13:49 |
| oSoMoN | lissyx, very nice! | 15:15 |
| bittin | https://www.phoronix.com/news/Ubuntu-22.04.1-LTS-Delayed | 16:55 |
| KGB-2 | geocode-glib signed tags 44dcb10 Jeremy Bicha upstream/3.26.3 * Upstream version 3.26.3 * https://deb.li/iWJAv | 20:52 |
| KGB-2 | geocode-glib pristine-tar 90a5d5d Jeremy Bicha geocode-glib_3.26.3.orig.tar.xz.delta geocode-glib_3.26.3.orig.tar.xz.id * pristine-tar data for geocode-glib_3.26.3.orig.tar.xz * https://deb.li/0OH2 | 20:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!