=== dps_ is now known as dps | ||
jibel | morning everyone | 05:14 |
---|---|---|
oSoMoN | good morning desktoppers | 06:52 |
didrocks | good morning | 07:05 |
jibel | salut didrocks & oSoMoN | 07:09 |
didrocks | salut jibel | 07:09 |
jibel | meh don't know how to use bzr anymore :/ | 07:19 |
didrocks | heh, using git too much :) | 07:28 |
didrocks | jibel: when you got a sec: https://github.com/ubuntu/zsys/pull/26 | 07:28 |
gitbot | ubuntu issue (Pull request) 26 in zsys "Fix many ineffectual assignments and typos" [Open] | 07:28 |
oSoMoN | salut jibel et didrocks | 07:31 |
didrocks | hey oSoMoN | 07:31 |
=== pstolowski|afk is now known as pstolowski | ||
jibel | didrocks, merged | 08:29 |
didrocks | thx! | 08:29 |
seb128 | gooood morning desktopers | 08:29 |
seb128 | salut les frenchies :-) | 08:30 |
didrocks | salut seb128 | 08:38 |
seb128 | lut didrocks, comment ça va aujourd'hui ? | 08:39 |
didrocks | ça va, et toi ? | 08:40 |
seb128 | ça va bien, entrain d'essayer de finir les formations :-/ | 08:44 |
marcustomlinson | good morning desktoppers | 08:50 |
didrocks | hey marcustomlinson | 08:51 |
oSoMoN | salut seb128 | 08:52 |
oSoMoN | good morning marcustomlinson | 08:52 |
seb128 | lut oSoMoN, hey marcustomlinson, how are you today? | 08:52 |
oSoMoN | feeling good, you? | 08:53 |
marcustomlinson | feeling better thanks | 08:53 |
seb128 | I'm good thx | 08:54 |
seb128 | brb, changing location | 08:56 |
=== jdstrand_ is now known as jdstrand | ||
seb128 | oSoMoN, just as a FYI, that librsvg build issue with rustcc 1.38 should be fixed by https://gitlab.gnome.org/GNOME/librsvg/commit/de26c4d8 | 15:34 |
seb128 | but it requires an update of the cssparser vendor copy included and I'm unsure how to properly do that | 15:34 |
seb128 | I downloaded a newer version, diffed the dir and added as a patch for testing but that fails because a .cargo-checksum is missing | 15:36 |
hellsworth | good morning folks | 15:39 |
oSoMoN | seb128, thanks, I'll take care of it | 15:39 |
seb128 | hey Heather! how are you today? | 15:40 |
oSoMoN | good morning hellsworth! | 15:40 |
seb128 | oSoMoN, thx, I'm not saying that it will fix the ppc64el issue though but let's see | 15:40 |
=== heather is now known as hellsworth | ||
=== pstolowski is now known as pstolowski|afk | ||
ijohnson | hi folks, I'm investigating slow startup performance in gnome snaps like gnome-calculator and I'm wondering what the purpose of the /usr/bin/gjs file being linked from the gnome-3-28-1804 content snap? | 20:28 |
hellsworth | hmm good question, let me go take a look at the content snap | 20:28 |
ijohnson | Specifically, I'm wondering: | 20:29 |
ijohnson | 1. can this file instead be located at /snap/gnome-calculator/current/usr/bin/gjs instead of /usr/bin/gjs? | 20:29 |
ijohnson | 2. that file in gnome 3.28 1804 content snap is a symlink to gjs-console, will it always be a symlink like that or could gjs point to something else like gjs-console2 ? | 20:29 |
ijohnson | thanks hellsworth | 20:30 |
hellsworth | where do you find gjs being linked? i mean where do you see this /usr/bin/gjs? | 20:31 |
ijohnson | hellsworth: it's in the gnome-3-28-1804 content snap at /snap/gnome-3-28-1804/current/usr/bin/gjs | 20:31 |
ijohnson | err that's the source | 20:32 |
ijohnson | it's being used in a layout in the gnome-calculator snap like this: | 20:32 |
ijohnson | https://www.irccloud.com/pastebin/ZfgDkUNp/ | 20:32 |
ijohnson | what that layout amounts to in the snap is that a symlink shows up at /usr/bin/gjs which points to `$SNAP/gnome-platform/usr/bin/gjs` | 20:34 |
ijohnson | (also sorry I had a typo in that snippet I pasted above, I updated it | 20:34 |
hellsworth | can you point me to the snapcraft yaml you're looking at? If I look at the gnome-calculator snapcraft in master and then do `snapcraft expand-extensions` there is no gjs mentioned: https://gitlab.gnome.org/GNOME/gnome-calculator/blob/master/snap/snapcraft.yaml | 20:38 |
hellsworth | so i'm wondering how gjs got in there | 20:38 |
sarnold | ijohnson: this is wild-guess territory as I don't know much about any of these tools -- my guess is this is to allow all the gnome snaps to share the same javascript interpreter, so that you don't have to have a dozen different JS interpreters loaded at once, but can share the memory for these processes (SHR column in top) | 20:39 |
hellsworth | hmm that would make sense | 20:40 |
ijohnson | sarnold: yes that makes sense, my question is more about why it needs to be mounted at /usr/bin/gjs and not at $SNAP/usr/bin/gjs or even just at $SNAP/gnome-platform/usr/bin/gjs | 20:42 |
ijohnson | hellsworth: I just ran `snap download gnome-calculator` and am looking at the file `meta/snap.yaml` from the snap | 20:42 |
ijohnson | unsquashed with `unsquashfs gnome-calculator_544.snap` | 20:43 |
sarnold | ijohnson: hmm. compounding my guesses at this point, in case something within the application tries to use the interpreter in a funny fashion? maybe libraries to spawn different things would call gjs directly through its well-known-location? | 20:43 |
ijohnson | sarnold: yeah that is usually why folks use layouts, but it's unfortunate here because this layout leads to 500-1000ms of snap-update-ns setting up various mounts | 20:44 |
ijohnson | hellsworth: indeed the snapcraft.yaml at that gitlab repo looks much different from what's on stabl in gnome-calculator indeed | 20:44 |
ijohnson | hellsworth: looks like gnome-calculator edge has the same as stable basically | 20:46 |
sarnold | ijohnson: ouch :( that's really painful for an interactive app. | 20:46 |
sarnold | not a big deal for IOT usecases, but .. ouch. | 20:46 |
ijohnson | sarnold: yes, hence why I'm looking into it :-| | 20:46 |
sarnold | ijohnson: maybe snapd could pre-create all the namespaces when snaps are installed or when snapd is started, and reuse those namespaces for execution? | 20:47 |
ijohnson | the namespaces do get re-used for subsequent launches, but for the first launch either right after install or first launch after boot | 20:47 |
ijohnson | for those cases we have to create these mounts | 20:48 |
hellsworth | well gjs is placed there because of this commit: https://github.com/snapcore/snapcraft/commit/55bfb2dfcf5f604ee2ebbf3970a4fd846b021771 | 20:52 |
hellsworth | i don't know why this was added tbh though | 20:54 |
ijohnson | hellsworth: thanks for that pointer, this at least gives me a test case I can look at for why the gjs binary is needed at all, because gnome-calculator doesn't seem to need it at all | 20:56 |
marcustomlinson | nice find hellsworth | 20:56 |
marcustomlinson | makes sense to me now why this doesn't make sense to me now :P | 20:57 |
hellsworth | lol | 20:57 |
sarnold | lol | 20:58 |
* ijohnson be back in a bit | 21:00 | |
hellsworth | https://github.com/snapcore/snapcraft/commit/55bfb2dfcf5f604ee2ebbf3970a4fd846b021771#r36058007 | 21:01 |
hellsworth | maybe the commiter can comment on why gjs was added. it looks very deliberate so there must have been some strong reason | 21:01 |
hellsworth | and that reason can be weighed with the 500-1000ms expense | 21:01 |
kenvandine | gnome-characters could use it | 21:06 |
kenvandine | But many other gjs apps | 21:06 |
kenvandine | It's for convenience | 21:07 |
kenvandine | Most gjs hard code the interpretor path for gjs | 21:07 |
kenvandine | So all gjs have to be patched without the layout | 21:08 |
kenvandine | Not just using the content snap | 21:08 |
kenvandine | Even trying to bundle gjs, they still needed patching | 21:08 |
kenvandine | ijohnson: ^^ | 21:09 |
ijohnson | kenvandine: hmm that's unfortunate, is this a script type situation like with python paths in the shebang where snapcraft could autonatically fix the location of gjs to somewhere else or even better just to search on $PATH? | 21:22 |
fidencio | people, we'd like to update the GNOME Boxes' recommended downloads so it shows the latest ubuntu release | 21:24 |
fidencio | however, in order to so, I really would need a review on https://gitlab.com/libosinfo/osinfo-db/merge_requests/54 | 21:24 |
gitbot | libosinfo issue (Merge request) 54 in osinfo-db "ubuntu: Add 19.10 entry" [Opened] | 21:24 |
fidencio | would be possible to have someone from canonical to take a look at that and tell me that I'm not missing something? | 21:24 |
hellsworth | jbicha: ^^ | 21:26 |
hellsworth | and didrocks isn't online atm | 21:26 |
fidencio | hellsworth: I've marked them on the MR when I opened it (a few weeks ago) | 21:27 |
hellsworth | i know but maybe they forgot so maybe it's time to ping them again :) | 21:27 |
fidencio | hellsworth: let's hope they see it (if not, I'll ping Didier Tomorrow on our timezone) | 21:28 |
hellsworth | good plan | 21:29 |
fidencio | hellsworth: thanks for the help! | 21:31 |
hellsworth | marcustomlinson: i finally got around to testing cherrytree with gtk2 stuff like gimp has and it doesn't look any different than it did when i was using the gnome extension | 22:24 |
hellsworth | https://github.com/hellsworth/cherrytree/tree/add-snap | 22:25 |
hellsworth | i really just don't know why the snap package looks different from the deb | 22:25 |
hellsworth | any thoughts you have are welcome :) | 22:26 |
marcustomlinson | hellsworth: have you tried deleting ~/snap/cherrytree and re running | 22:29 |
hellsworth | hmm no | 22:32 |
hellsworth | i should try that :) | 22:32 |
hellsworth | yeah even that didn't change things | 22:33 |
marcustomlinson | Interesting, I’m not at a computer right now, I can take a look tomorrow | 22:33 |
hellsworth | no rush. i could use your expertise when you have time :) | 22:35 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!