Thursday, 2025-12-04

*** croeland1 is now known as croelandt14:00
*** croelandt is now known as croeland114:48
*** croeland1 is now known as croelandt14:48
opendevreviewClark Boylan proposed openstack/project-config master: Update Jeepyb's Gerrit builds to Gerrit 3.11  https://review.opendev.org/c/openstack/project-config/+/96984619:43
sean-k-mooneyby the way we have been noticing strage behvior with gerrit comment not properly beign render and poping in if you referesh for at least the last few days19:53
fungisean-k-mooney: that could be a browser permissions problem, i think gerrit needs notify perms to immediately receive updates in an already open page19:54
sean-k-mooneyfungi: this is not for already opened pages19:56
sean-k-mooneythis is on opening a page for the first time19:56
fungioh, interesting19:56
clarkbsean-k-mooney: is that something you might be able to shrae a screenshot of? I'm not sure I fully understand19:57
sean-k-mooneyi was still wonderign if it was a caching issue19:57
fungii wonder if some rest api calls being made by the js frontend are failing silently. maybe a js console in the browser would show them if so19:57
clarkbis it basically an empty yellow comment box?19:57
clarkbbut yes the next step is likely opening the browser console to see if there was a failed request for data19:58
sean-k-mooneyno comment form previous revsiosn are not show at all19:58
sean-k-mooneyonly the comemnt form the current patchset19:58
sean-k-mooneythen you refresh and they are all there19:58
clarkboh got it. I think there is actually some algorithm there. Like it only shows unresolved comments from old patchsets by default19:58
sean-k-mooneyright its the unresloved ones im refering too19:59
clarkband then I think if you drop a file from the change you can only see those comments by loading the old patchset (this might be a bug)19:59
clarkbsean-k-mooney: do you have an example change? I'm curious if I can reproduce it in one of the three browsers I'm currently running19:59
clarkbbut yes, if a refresh fixes it then a likely problem is failed network requests20:00
clarkband I think those should show up in your browser console log20:00
sean-k-mooneyi was seeing it on https://review.opendev.org/c/openstack/nova-specs/+/967712/9/specs/2026.1/approved/resource-notifications.rst bu ti have resove most of the older commeents at thsi point 20:00
clarkbthe comments from ps7 and ps8 appear to load for me on initial load (ps9 is latest) in firefox.20:01
clarkbthough I guess I don't know how many there should be just that there are non zero quantities20:01
clarkbchrome and chromium seem to match firefox.20:01
clarkbIf you see it happen again can you check your console and share any tracebacks if they exist? And maybe also see if anyone can reproduce for that change? if it is a cache issue I would expect it to be global for a specific change20:02
clarkbsince there aren't user specific caches for things like that20:02
sean-k-mooneyyes they loaded for me as weel this time 20:03
sean-k-mooneyalthough 8 https://review.opendev.org/c/openstack/nova-specs/+/967712/8/specs/2026.1/approved/resource-notifications.rst20:04
sean-k-mooneyhas no comment form 7 when i loaded it just now20:04
sean-k-mooneyv9 has comemnt form v8 and v720:04
sean-k-mooneyso v8 shoudl have the coment form v7 too20:05
sean-k-mooneythat was repatable for me and then i refest v8 and now the v7 comment are thre too20:06
sean-k-mooneyand if i go back go v8 they are missing again20:07
sean-k-mooneyhttps://imgur.com/a/sZxE4Hg20:11
clarkbthe two commetn chains from ps7 loaded for me in firefox when I pulled up the ps8 link above20:11
clarkbI have lunch now but can look closer after20:11
sean-k-mooneyack no worries20:12
clarkbanything in your browser console?20:12
sean-k-mooneyit actully yes20:13
sean-k-mooneythre is a lot of output form somethign call gr-app.js20:14
sean-k-mooneyand some warning about preloading fonts20:14
sean-k-mooney"The resource at “https://review.opendev.org/changes/?q=change:967712” preloaded with link preload was not used within a few seconds. Make sure all attributes of the preload tag are set correctly"20:14
clarkbwhile I could see font loading affectin comments I would be surprised if any comments load at all20:22
clarkblet me see if I get that message (fwiw I have a long standing change upstream to not use web fonts but now I don't have a valid cla with google...)20:22
sean-k-mooneywell i think the font message is just saying this was preloaded but not required20:23
sean-k-mooneyor similar so i dont think thats is realted20:23
sean-k-mooneyso on firefox i see lots of debug logs form gr-app.js which i assume is gerrti javascrip component20:24
clarkbya I get the same thing in my console where the comments have loaded20:24
clarkbyes I believe that is the part of gerrit that acts like a single page app or whatever20:26
sean-k-mooneyya it odd i can get back to the "missing v7" state just by uing the drop down to go form v9 to v820:26
clarkbsean-k-mooney: let me try that20:27
sean-k-mooneyi wonder are they trying to be fancy and not actully reirect the page20:27
clarkbyes that reproduces20:27
clarkbso loading the expected patchset directly seems to work fine20:27
clarkbbut using the in diff view patchset selector seems to lose some comments.20:28
sean-k-mooneyyep20:28
clarkband doing that produces no tracebacks in the console20:28
clarkbsean-k-mooney: I don't know what is happening and without tracebacks we probably have to go and read and understand some js/ts code. Probably best to start with a bug report. Is that something you want to do or would you prefer I do it? I think due to spam there might be initial submission issues for new users so may be best if I do it20:30
clarkbsean-k-mooney: it looks like if you view the file at the top level of the chagn and then change patchsets there that comments render properly in that view20:33
clarkbsean-k-mooney: so it seems like this could be specific to the file view. But also ^ may be a workaround?20:33
clarkb(I personally prefer view things in distinct file views as I think it helps me organize better, but I know top level view is something people like too so maybe its good enough for now)20:34
clarkbI'll go ahead and do my best to file an issue when I'm back from todays school run21:05
sean-k-mooneyclarkb: thanks its normally not an isseu21:08
sean-k-mooneywe were tryign to do a spec freeze today21:09
sean-k-mooneyso i kept jumpin gbetween revisoins of sepc to adress diffent comments21:09
sean-k-mooneyhitting F5 or ctrl-R whatever refhest is bound too works find too21:10
clarkbsean-k-mooney: I spent a bit more time fiddling with it to try and write a complete bug report and now I can't reproduce anymore22:52
clarkbsean-k-mooney: I think what was happenign was the line numbers were shifting and I wasn't checking comment content properly but when I do what we thought reproduced it the comments are all there now22:53
clarkbthe first moves from line 206 to 213 and the other from 226 to 23222:53
clarkbbased on your screenshot it does seem like something is happening but I don't think this is a reliable reproduction case22:54
clarkbI'm going to hold off on filing a bug to avoid leading gerrit down some dead end path. If you think you see it again can we double check the console log again?22:55
clarkbmostly I'm just hoping that if it isn't a consistent problem then finding some error would be really helpful (if there is one anyway)22:55

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!