← Previous day | Today | Next day → | Search | Index
All times shown according to UTC.
Time | Nick | Message |
---|---|---|
00:03 | lukeg joined #koha | |
07:37 | reiveune joined #koha | |
07:37 | reiveune | hello |
07:37 | thibaud_g joined #koha | |
07:53 | alex_ joined #koha | |
07:53 | alex_ | bonjour |
07:53 | wahanui | hi, alex_ |
08:09 | thibaud_g | hello all |
08:37 | thibaud_g joined #koha | |
09:23 | philor joined #koha | |
09:36 | alex_ joined #koha | |
11:22 | oleonard | o/ |
11:33 | tcohen | hola #koha o/ |
11:33 | cait[m] | \o |
12:06 | oleonard | https://snipboard.io/AMcxhn.jpg |
12:09 | cait[m] | lol |
12:21 | tcohen | hahaha |
12:22 | oleonard | Why do I have to RECHOWN my clone every time I want to run the qa tool? |
12:23 | cait[m] | finally someone else is asking this! |
12:23 | tcohen | oleonard: because you're running it as root inside a container? |
12:24 | cait[m] | nope |
12:24 | tcohen | do you have |
12:24 | export LOCAL_USER_ID=$(id -u) | |
12:25 | in your bashrc | |
12:25 | cait[m] | what I mean is that I still see that beaviour in hte most recent where you cant run it outside kshell |
12:25 | tcohen | my bet is your host OS has more than user configured, and yours is not the first one created |
12:26 | and thus doesn't have the uid == 1000 | |
12:26 | it would be interesting to see the permissions for the conflicted files before you chown | |
12:26 | I can help with that info | |
12:26 | cait[m] | i have that line |
12:27 | tcohen | it is not common for us to introduce bugs LOL |
12:27 | but if you give me more info we can track it | |
12:27 | cait[m] | I'll check on them next time i thappens |
12:27 | but maybe oleonard is faster? | |
12:28 | tcohen | at some point I've been carrying permissions issues for days before noticing |
12:28 | oleonard | If I "echo $(id -u)" in my macOS terminal it says 501 |
12:28 | tcohen | that's fine, oleonard |
12:28 | as long as the LOCAL_USER_ID variable is set | |
12:29 | my mac is upgrading, but can check in a bit | |
12:29 | I use it for the RM tasks often | |
12:29 | with no issues, so there's hope :-D | |
12:29 | cait[m] | 1001 for me |
12:30 | echo $LOCAL_USER_ID says 1001 too | |
12:32 | alex_ joined #koha | |
12:34 | oleonard | Looks like my files are a mix of kohadev-koha and root |
12:49 | cait[m] | tcohen: ^ |
12:50 | tcohen | yeah, that's known |
12:50 | running things as root yields that | |
12:50 | yarn, git, qa, etc | |
12:51 | all should be run as the instance user | |
12:51 | if there's a pattern you're following that yields that situation, file an issue on KTD | |
12:51 | and maybe we find a way to make it work | |
12:52 | for node deps | |
12:52 | we set a separate dir, for example | |
12:52 | cait[m] | so i should run yarn buld from kshell? |
12:52 | tcohen | in /kohadevbox/node_modules, and commands were adjusted to use that |
12:53 | only admin commands should be run as root | |
12:53 | cait[m] | so what is an admin command? |
12:53 | tcohen | the kohadevbox used the vagrant user as the default |
12:54 | koha-plack --restart kohadev? | |
12:54 | cait[m] | ok, all the koha- commands |
12:54 | tcohen | it is easy, anything that touches files will probably generate files with the current user as owner |
12:55 | if it is the root user, it gets problematic outside the container, because most of us don't use the root user | |
12:55 | for our regular tasks | |
12:55 | so, avoid the root user inside the container | |
12:55 | cait[m] | hm |
12:55 | git bz | |
12:56 | tcohen | ashimema proposed we somehow made ktd --shell jump directly into kshell |
12:56 | cait[m] | i use that inside the container... it sets it up for me, it's nice |
12:56 | when you are inside kshell you don#t see what branch you are on | |
12:56 | etc. | |
12:56 | running git ocmmands in there is not nice right now | |
12:56 | also git so doesn't set the right user etc. | |
12:57 | or doesn't even work I think | |
12:57 | tcohen | we could add a nicer prompt in the instance shell |
12:58 | but those permissions issues are related to running things as root | |
12:58 | so the instance user cannot manipulate the git db | |
12:58 | I'm not sure about git bz | |
12:58 | if it is set correctly or not | |
12:58 | cait[m] | I seem to remember it doesn't work |
12:58 | * tcohen | tries |
12:59 | tcohen | nope, it is not set |
12:59 | file an issue | |
12:59 | cait[m] | try git so - it won't work either |
12:59 | if we are supposed to run things inside kshell there is a lot left to do | |
13:00 | tcohen | I only do qa -c XX --run-tests inside the container |
13:00 | all my work I do it on the host OS | |
13:00 | cait[m] | then you have to set up more tools |
13:01 | like git bz and hooks | |
13:01 | if we aim to provide a dev env that works out of the box, that doesn't make sense to me | |
13:01 | tcohen | hooks are set at the clone level |
13:01 | cait[m] | i mean git so |
13:01 | sorry | |
13:01 | alias? | |
13:01 | wahanui | alias are ok |
13:01 | tcohen | I understand |
13:01 | cait[m] | so what I am going to file? |
13:01 | make git bz work in kshell for a start? | |
13:02 | make git so work in kshell? | |
13:02 | tcohen | maybe |
13:02 | yes | |
13:02 | we just need to c&p what we do for the root user, I guess | |
13:02 | or just move it to avoid confusing devs | |
13:02 | and we need help with the docs | |
13:03 | but I think I covered this topic to some extent | |
13:04 | cait[m] | hope i said it right: https://gitlab.com/koha-commun[…]cker/-/issues/331 |
13:05 | tcohen | does anyone have marcel's phone nomber? |
13:05 | *number | |
13:06 | cait[m] | sorry pass |
13:06 | tcohen | we have a problem with his patches, when applying to 21.11 |
13:31 | oleonard | tcohen: You say avoid the root user inside the container, but what does that mean? I have to use kshell to run the qa tool, and I apparently have to chown everything to root to get the qa tool to work |
13:32 | cait[m] | it usually all works for me if I own it all to my user outside of ktd |
13:32 | but have to do that oftne | |
13:34 | oleonard | cait[m]: When I own it all to my user outside of ktd that results in the files being owned by root inside ktd. Is that expected? |
13:35 | cait[m] | maybe? I am not sure about those permission things ,just sharing what works for me so far and how I was told to do it |
13:36 | oleonard | Yeah it sounds like we're doing the same thing... chowning everything often |
13:36 | cait[m] | possibly |
13:36 | i filed the 2 issues as discussed earlier - maybe switching it all to kshell would not be too bad | |
13:38 | if that helped the permissions and allowed access to all needed scripts/tools | |
13:57 | marie-luce joined #koha | |
14:08 | oleonard | Surely there is a way to globally set things like [% USE raw %] in the templates |
14:12 | Yep, of course there is. | |
14:18 | Oh wait I'm misreading... | |
15:01 | lukeg joined #koha | |
15:03 | oleonard | tcohen: Master needs CSS to be rebuilt |
15:24 | Dyrcona joined #koha | |
16:00 | bag joined #koha | |
16:09 | reiveune | bye |
16:09 | reiveune left #koha | |
16:10 | thibaud_g joined #koha | |
17:04 | tcohen | oleonard: we decided not to push css and related things to the repo |
17:05 | cait[m] | did we update documentation yet? |
17:05 | I am ont sure what I need to do now | |
17:05 | tcohen | oleonard: on macOS it is weird |
17:05 | because it is a VM, as in Windows | |
17:05 | in my case, I | |
17:05 | - have LOCAL_USER_ID set correctly | |
17:05 | - always pull the latest image | |
17:06 | - just in case, chown everything (outside KTD) before launching | |
17:06 | you need to be aware run.sh (the startup script for KTD) | |
17:07 | will pick LOCAL_USER_ID and change the UID for the kohadev user | |
17:07 | to match | |
17:07 | so, in general | |
17:07 | I'd ktd down | |
17:07 | then chown -r <your user> ~/git/koha | |
17:07 | then start over | |
17:08 | cait[m] | what if I have a patch with CSS changes? any changes to the yarn build commands and such I need to be aware of when testing? |
17:10 | tcohen | those things need to be built |
17:11 | KTD already does on startup, for Vue | |
17:11 | we could add for the others | |
17:11 | the current status is not ideal, but ideas are welcome | |
17:11 | cait[m] | i thought that was done already |
17:13 | so right now it's a must to run the yarn builds every time I do a ktd up? | |
17:13 | I noticed that there were css changes when I rebaesd my 22.11 branch that I needed to submit, so there is an additional step now too when you built your own packages | |
17:13 | we probably need to communicate this change better | |
17:20 | tcohen | I think I will go push a commit with the built CSS for now |
17:20 | and next week, with some fresh brain I'll comment about options and propose a discussion/vote on a dev meeting | |
17:20 | cait[m] | sounds good thank you |
17:21 | should also make sur eit makes it to the 'how to write a patch' things in the wiki | |
17:21 | people might be using jus tgit without ktd | |
17:21 | need to make sure we tell them how to make CSS changes | |
17:22 | tcohen | FTR: yarn commands need to be run inside kshell because they implant new versions of files |
17:22 | in the git clone | |
17:22 | cait[m] | hm I can check if that works for me sec |
17:22 | yep that works ok | |
17:22 | hm ok, not totally | |
17:23 | Error in plugin 'sass'... (full message at <https://matrix.org/_matrix/med[…]sTpmzvjjjEtkWBleg>) | |
17:23 | that's for yarn build --view opac | |
17:25 | tcohen | oleonard-away: thanks for reporting it, and sorry for the confussion |
17:26 | pastebot | "tcohen" at 127.0.0.1 pasted "Yarn on kshell" (15 lines) at https://paste.koha-community.org/135 |
17:26 | cait[m] | I am running a yarn install... to see if that makes it go away, not sure if it's a real error - taking forever tho |
17:27 | .... and it did make the error go away, we are good | |
17:27 | tcohen | Another comment about yarn and friends: node dependencies will typically be bundled already in KTD if you use a recent image. If you are facing issues, it might be just that your stray node_modules/ in your clone is getting in the middle |
17:28 | cait[m]: don't do a yarn install | |
17:28 | what I was saying | |
17:28 | that's problematic, because (say) we change the node.js version from node v12 to v14 | |
17:29 | and some libs are pre-compiled for an older runtime, things fail | |
17:29 | we wanted some consistency between dev environment and CI | |
17:29 | so we bundled in KTD | |
17:30 | talking about master | |
17:30 | ^^ just in case | |
17:32 | cait[m] | we should write all that up somewhere :) |
17:32 | ok, I'll do a fresh ktd pull instead now - think that is what you were saying? | |
17:32 | tcohen | nope |
17:32 | that's always good | |
17:33 | but the important bit was to clear your node_modules dir | |
17:33 | cait[m] | should the rebuilding of the env have the same effect? or do i need to do something still then? |
17:33 | oleonard joined #koha | |
17:33 | tcohen | cait[m]: I'm not sure |
17:34 | but please do inside kshell :-D | |
17:34 | cait[m] | i meant doing a ktd down ktd pull ktd up... cannot do that in kshell ;) |
17:35 | tcohen | https://gitlab.com/koha-commun[…]ckerfile#L147-151 |
17:35 | cait[m] | i got that I should prefer kshell, but we also established not everything works there today :D |
17:35 | tcohen | to make it clear |
17:35 | we already do the yarn install step | |
17:35 | in the build step | |
17:35 | and put the deps in /kohadevbox/node_modules | |
17:35 | I'm repeating myself as my grandpa does | |
17:35 | sorry | |
17:35 | cait[m] | nah that's good; knowing we do the yarn helps |
17:36 | i thought it was only the vue | |
17:36 | tcohen | one thing is Docker image build time... |
17:36 | which I'm showing | |
17:36 | the other is ktd up, which is run time | |
17:36 | vue is done when starting ktd | |
17:36 | and I'm not sure about that, but we keep it for now | |
17:36 | cait[m] | yes, understood |
17:37 | trying to follow | |
17:37 | tcohen | we really need to think of a proper way to work with this built things, that traditionally don't go to the repo on other projects, so not kosher what we've been doing |
17:37 | we need ideas | |
17:37 | vue has a gulp watch or smth | |
17:37 | we could do that for all things taht are built | |
17:38 | cait[m] | my main concern is things changing fast and people getting stuck - documentation etc |
17:38 | tcohen | and there's been talks with mtj about adding this build step in the packaging |
17:38 | yeah | |
17:38 | as long as I'm the only one patching the README we are in trouble :-D | |
17:38 | cait[m] | I'll keep poking and nagging |
17:38 | i'd do it everytime i am sure i could do it right | |
17:39 | but most of the time I don't undestand it well enough yet to be the one documenting | |
17:39 | tcohen | all the aliases section in the README have too much precedence |
17:39 | compared to more useful things like | |
17:39 | the included aliases and which user needs to be used to run | |
17:39 | inside ktd | |
17:39 | which is documented far below | |
17:40 | cait[m] | maybe something like a 'quick guide' to get started at the top? |
17:40 | for hte most common use case | |
17:40 | tcohen | next week |
17:40 | wahanui | next week is unknown for me. |
17:40 | tcohen | :-D |
17:41 | I know how I'd change the README | |
17:41 | he | |
17:41 | just lack of brain at the moment | |
17:43 | jzairo joined #koha | |
17:53 | cait[m] | heh |
17:53 | enjoy your weekend | |
19:10 | jzairo | caroline++ |
20:28 | cait[m] | caroline++ jzairo++ |
20:42 | philor | guessing not many people cancel background jobs? |
21:54 | philor joined #koha | |
22:45 | tuxayo | hi # koha :) |
22:46 | davidnind: around? I was wondering if from your experience of the meeting script, only the connection with the calendar had issues? | |
22:50 | At some point I was leaving two #info next meeting in the meeting due to the two timeslots but I stop doing that a few months ago (I made the N+2 different enough so the script doesn't pick it) | |
23:09 | davidnind | tuxayo: it used to work up to updating and creating the pages on the Wiki, but not updating the calendar (mainly because I couldn't get that bit to work with th elibrary used) |
23:09 | The last time I tried all it would do was update the Next IRC meeting page.... | |
23:10 | The last couple of meetings I've setup, I've just done manually. | |
23:11 | tuxayo | > The last time I tried all it would do was update the Next IRC meeting page.... |
23:11 | So it missed: the calendar, the current page to add the logs, creating the next meeting page | |
23:13 | davidnind | Yep - the calendar part was because I couldn't get the library/script used to connect to the Google Calendar to work |
23:14 | I was trying to find the URL for the test Wiki, but it is escapong me at the moment | |
23:15 | tuxayo | Ah it was during the meeting switch? |
23:18 | davidnind | ? |
23:21 | tuxayo | > I was trying to find the URL for the test Wiki |
23:21 | Where your issues related to the ongoing upgrade of the wiki? | |
23:23 | davidnind | with the calendar, no - I think the script stopped working as expected for updating the Wiki after the Wiki upgrade |
23:26 | tuxayo | > I think the script stopped working as expected for updating the Wiki after the Wiki upgrade |
23:26 | oh this is bad :o | |
23:27 | davidnind | I think it is open for someone to fix - https://irc.koha-community.org[…]2-12-19#i_2467648 |
23:27 | tuxayo | Is the script code hosted somewhere? So I can check if it has an outdated dependency to interact with mediawiki |
23:32 | davidnind | it used to be here as koha-meetings - but it isn't showing for me.... https://gitlab.com/koha-community |
23:36 | just to clarify that the meeting script used gcalcli to connect to the Google Calendar API - that was the part I couldn't install/get to work (as I was previously on Ubuntu 18/.04) | |
23:39 | Here it is (I wasn't logged in to GitLab) https://gitlab.com/koha-community/koha-meetings - can you see it? | |
23:42 | here is the test Wiki https://wiki.test.koha-community.org | |
23:46 | tuxayo | davidnind++ I was digging everywhere but this didn't pop on searches |
23:47 | even though it's indeed named koha-meetings.pl | |
23:53 | I'll try the script on next meeting for the wiki part and check if more recent versions of MediaWiki::API will work | |
23:53 | davidnind | you could test on the test wiki if you are really keen.... |
23:56 | tuxayo | Ah yes, I don't have to wait and risk having mess to clean up |
23:57 | yay I can login there | |
23:59 | So to test I should... (full message at <https://matrix.org/_matrix/med[…]FkrrynxfAkuyVsKaQ>) | |
23:59 | Anything more? |
← Previous day | Today | Next day → | Search | Index