Time  Nick     Message
01:08 mtj      hi tuxayo, there is a reason why the docker pipeline has not been triggered...
01:10 mtj      ..its waiting for the 21.05.d10 jenkins build to be successful
01:11 mtj      (i programmed this for all branches)
01:22 tuxayo   mtj: thanks
01:22 tuxayo   What about the case from here https://wiki.koha-community.org/wiki/Release_maintenance#Backporting_a_patch_with_a_new_perl_dependency
01:22 tuxayo   > 4.  Disregard the CI that will fail
01:23 tuxayo   Many dep changes will cause the build to fail due to missing lib on the old OS image
01:24 tuxayo   Ok, in this case this case, the unrelated t/db_dependent/Koha/Items.t failure dooms us ^^"
01:26 mtj      yeah, we hit a problem - because d10 build is already failing
01:27 mtj      easy fix here is to manually start the docker image building
01:33 mtj      i think the best solution here would be to have 2 sets of docker images, for each branch
01:33 mtj      both koha-staging and koha-testing images
01:34 mtj      the koha-staging images would be used for jenkins testing
01:34 mtj      the koha-testing images would be used for developers
01:35 mtj      once a jenkins build passes using the koha-staging image - the image gets pushed to koha-testing
01:40 mtj      ...this means that we can guarantee that developers will always have working 'koha-testing' images, even when RMaints have pushed commits that break the latest docker images
01:51 tuxayo   Ok I get it, nice idea!
01:52 tuxayo   > RMaints have pushed commits that break the latest docker images
01:52 tuxayo   When that happen, wouldn't it be always the case to be a dep change, and the old image won't work anyway with the new code that requieres it. So it's still broken.
01:53 tuxayo   Not sure something can be done for that.
02:30 mtj      tuxayo: breakage can possibly be caused by anything, not just dep changes
02:30 tuxayo   mtj: right, I meant image breakage. Like ktd doesn't start
02:32 mtj      hmm... if i think about it more, my suggestion is not perfect - but its better than what we have currently
02:33 mtj      even if jenkins is broken, developers have a docker image that is known to work
02:34 tuxayo   > but its better than what we have currently
02:34 tuxayo   totally
02:34 tuxayo   > even if jenkins is broken, developers have a docker image that is known to work
02:34 tuxayo   They have go back a few commits to still have code that matches the image in terms of dependencies.
02:35 tuxayo   mtj++ thanks, for the help, I'll try tomorrow if I can start 21.05.x, my worktree is a conflicting rebase mess now ^^"
02:41 mtj      tuxayo: i think what we want to do is allow Rrmaints to test breaking stuff - without worrying about breaking things for developers
07:05 alex_    Bonjour
07:05 wahanui  que tal, alex_
07:17 tuxayo   mtj: "Rrmaints to test breaking stuff" Speaking of that, the other day I pushed a 21.05.x-test branch and changed the branch in two CI jobs to use it and then reverted it. To test if some backport could have cause the Items.t failure without doing reverts and rebackport on 21.05.x
10:50 oleonard o/
11:30 tcohen   hola #koha
12:16 marcelr  o/
12:24 tcohen   \o
12:24 magnuse  o/
13:38 tcohen   eyes welcome on https://gitlab.com/thekesolutions/Koha/-/commits/oauth_client
13:39 thd      Good morning tcohen
13:40 tcohen   hey thd
13:41 thd      I have been experimenting with Canasta for MediaWiki.
13:41 tcohen   how are things
13:43 thd      To connect to the MySQL Docker container from Canasta, instead of using localhost as the hostname for MySQL one uses the MySQL container ID.
13:45 thd      The MySQL database is stored in a persistent volume but if the Canasta container is shut down and restarted the MySQL container ID changes.
13:45 tcohen   in the docker-compose file, it is called 'the service name'
13:45 tcohen   I insist we should meet and split the load
13:46 tcohen   I can explain you those details
13:46 thd      Yes.
13:46 tcohen   and as I said, I already have a working install
13:46 thd      We can meet now.
13:46 tcohen   which we should tweak with your findings
13:47 tcohen   give me a few minutes so I change room
13:47 tcohen   ashimema: are you around and available?
13:47 ashimema yup
13:50 tcohen   hey
13:51 ashimema I'm more than happy to sit in and soak up the knowledge (and minute it)
13:51 tcohen   https://us02web.zoom.us/j/9964649021
13:52 thd      I have never run zoom.
13:52 tcohen   I think you can use a browser
13:55 tcohen   thd: are you coming?
13:56 tcohen   prefer another way to do this?
13:56 thd      Do you have another way.
13:57 tcohen   are you able to do Google Meet?
13:57 thd      Zoom is non-free but if that is convenient for you at the moment, I am installing zoom.
13:58 tcohen   ah, I thought you were able to use it within a browser
13:58 thd      I do not have much set up on the Debian 10 based laptop which I am using at the moment.
13:59 thd      The browser link gives a link to the download.
14:01 tcohen   ok
14:01 tcohen   we wait for you
14:08 thd      installing some dependencies
14:14 thd      zoom is installed and all dependencies are satisfied.  I am having trouble pasting the link into the zoom dialog box from the zoom program where it is Enter meeting ID or personal link name.
14:16 thd      I can see you but not hear you.  I may have pressed the wrong button.
14:18 thd      I can hear you but video has a red mark through it.  I also have no microphone.  I was not even thinking of that.
14:19 thd      I can see you and here you.  But I have no microphone maybe.
14:20 thd      I can hear you.
15:35 thd      tcohen: LocalSettings.php has a couple of things which I redacted when I put them in git including $wgSecretKey which should be regenerated if exposed publicly.
15:49 tuxayo   > The browser link gives a link to the download.
15:49 tuxayo   That might be a trick, they try to push people to install the client, but there is always a small option to join via browser. I had to use zoom this morning and was able to join with Firefox
16:15 thd      tcohen: I am making some notes in a README-Koha.txt file mostly about LocalSettings.php so that you understand the current state of what I was doing for debugging that file with Canasta.
16:30 caroline Is there a way for me to recover this commit into my own repository, add to it and then create a merge request to the manual for the two commits (David's and mine)? https://gitlab.com/gnucifer/koha-manual/-/merge_requests/1/commits?commit_id=eb07dd1abcb357de27d294cee49c287c6f5bfe6c
16:30 caroline I'm not super fluent in git however
16:33 tuxayo   caroline: create a new branch from main/master, then `git cherry pick eb07dd1abcb357de27d294cee49c287c6f5bfe6c`, then push the branch and create the MR
16:34 caroline tuxayo, it says "fatal: unknown commit eb07dd1abcb357de27d294cee49c287c6f5bfe6c" I feel like I should fetch something, but I'm not sure
16:34 tuxayo   :o
16:36 tuxayo   ok then in https://gitlab.com/gnucifer/koha-manual/-/merge_requests/1
16:36 tuxayo   - click code
16:36 tuxayo   - checkout branch
16:36 tuxayo   It tells to
16:36 tuxayo   `git fetch origin`
16:36 tuxayo   `git checkout -b 'bug-14957-marc-merge-rules' 'origin/bug-14957-marc-merge-rules'`
16:37 tuxayo   Then go back to your new branch based on main/master and retry the cherry pick
16:37 caroline hm, the git checkout gives me an error again "fatal: 'origin/bug-14957-marc-merge-rules' is not a commit and a branch 'bug-14957-marc-merge-rules' cannot be created from it"
16:38 caroline I downloaded a .patch file, I thought maybe I could upload it?
16:38 caroline Code > Email patches downloads a .patch file
16:39 tuxayo   Yes, go to your new branch based on main/master and do `git am patch/of/my/patch/file`
16:40 caroline mkay, that seems to want to work, but "patch does not apply" I'm guessing I have to do some manual work
16:42 tuxayo   :o
16:43 tuxayo   The 1st way might work. It's just that origin isn't the community but rather your repo I guess.
16:43 tuxayo   `git remote -v` to confirm
16:44 caroline origin is indeed my own repo... but I have the community repo as a second remote called "community"... I'll try changin the origin to community and see what it says
16:45 caroline hm, nope same error (fatal xxxx is not a commit blah blah)
16:47 tuxayo   Did you do `git fetch community`?
16:47 caroline yes
16:47 tuxayo   Then `git checkout -b 'bug-14957-marc-merge-rules' 'community/bug-14957-marc-merge-rules'`
16:47 caroline ah, I forgot the second origin there
16:48 caroline still fatal
16:49 tuxayo   T_T
16:50 caroline Ok I think I managed to get the downloaded patch onto my branch with `git apply --3way ~/Downloads/1.patch`
16:50 caroline It gives me the <<< >>> where the conflicts are so I can resolve them
16:51 tuxayo   great :D
16:53 caroline found it here if someone else ever needs it https://www.delftstack.com/howto/git/git-patch-does-not-apply/#troubleshoot-git-patch-error-if-none-of-the-above-commands-work
16:53 tuxayo   thanks!
16:55 caroline ugh, *facepalm* it was already there... just not where I was looking
16:59 caroline thanks for your help tuxayo++ :)
18:05 tuxayo   caroline: oh what happened?
18:06 tuxayo   you're welcome, even if it was only dead ends, we don't know until we are there ^^"