[04:04:27]
<Aleks (he/him/il/lui)> or look into how to implement proper locking mechanism using files /o\
[04:10:29]
<Tag> the analyseCI was still running, I have to kill a subprocess(?) and the script ends and release the lock
[04:11:37]
<Tag> then the CI kill the old container if it's still up when a new job is created
[04:13:13]
<Aleks (he/him/il/lui)> ah /o\
[04:14:23]
<Aleks (he/him/il/lui)> or maybe that should be handled by yunorunner
[04:15:16]
<Aleks (he/him/il/lui)> there's still this weird CI_package_check script between yunorunner and package_check :|
[04:15:58]
<Tag> yup, that could need a massive refactor but CI_package_check is huge
[04:16:26]
<Tag> or maybe not that huge ?
[04:16:58]
<Tag> ah, yep, not that huge
[04:17:12]
<Aleks (he/him/il/lui)> it's mainly https://github.com/YunoHost/CI_package_check/blob/master/analyseCI.sh
[04:17:52]
<Aleks (he/him/il/lui)> but still at least one full day of work or two to convert to python
[04:18:07]
<Tag> yup
[04:18:17]
<Aleks (he/him/il/lui)> though the real important thing is only 1-2 lines :D https://github.com/YunoHost/CI_package_check/blob/master/analyseCI.sh#L161-L162
[04:18:44]
<Aleks (he/him/il/lui)> well there's also some script in lib/ to auto push the level stuff every friday
[04:18:55]
<Aleks (he/him/il/lui)> anyway
[04:19:12]
<Aleks (he/him/il/lui)> someday™
[04:19:22]
<Aleks (he/him/il/lui)> but maybe i shall work on trying to implement the new tests.toml tonight
[04:19:43]
<Aleks (he/him/il/lui)> annnnd i realize it's 11pm already
[04:19:47]
<Tag> I was in a Aquilenet Meeting and
[04:20:09]
<Tag> le prix de l'électricité augmente
[04:20:20]
<Tag> xD
[04:20:33]
<Tag> Pour l'instant ça va, mais on va mettre en place des mesures sur les équipements de la Mezzanine et on y verra plus clair par rapport à la conso électrique du DC assoc'
[04:21:14]
<Aleks (he/him/il/lui)> tiens btw ça se passe comment pour les frais d'hébergement ? on paye un truc ?
[04:21:42]
<Aleks (he/him/il/lui)> je sais plus combien on économise par rapport à scaleway mais moi ça me choque pas si YunoHost paye genre chépa 30 ou 50€/mois
[04:22:18]
<Tag> on avait prévu de compter 35€/mois une fois que tout serait à cogent pour "la moitié du serveur" et j'aurais pris l'autre à 35€/mois (1U à Cogent 70€/mois)
[04:22:31]
<Tag> et pour l'instant qu'il était à la mezzanine aquilenet m'en demandait 35€/mois
[04:22:44]
<Tag> que pour l'instant je paye totalement du coup
[04:22:52]
<Tag> > <@Alekswag:matrix.org> je sais plus combien on économise par rapport à scaleway mais moi ça me choque pas si YunoHost paye genre chépa 30 ou 50€/mois
faut juste qu'on fasse le claire chez aquilenet sur la facture d'élec, parce que pour l'instant elle regroupe la conso du local, des communs de l'immeuble et de l'hébergement ^^
[04:23:34]
<Aleks (he/him/il/lui)> alrighty
[04:23:51]
<Tag> mais oui à première vue passer à 50€ c'est pas déconnant
[04:24:21]
<Tag> (mais du coup divisé en deux, avec ma part)
[04:24:33]
<Émy - OniriCorpe> https://ci-apps-dev.yunohost.org/ci/job/4704
the job is stuck since one hour 😅
[09:45:36]
<@oufmilo:mtx.travelpandas.eu> i have the same problem if i cancel a job in my CI, i must restart yunorunner.service
[09:57:32]
<Guillaume Bouzige> > i have the same problem if i cancel a job in my CI, i must restart yunorunner.service
if you have been able to reproduce the issue, please feel free to update this issue with the details of your tests https://github.com/YunoHost/CI_package_check/issues/47
[09:59:33]
<Guillaume Bouzige> from my own experience last week, by being very very patient I manage to get my tests through the CI...I always get stuck when I try to look at it too much, cancel restart etc...at the end just let it roll slowly and you shall get through
[10:02:30]
<@oufmilo:mtx.travelpandas.eu> > from my own experience last week, by being very very patient I manage to get my tests through the CI...I always get stuck when I try to look at it too much, cancel restart etc...at the end just let it roll slowly and you shall get through
For me it's happen if i cancel a job. But if i don't cancel it everything is ok
[10:03:07]
<Guillaume Bouzige> for me if I cancel, go for 4h bike ride, come back start the job is ok
[10:17:34]
<@oufmilo:mtx.travelpandas.eu> > for me if I cancel, go for 4h bike ride, come back start the job is ok
I have not tested waiting before restarting the job
[14:11:24]
<Tio> I think the Friendica package is again wrongly updated to a dev version https://github.com/YunoHost-Apps/friendica_ynh/commit/9239a06f8ce36d32c4626360828e0649763c11d3
[14:11:25]
<Tio> because of wrong commit versions that were added to the ynh package
[14:11:26]
<Tio> if anyone can please check I left a comment on the commit page
[14:11:31]
<Tio> so it is something that should be addressed very fast so that not many are updating their friendica versions
[14:12:19]
<Tio> if this is the case, it is the second time this is happening thus making it impossible for us to update to the latest friendica stable, and we were unable for the past months because of the same error. Second, the ones that have already updated wont be able to go back to stable since that will break their install
[14:16:02]
<Aleks (he/him/il/lui)> zblerg
[14:16:11]
<Aleks (he/him/il/lui)> can you maybe make a pull request with the appropriate commit change ? :|
[14:16:56]
<Tio> let me see
[14:27:20]
<Tio> https://github.com/YunoHost-Apps/friendica_ynh/pull/89 created bu since am not familiar with github i could only change 1 file
[14:27:40]
<Tio> i guess that's all that has to be changed
[14:31:44]
<Tio> and btw I'd like to also test these on my server but without making a backup first since that will take ages and I have external backups. however i do not know how to trigger an update without backups `sudo NO_BACKUP_UPGRADE=1 yunohost app upgrade appname --force` used to work but seems not to anymore
[14:32:06]
<tituspijean> it's with the `-b` flag
[14:33:41]
<Tio> ah i remember something about that thanks
[14:36:42]
<Tio> worked thanks
[14:37:28]
<Aleks (he/him/il/lui)> hmmmm my !testme doesnt work ? 🤔🙀
[14:37:56]
<Tio> maybe because the version of friendica is the same for the ynh package?
[14:38:01]
<Tio> as it was before
[14:38:04]
<Tio> or you are trying to install as new
[14:38:23]
<Aleks (he/him/il/lui)> hmnah !testme isnt supposed to care about that
[14:39:13]
<Aleks (he/him/il/lui)> hmmm could be the webhook system being down ?
[14:39:15]
<Aleks (he/him/il/lui)> it's supposed to be handled by ci-apps-dev yunorunner tho
[14:39:37]
<Aleks (he/him/il/lui)> blerg anyway, gotta go for groceries..
[14:39:53]
<Tio> can you please confirm that it is ok to get the hashes from here https://github.com/friendica/friendica/commits/stable ? i grabbed the latest commit https://github.com/friendica/friendica/commit/eeadc00e83d8bc7144b0edf74c0fe851ef8636a6 - that's how i think you guys are doing it
[14:40:15]
<Aleks (he/him/il/lui)> yup probably
[14:40:42]
<Aleks (he/him/il/lui)> i'm not sure why this packages fetches sources from git though
[14:41:23]
<Tio> I think we changed it like that since a year or so ago it was an issue with updating it
[14:41:30]
<Tio> https://github.com/YunoHost-Apps/friendica_ynh/pull/60 see
[14:41:41]
<Tio> but should work just fine if we add the proper commits
[14:41:47]
<Aleks (he/him/il/lui)> like, one implicit advantage of using `app.src` with a link to an archive (+hash) is that usually the archive tends to easily correspond to an explicit release from dev folks, which is clearer than a random commit where it's always a bit fuzzy
[14:42:37]
<Aleks (he/him/il/lui)> i mean you'd usually go to e.g. https://github.com/friendica/friendica/releases/tag/2022.12, scroll down, use that archive
[14:42:54]
<Aleks (he/him/il/lui)> but idk maybe there's a good reason
[14:43:26]
<Aleks (he/him/il/lui)> alternatively, maybe using the tag ref (so in that case directly 2022.12) instead of the commit hash could work too
[14:44:02]
<Tio> sure, but since friendica rarely releases stable versions the one thing we have to pay attention is the proper commit hash
[14:44:06]
<Aleks (he/him/il/lui)> again that would maybe remove some ambiguity
[14:44:08]
<Tio> i can commit to that from now on
[14:44:41]
<Tio> once we fix this I can make sure this will go well with proper commit hashes
[14:44:42]
<Aleks (he/him/il/lui)> (though the issue with tag ref is if the dev changes what commit is pointed by the ref ... but that's another story)
[14:44:49]
<Tio> I rely on Friendica so I want to make sure all works
[14:46:08]
<Tio> I can do a pull request with proper hashes every time friendica releases a stable. So from now on things should be fine if that's all is needed.
[14:46:15]
<Tio> we just need to fix this one 😄
[15:12:37]
<tituspijean> Pas de soucis, prends soin de toi :)
Je laisse la main à oufmilo sur sa branche, et si ça roule on écrasera tes premiers commits
[15:12:38]
<Tag> J'ai fais le tour de https://github.com/search?o=desc&q=org%3AYunoHost-Apps+%21testme&s=created&type=Issues et relancé les tests que le webhook n'a pas pu catch lors de la coupure
[15:12:38]
<tituspijean> Pas de soucis, prends soin de toi :)
Je laisse la main à oufmilo sur sa branche, et si c'est roule on écrasera tes premiers commits
[15:12:39]
<Tag> aussi j'ai relancé deux tests (pour nextcloud et plateau) qui ont fail probablement à cause de la coupure
[15:12:43]
<tituspijean> then cancel and restart, and report here if the issue persists
[15:12:45]
<tituspijean> 🤞
[15:12:49]
<Guillaume Bouzige> https://ci-apps-dev.yunohost.org/ci/job/4616
[15:12:51]
<Tag> well it seems to have repaired on its own...
[15:12:51]
<Guillaume Bouzige> indeed sorry for the ping
[15:12:51]
<Tag> My guess is the container 1 has finally done with the jobs that wasn't killed properly
[15:12:51]
<Tag> I'll restart failed jobs
[15:14:20]
<Guillaume Bouzige> weird my latest plateau job went timeout again, i have restarted it but my faith in CI magics is low
[15:14:45]
<Tag> #4619 ?
[15:14:50]
<Guillaume Bouzige> yup
[15:14:50]
<Guillaume Bouzige> 🤞
[15:14:59]
<tituspijean> It should be a direct copy and paste of the yml.
[15:14:59]
<tituspijean> You need the .yml for the Github Action to run. If you want to test the .sh directly starting it should work, albeit with a bunch of warnings regarding the Github Actions environment
[15:15:00]
<tituspijean> The only useful thing to tune in there is the cron frequency.
[15:15:00]
<tituspijean> Best course would be :
1. Locally with `bash .github/workflows/updater.sh`
2. If satisfied with the result upload both .sh and .yml to the repo and try to run the action manually
[15:15:00]
<tituspijean> Dammit.
[15:15:00]
<tituspijean> In YunoHost/example_ynh for instance
[15:15:00]
<tituspijean> Ah sorry xD
[15:15:01]
<tituspijean> Look into flarum_ynh
[15:15:01]
<tituspijean> *where* did I put the default yml...
[15:15:06]
<Tag> (on va éviter un allé-retour paris-bordeaux)
[15:15:07]
<Tag> et voilà $ dig ci-apps-dev.yunohost.org
ci-apps-dev.yunohost.org. 150 IN A 46.231.241.32
[15:15:07]
<Tag> Je vais changer l'IP de ci-apps-dev pour une un peu mieux routée
[15:15:08]
<tituspijean> En parlant d'IP, quelqu'un récemment se plaignait que le forum ou la doc shaiplu n'était pas accessible en IPv6
[15:15:08]
<tituspijean> (Et peut-être pour les déclenchements manuels aussi, j'avoue ne plus savoir)
[15:15:08]
<tituspijean> > <Krakinou> c'est `base: testing` dans le yml?
Ça ça veut dire que l'action va créer une nouvelle branche à partir de testing.
Par contre l'action prendra toujours le .yml de la branche par défaut pour le cron.
[15:15:09]
<tituspijean> Je voulais regarder ça dans Samurai mais j'ai peur de tout péter
[15:15:09]
<Guillaume Bouzige> > <@tag:lostpod.me> #4619 ?
https://ci-apps-dev.yunohost.org/ci/job/4619 soon to be timeout...?
[15:15:09]
<Tag> Ah, it's fully stuck
[15:15:09]
<tituspijean> Aaaaaah oui c'est vrai je l'avais mis en Action de l'organisation YunoHost-Apps
[15:15:39]
<Tag> /o\
[15:15:58]
<Tag> ok CI-2 is not locked anymore, I restarted the failed builds
[15:15:58]
<Aleks (he/him/il/lui)> i was surprised it worked before
[15:15:58]
<Aleks (he/him/il/lui)> hmyeah that part of the code is a bit shaky if i remember correctly
[15:16:00]
<Tag> Not sure that it's the "cancel" button only... I noticed this happen when a job is timed out after 180mins
[15:16:00]
<tituspijean> It will run on the default branch by default. But you can change it temporarily to test it out ;)
[15:17:48]
<@rosbeefandino:3cmr.fr> hooo sorry i read better and matrix discord bridge already do that with the domain value
[15:17:48]
<@rosbeefandino:3cmr.fr> how to ?
[15:17:48]
<@rosbeefandino:3cmr.fr> for a next update matrix discord bridge could automaticaly get users domain value from the synapse config .
[15:17:48]
<@rosbeefandino:3cmr.fr> hello is it possible to get config values "securely" of a third party package ?
[15:19:27]
<Aleks (he/him/il/lui)> https://aria.im/_matrix/media/v1/download/matrix.org/SVDGSJOvOScIZQMxCCjOkCac
[15:19:27]
<Aleks (he/him/il/lui)> am i working on trying to get a set of unified app logos ? hmmmm maybe 👀
[15:20:07]
<Aleks (he/him/il/lui)> https://aria.im/_matrix/media/v1/download/matrix.org/pwKhCvoxKuQjQEXZBNTbgTkG
[15:20:08]
<Guillaume Bouzige> 🤩
[15:20:08]
<@rosbeefandino:3cmr.fr> 206 😱👏
[15:20:08]
<Aleks (he/him/il/lui)> (the full set of 200ish icons i spent my entire sunday on :D)
[15:20:08]
<Guillaume Bouzige> > <Krakinou> I don't know if it's only that, but I'm sure it happens each time 🙂
I have been battling around with CI those days and what I learned is that you must not be in a hurry. the lock get released like much much later on if only the whole thing is not stuck...i will be willing in a nearby future to run some more tests inVM/VPS to see a bit more in details what is going on...globally it works great when you don't look at it too much and you don't expect any result soon. In fact testing shall be done outside of CI in a reproducible env instead so we don't "look for" or "discover" errors in CI.
[15:20:08]
<Pierre (plabuse)> Aleks (he/him/il/lui): nice, beau boulot, nous allons pouvoir mettre des icônes sur le dashboard de Yunohost maintenant, ;-)
[15:20:11]
<Pierre (plabuse)> ok, c'est quoi la taille de tes icônes (140x140), j'aimerais bien faire un sprite css, genre https://w3bits.com/css-sprites/ pour utiliser ensuite dans le code.
[15:20:12]
<Aleks (he/him/il/lui)> Hmyeah if i dont procrastinate and start yet another thing instead of finishing what i already started :D
[15:20:15]
<Aleks (he/him/il/lui)> ébenheu j'ai pas de taille vraiment fixée, mais c'est des carrés, la taille m'importe peu (sauf pour la taille du fichier) car en CSS je resize tout en 200x200
[15:20:15]
<Aleks (he/him/il/lui)> discussing with axolotle i think i'll end up making a PR to host the icons in the apps repo at some point ... the biggest question is wat do with the apps that have no icon/logo, we need some sort of placeholder system, maybe with the icon of the category or tags
[15:20:17]
<axolotle> yep a generated sprite can be a good idea, a single file to fetch from apps
[15:20:18]
<axolotle> that would require to store icon position in sprite in the apps json, with maybe a fallback to a forkawesome icon name indeed
[15:20:20]
<Aleks (he/him/il/lui)> like if you modify a single logo you gotta refetch the entire huge file 🤔
[15:20:21]
<axolotle> yup
[15:20:21]
<Aleks (he/him/il/lui)> hmmm not sure to understand the benefit of a sprite file, that sounds overly complex for just having a single file to fetch 🤔
[15:20:28]
<Pierre (plabuse)> Aleks (he/him/il/lui): ce n'est pas faux, dans notre cas le sprite n'est pas judicieux.
[15:20:41]
<Aleks (he/him/il/lui)> speaking of finishing what I started: before we start packaging too many apps in v2 format, I'm wondering if we shouldnt simplify the notification architecture, ie instead of having:
```
doc/
- DESCRIPTION.md
- ADMIN.md
- notifications/
- pre_install.md
- post_install.md
- etc...
```
I think we should simply have:
```
doc/
- DESCRIPTION.md
- ADMIN.md
- PRE_INSTALL.md
- POST_INSTALL.md
- ...
```
whatcha think folks ?
[15:20:41]
<Aleks (he/him/il/lui)> the idea being to avoid having yet another subfolders, and keeping all filename in caps lock, and this makes it a bit more intuitive that all files support the `__FOOBAR__` syntax for variables/settings
[15:20:51]
<@rosbeefandino:3cmr.fr> > <@Alekswag:matrix.org> discussing with axolotle i think i'll end up making a PR to host the icons in the apps repo at some point ... the biggest question is wat do with the apps that have no icon/logo, we need some sort of placeholder system, maybe with the icon of the category or tags
Why not to put the logo in background of the tile letting the capital letter over the logo?
[15:20:52]
<Aleks (he/him/il/lui)> yeah YunoHost is infinite
[15:20:52]
<Aleks (he/him/il/lui)> ah yeah i was more thinking about integrating this in the webadmin first, not in user portal (for now)
[15:20:52]
<@rosbeefandino:3cmr.fr> > <@Alekswag:matrix.org> ah yeah i was more thinking about integrating this in the webadmin first, not in user portal (for now)
😆one more task ;)
[15:20:52]
<@rosbeefandino:3cmr.fr> Just for my information what is the v2 app?
[15:20:52]
<Aleks (he/him/il/lui)> Countable infinites < Uncountable infinites < Yunohost infinites
[15:20:52]
<Aleks (he/him/il/lui)> (though 11.1 stays compatible with the old packaging format of course)
[15:20:52]
<Aleks (he/him/il/lui)> but indeed we could create a placeholder (or alternative logo) using the good old "periodic table" look
[15:20:52]
<Aleks (he/him/il/lui)> the corresponding doc is in a pull request for now, but you can have a look at https://github.com/YunoHost/doc/blob/8beb56a749efec6dd2a3cf9a7b847fac8d8d1930/pages/06.contribute/10.packaging_apps/14.advanced/03.packaging_v2/packaging_v2.md
[15:20:52]
<Aleks (he/him/il/lui)> a major evolution of the packaging format coming in yunohost 11.1
[15:20:53]
<@rosbeefandino:3cmr.fr> Why changing from json to toml?
[15:20:53]
<@rosbeefandino:3cmr.fr> Readability?
[15:20:53]
<@rosbeefandino:3cmr.fr> If no logo at least it keeps the colored background and the capital letter
[15:20:53]
<Aleks (he/him/il/lui)> there is a also a bunch of PR porting some apps to v2, for example https://github.com/YunoHost-Apps/castopod_ynh/tree/v2
[15:20:53]
<@rosbeefandino:3cmr.fr> > <@Alekswag:matrix.org> hmmm not sure to understand the benefit of a sprite file, that sounds overly complex for just having a single file to fetch 🤔
👍 more i have only 5 apps on m'y ynh why to load all logos. And and who administrate the logo position in the sprite.
[15:20:54]
<@rosbeefandino:3cmr.fr> Hooo, i'm so disconected, i'm packaging only what i need when it's broken. I will throw an eye 👁️🗨️
[15:20:54]
<gredin67> would you prefer JWT secret disabled or a nextcloud_ynh action to write the token in nextcloud config from onlyoffice installer ? https://github.com/YunoHost-Apps/nextcloud_ynh/issues/537
[15:20:55]
<Aleks (he/him/il/lui)> and also to be consistent with `config_panel.toml` being written in toml
[15:20:55]
<Aleks (he/him/il/lui)> (and the future `tests.toml` replacing that horendous `check_process`)
[15:20:55]
<Aleks (he/him/il/lui)> > <@rosbeefandino:3cmr.fr> Why changing from json to toml?
json aint meant to be written by humans, its syntax is annoying, toml is just more lightweight
[15:20:56]
<Guillaume Bouzige> toml is great ! yes
[15:21:13]
<Aleks (he/him/il/lui)> yalh76: hmmm what's the status of https://github.com/YunoHost/package_check/pull/130, is this something we could merge ? 👀
[15:21:25]
<@oufmilo:mtx.travelpandas.eu> Bonsoir,
[15:21:27]
<@oufmilo:mtx.travelpandas.eu> Mon Yunorunner fonctionne ! 🥳🥳
Est ce que c'est bon si je fais un PR avec le résultat comme ici -> https://github.com/YunoHost-Apps/peertube\_ynh/pull/361
Ou alors je dois quand même passer par le votre -> https://ci-apps-dev.yunohost.org/ci/
[15:21:39]
<Aleks (he/him/il/lui)> 🅞🅤🅕🅜🅘🅛🅞: benheu ouai idéalement c'est bien de faire tourner sur la CI classique à moins que y'ai une raison vraiment précise de pas passer par la CI classique
[15:21:39]
<Aleks (he/him/il/lui)> genre ça évite de devoir faire confiance à des CI hébergées par pleins de gens différent
[15:21:40]
<@oufmilo:mtx.travelpandas.eu> > <@Alekswag:matrix.org> 🅞🅤🅕🅜🅘🅛🅞: benheu ouai idéalement c'est bien de faire tourner sur la CI classique à moins que y'ai une raison vraiment précise de pas passer par la CI classique
Ok pas de soucis, c'est juste que parfois ça permet de désencombrer la CI classique.
Je viens de relancer le test sur la CI classique 👍
[15:21:40]
<@oufmilo:mtx.travelpandas.eu> > <@Alekswag:matrix.org> genre ça évite de devoir faire confiance à des CI hébergées par pleins de gens différent
Oui je comprend
[15:22:46]
<nicofrand> Bonjour ! J'aurais éventuellement besoin de créer un script pour l'app Kresus, pour permettre de mettre une dépendance à jour (via pip, dans un virtualenv) et ne souhaite pas créer une màj du paquet à chaque màj de la dépendance. Y'a un truc prévu pour rendre ce script accessible à l'utilisateur (un bouton quelque part ou je ne sais quoi) ?
[15:23:07]
<Aleks (he/him/il/lui)> Hmmm en 11.1 nornalement tu peux faire un bouton dans le config panel
[15:23:08]
<Aleks (he/him/il/lui)> (Est-ce qu'on a fait le truc qu'il faut pour les apps ? Chéplu)
[15:23:10]
<nicofrand> OK, une bonne raison pour migrer vers le nouveau système alors
[15:23:10]
<axolotle> Yep, something like that should work:
```toml
# config_panel.toml
[panel]
[panel.section]
[panel.section.action_name]
ask = "Click me"
type = "button"
icon = "star"
style = "success"
```
```bash
# scripts/config
run__action_name() {
echo "ACTION!"
}
```
[15:23:10]
<nicofrand> merci !
[15:23:51]
<Émy - OniriCorpe> le CI semble planté
`!!! Another analyseCI process is currently using the lock ./CI-1.lock !!!`
[15:24:04]
<Tag> aaaaaa
[15:24:06]
<Tag> Émy - OniriCorpe: all good now
[15:24:13]
<Aleks (he/him/il/lui)> zblerg we could probably add some trick like "if there's no analyseCI running in `ps -ef`, delete the lock"
[15:24:14]
<Aleks (he/him/il/lui)> or look into how to implement proper locking mechanism using files /o\
[15:24:14]
<Aleks (he/him/il/lui)> or store the PID in the lock and check that PID is still running
[15:24:19]
<Tag> then the CI kill the old container if it's still up when a new job is created
[15:24:19]
<Tag> the analyseCI was still running, I have to kill a subprocess(?) and the script ends and release the lock
[15:24:20]
<Tag> yup, that could need a massive refactor but CI_package_check is huge
[15:24:20]
<Tag> or maybe not that huge ?
[15:24:20]
<Aleks (he/him/il/lui)> or maybe that should be handled by yunorunner
[15:24:20]
<Aleks (he/him/il/lui)> there's still this weird CI_package_check script between yunorunner and package_check :|
[15:24:20]
<Aleks (he/him/il/lui)> it's mainly https://github.com/YunoHost/CI_package_check/blob/master/analyseCI.sh
[15:24:20]
<Aleks (he/him/il/lui)> ah /o\
[15:24:20]
<Tag> ah, yep, not that huge
[15:24:21]
<Aleks (he/him/il/lui)> but still at least one full day of work or two to convert to python
[15:24:21]
<Aleks (he/him/il/lui)> though the real important thing is only 1-2 lines :D https://github.com/YunoHost/CI_package_check/blob/master/analyseCI.sh#L161-L162
[15:24:21]
<Tag> yup
[15:24:22]
<Aleks (he/him/il/lui)> but maybe i shall work on trying to implement the new tests.toml tonight
[15:24:22]
<Aleks (he/him/il/lui)> anyway
[15:24:22]
<Aleks (he/him/il/lui)> annnnd i realize it's 11pm already
[15:24:22]
<Tag> le prix de l'électricité augmente
[15:24:22]
<Aleks (he/him/il/lui)> well there's also some script in lib/ to auto push the level stuff every friday
[15:24:22]
<Aleks (he/him/il/lui)> someday™
[15:24:23]
<Tag> xD
[15:24:23]
<Tag> Pour l'instant ça va, mais on va mettre en place des mesures sur les équipements de la Mezzanine et on y verra plus clair par rapport à la conso électrique du DC assoc'
[15:24:23]
<Tag> I was in a Aquilenet Meeting and
[15:24:24]
<Tag> et pour l'instant qu'il était à la mezzanine aquilenet m'en demandait 35€/mois
[15:24:24]
<Tag> on avait prévu de compter 35€/mois une fois que tout serait à cogent pour "la moitié du serveur" et j'aurais pris l'autre à 35€/mois (1U à Cogent 70€/mois)
[15:24:24]
<Aleks (he/him/il/lui)> je sais plus combien on économise par rapport à scaleway mais moi ça me choque pas si YunoHost paye genre chépa 30 ou 50€/mois
[15:24:24]
<Aleks (he/him/il/lui)> tiens btw ça se passe comment pour les frais d'hébergement ? on paye un truc ?
[15:24:25]
<Tag> que pour l'instant je paye totalement du coup
[15:24:25]
<Aleks (he/him/il/lui)> alrighty
[15:24:25]
<Tag> mais oui à première vue passer à 50€ c'est pas déconnant
[15:24:26]
<Tag> > <@Alekswag:matrix.org> je sais plus combien on économise par rapport à scaleway mais moi ça me choque pas si YunoHost paye genre chépa 30 ou 50€/mois
faut juste qu'on fasse le claire chez aquilenet sur la facture d'élec, parce que pour l'instant elle regroupe la conso du local, des communs de l'immeuble et de l'hébergement ^^
[15:24:26]
<Tag> (mais du coup divisé en deux, avec ma part)
[15:24:26]
<Émy - OniriCorpe> https://ci-apps-dev.yunohost.org/ci/job/4704
the job is stuck since one hour 😅
[15:25:56]
<@oufmilo:mtx.travelpandas.eu> i have the same problem if i cancel a job in my CI, i must restart yunorunner.service
[15:25:58]
<Guillaume Bouzige> > i have the same problem if i cancel a job in my CI, i must restart yunorunner.service
if you have been able to reproduce the issue, please feel free to update this issue with the details of your tests https://github.com/YunoHost/CI_package_check/issues/47
[15:25:58]
<Guillaume Bouzige> from my own experience last week, by being very very patient I manage to get my tests through the CI...I always get stuck when I try to look at it too much, cancel restart etc...at the end just let it roll slowly and you shall get through
[15:26:00]
<Guillaume Bouzige> for me if I cancel, go for 4h bike ride, come back start the job is ok
[15:26:00]
<@oufmilo:mtx.travelpandas.eu> > from my own experience last week, by being very very patient I manage to get my tests through the CI...I always get stuck when I try to look at it too much, cancel restart etc...at the end just let it roll slowly and you shall get through
For me it's happen if i cancel a job. But if i don't cancel it everything is ok
[15:26:09]
<@oufmilo:mtx.travelpandas.eu> > for me if I cancel, go for 4h bike ride, come back start the job is ok
I have not tested waiting before restarting the job
[15:26:38]
<Tio> I think the Friendica package is again wrongly updated to a dev version https://github.com/YunoHost-Apps/friendica_ynh/commit/9239a06f8ce36d32c4626360828e0649763c11d3
[15:26:39]
<Tio> so it is something that should be addressed very fast so that not many are updating their friendica versions
[15:26:39]
<Aleks (he/him/il/lui)> can you maybe make a pull request with the appropriate commit change ? :|
[15:26:39]
<Tio> if this is the case, it is the second time this is happening thus making it impossible for us to update to the latest friendica stable, and we were unable for the past months because of the same error. Second, the ones that have already updated wont be able to go back to stable since that will break their install
[15:26:39]
<Tio> let me see
[15:26:39]
<Tio> if anyone can please check I left a comment on the commit page
[15:26:39]
<Tio> because of wrong commit versions that were added to the ynh package
[15:26:39]
<Aleks (he/him/il/lui)> zblerg
[15:26:40]
<tituspijean> it's with the `-b` flag
[15:26:40]
<Tio> i guess that's all that has to be changed
[15:26:40]
<Tio> https://github.com/YunoHost-Apps/friendica_ynh/pull/89 created bu since am not familiar with github i could only change 1 file
[15:26:40]
<Tio> ah i remember something about that thanks
[15:26:40]
<Tio> worked thanks
[15:26:40]
<Aleks (he/him/il/lui)> hmmmm my !testme doesnt work ? 🤔🙀
[15:26:40]
<Tio> and btw I'd like to also test these on my server but without making a backup first since that will take ages and I have external backups. however i do not know how to trigger an update without backups `sudo NO_BACKUP_UPGRADE=1 yunohost app upgrade appname --force` used to work but seems not to anymore
[15:26:41]
<Tio> or you are trying to install as new
[15:26:42]
<Tio> maybe because the version of friendica is the same for the ynh package?
[15:26:42]
<Tio> can you please confirm that it is ok to get the hashes from here https://github.com/friendica/friendica/commits/stable ? i grabbed the latest commit https://github.com/friendica/friendica/commit/eeadc00e83d8bc7144b0edf74c0fe851ef8636a6 - that's how i think you guys are doing it
[15:26:42]
<Aleks (he/him/il/lui)> yup probably
[15:26:42]
<Aleks (he/him/il/lui)> like, one implicit advantage of using `app.src` with a link to an archive (+hash) is that usually the archive tends to easily correspond to an explicit release from dev folks, which is clearer than a random commit where it's always a bit fuzzy
[15:26:42]
<Aleks (he/him/il/lui)> hmmm could be the webhook system being down ?
[15:26:42]
<Aleks (he/him/il/lui)> hmnah !testme isnt supposed to care about that
[15:26:42]
<Aleks (he/him/il/lui)> blerg anyway, gotta go for groceries..
[15:26:42]
<Tio> https://github.com/YunoHost-Apps/friendica_ynh/pull/60 see
[15:26:42]
<Tio> but should work just fine if we add the proper commits
[15:26:42]
<Tio> as it was before
[15:26:42]
<Aleks (he/him/il/lui)> it's supposed to be handled by ci-apps-dev yunorunner tho
[15:26:43]
<Aleks (he/him/il/lui)> i'm not sure why this packages fetches sources from git though
[15:26:44]
<Aleks (he/him/il/lui)> again that would maybe remove some ambiguity
[15:26:44]
<Aleks (he/him/il/lui)> (though the issue with tag ref is if the dev changes what commit is pointed by the ref ... but that's another story)
[15:26:44]
<Tio> once we fix this I can make sure this will go well with proper commit hashes
[15:26:44]
<Tio> i can commit to that from now on
[15:26:44]
<Tio> I think we changed it like that since a year or so ago it was an issue with updating it
[15:26:44]
<Aleks (he/him/il/lui)> alternatively, maybe using the tag ref (so in that case directly 2022.12) instead of the commit hash could work too
[15:26:44]
<Tio> I can do a pull request with proper hashes every time friendica releases a stable. So from now on things should be fine if that's all is needed.
[15:26:44]
<Aleks (he/him/il/lui)> i mean you'd usually go to e.g. https://github.com/friendica/friendica/releases/tag/2022.12, scroll down, use that archive
[15:26:44]
<Aleks (he/him/il/lui)> but idk maybe there's a good reason
[15:26:45]
<Tio> sure, but since friendica rarely releases stable versions the one thing we have to pay attention is the proper commit hash
[15:26:45]
<Tio> I rely on Friendica so I want to make sure all works
[15:26:45]
<Tio> we just need to fix this one 😄
[17:32:24]
<gredin67> would you prefer JWT secret disabled or a nextcloud_ynh action to write the token in nextcloud config from onlyoffice installer ? https://github.com/YunoHost-Apps/nextcloud_ynh/issues/537
[17:33:27]
<tituspijean> (Et peut-être pour les déclenchements manuels aussi, j'avoue ne plus savoir)
[17:33:27]
<tituspijean> then cancel and restart, and report here if the issue persists
[17:33:27]
<@oufmilo:mtx.travelpandas.eu> Mon Yunorunner fonctionne ! 🥳🥳
Est ce que c'est bon si je fais un PR avec le résultat comme ici -> https://github.com/YunoHost-Apps/peertube\_ynh/pull/361
Ou alors je dois quand même passer par le votre -> https://ci-apps-dev.yunohost.org/ci/
[17:33:27]
<tituspijean> Best course would be :
1. Locally with `bash .github/workflows/updater.sh`
2. If satisfied with the result upload both .sh and .yml to the repo and try to run the action manually
[17:33:27]
<tituspijean> You need the .yml for the Github Action to run. If you want to test the .sh directly starting it should work, albeit with a bunch of warnings regarding the Github Actions environment
[17:33:27]
<tituspijean> > <Krakinou> c'est `base: testing` dans le yml?
Ça ça veut dire que l'action va créer une nouvelle branche à partir de testing.
Par contre l'action prendra toujours le .yml de la branche par défaut pour le cron.
[17:33:27]
<tituspijean> *where* did I put the default yml...
[17:33:27]
<nicofrand> Bonjour ! J'aurais éventuellement besoin de créer un script pour l'app Kresus, pour permettre de mettre une dépendance à jour (via pip, dans un virtualenv) et ne souhaite pas créer une màj du paquet à chaque màj de la dépendance. Y'a un truc prévu pour rendre ce script accessible à l'utilisateur (un bouton quelque part ou je ne sais quoi) ?
[17:33:27]
<tituspijean> Pas de soucis, prends soin de toi :)
Je laisse la main à oufmilo sur sa branche, et si ça roule on écrasera tes premiers commits
[17:33:27]
<@oufmilo:mtx.travelpandas.eu> Bonsoir,
[17:33:27]
<tituspijean> Je voulais regarder ça dans Samurai mais j'ai peur de tout péter
[17:33:28]
<Aleks (he/him/il/lui)> am i working on trying to get a set of unified app logos ? hmmmm maybe 👀
[17:33:28]
<Aleks (he/him/il/lui)> (though 11.1 stays compatible with the old packaging format of course)
[17:33:28]
<Aleks (he/him/il/lui)> Countable infinites < Uncountable infinites < Yunohost infinites
[17:33:28]
<@oufmilo:mtx.travelpandas.eu> > <@Alekswag:matrix.org> 🅞🅤🅕🅜🅘🅛🅞: benheu ouai idéalement c'est bien de faire tourner sur la CI classique à moins que y'ai une raison vraiment précise de pas passer par la CI classique
Ok pas de soucis, c'est juste que parfois ça permet de désencombrer la CI classique.
Je viens de relancer le test sur la CI classique 👍
[17:33:28]
<Aleks (he/him/il/lui)> ébenheu j'ai pas de taille vraiment fixée, mais c'est des carrés, la taille m'importe peu (sauf pour la taille du fichier) car en CSS je resize tout en 200x200
[17:33:28]
<nicofrand> OK, une bonne raison pour migrer vers le nouveau système alors
[17:33:28]
<Aleks (he/him/il/lui)> the corresponding doc is in a pull request for now, but you can have a look at https://github.com/YunoHost/doc/blob/8beb56a749efec6dd2a3cf9a7b847fac8d8d1930/pages/06.contribute/10.packaging_apps/14.advanced/03.packaging_v2/packaging_v2.md
[17:33:28]
<Aleks (he/him/il/lui)> (Est-ce qu'on a fait le truc qu'il faut pour les apps ? Chéplu)
[17:33:28]
<Aleks (he/him/il/lui)> there is a also a bunch of PR porting some apps to v2, for example https://github.com/YunoHost-Apps/castopod_ynh/tree/v2
[17:33:28]
<Aleks (he/him/il/lui)> genre ça évite de devoir faire confiance à des CI hébergées par pleins de gens différent
[17:33:28]
<Aleks (he/him/il/lui)> yalh76: hmmm what's the status of https://github.com/YunoHost/package_check/pull/130, is this something we could merge ? 👀
[17:33:28]
<Aleks (he/him/il/lui)> > <@rosbeefandino:3cmr.fr> Why changing from json to toml?
json aint meant to be written by humans, its syntax is annoying, toml is just more lightweight
[17:33:28]
<tituspijean> The only useful thing to tune in there is the cron frequency.
[17:33:29]
<Aleks (he/him/il/lui)> hmyeah that part of the code is a bit shaky if i remember correctly
[17:33:29]
<tituspijean> Ah sorry xD
[17:33:29]
<tituspijean> Aaaaaah oui c'est vrai je l'avais mis en Action de l'organisation YunoHost-Apps
[17:33:29]
<Aleks (he/him/il/lui)> 🅞🅤🅕🅜🅘🅛🅞: benheu ouai idéalement c'est bien de faire tourner sur la CI classique à moins que y'ai une raison vraiment précise de pas passer par la CI classique
[17:33:50]
<Guillaume Bouzige> > <Krakinou> I don't know if it's only that, but I'm sure it happens each time 🙂
I have been battling around with CI those days and what I learned is that you must not be in a hurry. the lock get released like much much later on if only the whole thing is not stuck...i will be willing in a nearby future to run some more tests inVM/VPS to see a bit more in details what is going on...globally it works great when you don't look at it too much and you don't expect any result soon. In fact testing shall be done outside of CI in a reproducible env instead so we don't "look for" or "discover" errors in CI.
[17:33:50]
<Rosbeef Andino 🐧> If no logo at least it keeps the colored background and the capital letter
[17:33:50]
<Aleks (he/him/il/lui)> (the full set of 200ish icons i spent my entire sunday on :D)
[17:34:08]
<🅞🅤🅕🅜🅘🅛🅞> > from my own experience last week, by being very very patient I manage to get my tests through the CI...I always get stuck when I try to look at it too much, cancel restart etc...at the end just let it roll slowly and you shall get through
For me it's happen if i cancel a job. But if i don't cancel it everything is ok
[17:34:09]
<Aleks (he/him/il/lui)> there's still this weird CI_package_check script between yunorunner and package_check :|
[17:34:15]
<Aleks (he/him/il/lui)> but still at least one full day of work or two to convert to python
[17:34:16]
<Aleks (he/him/il/lui)> or look into how to implement proper locking mechanism using files /o\
[17:34:18]
<tituspijean> It will run on the default branch by default. But you can change it temporarily to test it out ;)
[17:34:18]
<tituspijean> Pas de soucis, prends soin de toi :)
Je laisse la main à oufmilo sur sa branche, et si c'est roule on écraser tes premiers commits
[17:34:18]
<Guillaume Bouzige> weird my latest plateau job went timeout again, i have restarted it but my faith in CI magics is low
[17:34:18]
<tituspijean> In YunoHost/example_ynh for instance
[17:34:18]
<tituspijean> Dammit.
[17:34:18]
<🅞🅤🅕🅜🅘🅛🅞> > <@Alekswag:matrix.org> genre ça évite de devoir faire confiance à des CI hébergées par pleins de gens différent
Oui je comprend
[17:34:18]
<tituspijean> En parlant d'IP, quelqu'un récemment se plaignait que le forum ou la doc shaiplu n'était pas accessible en IPv6
[17:34:18]
<nicofrand> merci !
[17:34:18]
<🅞🅤🅕🅜🅘🅛🅞> Je crois que j'ai enfin réussi a faire fonctionner mon yunorunner haha
Est ce que c'est bon si je fais un PR avec le résultat comme ici -> https://github.com/YunoHost-Apps/peertube_ynh/pull/361
Ou alors je dois quand même passer par le votre -> https://ci-apps-dev.yunohost.org/ci/
[17:34:18]
<tituspijean> 🤞
[17:34:18]
<tituspijean> Look into flarum_ynh
[17:34:18]
<tituspijean> Pas de soucis, prends soin de toi :)
Je laisse la main à oufmilo sur sa branche, et si c'est roule on écrasera tes premiers commits
[17:34:19]
<Rosbeef Andino 🐧> how to ?
[17:34:19]
<Aleks (he/him/il/lui)> i was surprised it worked before
[17:34:19]
<Rosbeef Andino 🐧> hello is it possible to get config values "securely" of a third party package ?
[17:34:19]
<Aleks (he/him/il/lui)> hmmm not sure to understand the benefit of a sprite file, that sounds overly complex for just having a single file to fetch 🤔
[17:34:19]
<Aleks (he/him/il/lui)> yeah YunoHost is infinite
[17:34:19]
<Aleks (he/him/il/lui)> https://aria.im/_matrix/media/v1/download/matrix.org/SVDGSJOvOScIZQMxCCjOkCac
[17:34:19]
<Rosbeef Andino 🐧> for a next update matrix discord bridge could automaticaly get users domain value from the synapse config .
[17:34:19]
<Rosbeef Andino 🐧> 206 😱👏
[17:34:19]
<Aleks (he/him/il/lui)> speaking of finishing what I started: before we start packaging too many apps in v2 format, I'm wondering if we shouldnt simplify the notification architecture, ie instead of having:
```
doc/
- DESCRIPTION.md
- ADMIN.md
- notifications/
- pre_install.md
- post_install.md
- etc...
```
I think we should simply have:
```
doc/
- DESCRIPTION.md
- ADMIN.md
- PRE_INSTALL.md
- POST_INSTALL.md
- ...
```
whatcha think folks ?
[17:34:19]
<Aleks (he/him/il/lui)> the idea being to avoid having yet another subfolders, and keeping all filename in caps lock, and this makes it a bit more intuitive that all files support the `__FOOBAR__` syntax for variables/settings
[17:34:19]
<Rosbeef Andino 🐧> Readability?
[17:34:19]
<Guillaume Bouzige> indeed sorry for the ping
[17:34:19]
<Aleks (he/him/il/lui)> a major evolution of the packaging format coming in yunohost 11.1
[17:34:19]
<Aleks (he/him/il/lui)> like if you modify a single logo you gotta refetch the entire huge file 🤔
[17:34:19]
<Rosbeef Andino 🐧> > <@Alekswag:matrix.org> discussing with axolotle i think i'll end up making a PR to host the icons in the apps repo at some point ... the biggest question is wat do with the apps that have no icon/logo, we need some sort of placeholder system, maybe with the icon of the category or tags
Why not to put the logo in background of the tile letting the capital letter over the logo?
[17:34:19]
<Aleks (he/him/il/lui)> but indeed we could create a placeholder (or alternative logo) using the good old "periodic table" look
[17:34:19]
<Aleks (he/him/il/lui)> and also to be consistent with `config_panel.toml` being written in toml
[17:34:20]
<Aleks (he/him/il/lui)> Hmyeah if i dont procrastinate and start yet another thing instead of finishing what i already started :D
[17:34:20]
<Aleks (he/him/il/lui)> ah yeah i was more thinking about integrating this in the webadmin first, not in user portal (for now)
[17:34:24]
<Aleks (he/him/il/lui)> tiens btw ça se passe comment pour les frais d'hébergement ? on paye un truc ?
[17:34:24]
<Tio> I think we changed it like that since a year or so ago it was an issue with updating it
[17:34:24]
<Aleks (he/him/il/lui)> or store the PID in the lock and check that PID is still running
[17:34:24]
<Tio> https://github.com/YunoHost-Apps/friendica_ynh/pull/89 created bu since am not familiar with github i could only change 1 file
[17:34:24]
<Tio> can you please confirm that it is ok to get the hashes from here https://github.com/friendica/friendica/commits/stable ? i grabbed the latest commit https://github.com/friendica/friendica/commit/eeadc00e83d8bc7144b0edf74c0fe851ef8636a6 - that's how i think you guys are doing it
[17:34:24]
<Tio> if anyone can please check I left a comment on the commit page
[17:34:24]
<Tio> https://github.com/YunoHost-Apps/friendica_ynh/pull/60 see
[17:34:24]
<Aleks (he/him/il/lui)> je sais plus combien on économise par rapport à scaleway mais moi ça me choque pas si YunoHost paye genre chépa 30 ou 50€/mois
[17:34:24]
<Tio> I can do a pull request with proper hashes every time friendica releases a stable. So from now on things should be fine if that's all is needed.
[17:34:25]
<Aleks (he/him/il/lui)> (and the future `tests.toml` replacing that horendous `check_process`)
[17:34:25]
<Pierre (plabuse)> Aleks (he/him/il/lui): ce n'est pas faux, dans notre cas le sprite n'est pas judicieux.
[17:34:25]
<Rosbeef Andino 🐧> Why changing from json to toml?
[17:34:25]
<Aleks (he/him/il/lui)> Hmmm en 11.1 nornalement tu peux faire un bouton dans le config panel
[17:34:25]
<Pierre (plabuse)> Aleks (he/him/il/lui): nice, beau boulot, nous allons pouvoir mettre des icônes sur le dashboard de Yunohost maintenant, ;-)
[17:34:25]
<tituspijean> It should be a direct copy and paste of the yml.
[17:34:25]
<Guillaume Bouzige> toml is great ! yes
[17:34:41]
<Aleks (he/him/il/lui)> speaking of finishing what I started: before we start packaging too many apps in v2 format, I'm wondering if we shouldnt simplify the notification architecture, ie instead of having:
```
doc/
- DESCRIPTION.md
- ADMIN.md
- notifications/
- pre_install.md
- post_install.md
- etc...
```
I think we should simply have:
```
- doc/
- DESCRIPTION.md
- ADMIN.md
- PRE_INSTALL.md
- POST_INSTALL.md
- ...
```
whatcha think folks ?
[17:34:42]
<Aleks (he/him/il/lui)> someday™
[17:34:43]
<axolotle> yup
[17:34:43]
<🅞🅤🅕🅜🅘🅛🅞> > for me if I cancel, go for 4h bike ride, come back start the job is ok
I have not tested waiting before restarting the job
[17:34:44]
<axolotle> that would require to store icon position in sprite in the apps json, with maybe a fallback to a forkawesome icon name indeed
[17:34:50]
<Aleks (he/him/il/lui)> zblerg we could probably add some trick like "if there's no analyseCI running in `ps -ef`, delete the lock"
[17:34:51]
<Tio> ah i remember something about that thanks
[17:34:51]
<Tio> sure, but since friendica rarely releases stable versions the one thing we have to pay attention is the proper commit hash
[17:34:51]
<Tio> once we fix this I can make sure this will go well with proper commit hashes
[17:34:59]
<Guillaume Bouzige> > i have the same problem if i cancel a job in my CI, i must restart yunorunner.service
if you have been able to reproduce the issue, please feel free to update this issue with the details of your tests https://github.com/YunoHost/CI_package_check/issues/47
[17:35:02]
<Aleks (he/him/il/lui)> or maybe that should be handled by yunorunner
[17:35:02]
<Aleks (he/him/il/lui)> hmmm could be the webhook system being down ?
[17:35:02]
<Aleks (he/him/il/lui)> anyway
[17:35:02]
<Aleks (he/him/il/lui)> blerg anyway, gotta go for groceries..
[17:35:02]
<Aleks (he/him/il/lui)> again that would maybe remove some ambiguity
[17:35:03]
<Tio> I think the Friendica package is again wrongly updated to a dev version https://github.com/YunoHost-Apps/friendica_ynh/commit/9239a06f8ce36d32c4626360828e0649763c11d3
[17:35:03]
<Aleks (he/him/il/lui)> it's mainly https://github.com/YunoHost/CI_package_check/blob/master/analyseCI.sh
[17:35:03]
<Tio> or you are trying to install as new
[17:35:03]
<Tio> but should work just fine if we add the proper commits
[17:35:03]
<Aleks (he/him/il/lui)> ah /o\
[17:35:03]
<Aleks (he/him/il/lui)> like, one implicit advantage of using `app.src` with a link to an archive (+hash) is that usually the archive tends to easily correspond to an explicit release from dev folks, which is clearer than a random commit where it's always a bit fuzzy
[17:35:03]
<Aleks (he/him/il/lui)> annnnd i realize it's 11pm already
[17:35:04]
<Aleks (he/him/il/lui)> yup probably
[17:35:04]
<Aleks (he/him/il/lui)> but idk maybe there's a good reason
[17:35:04]
<Aleks (he/him/il/lui)> i'm not sure why this packages fetches sources from git though
[17:35:04]
<Aleks (he/him/il/lui)> hmmmm my !testme doesnt work ? 🤔🙀
[17:35:04]
<Aleks (he/him/il/lui)> it's supposed to be handled by ci-apps-dev yunorunner tho
[17:35:12]
<Aleks (he/him/il/lui)> (though the issue with tag ref is if the dev changes what commit is pointed by the ref ... but that's another story)
[17:35:13]
<tituspijean> it's with the `-b` flag
[17:35:13]
<Guillaume Bouzige> from my own experience last week, by being very very patient I manage to get my tests through the CI...I always get stuck when I try to look at it too much, cancel restart etc...at the end just let it roll slowly and you shall get through
[17:35:13]
<Aleks (he/him/il/lui)> zblerg
[17:35:13]
<Aleks (he/him/il/lui)> alternatively, maybe using the tag ref (so in that case directly 2022.12) instead of the commit hash could work too
[17:35:13]
<Aleks (he/him/il/lui)> but maybe i shall work on trying to implement the new tests.toml tonight
[17:35:54]
<Tio> as it was before
[17:35:54]
<Tio> i guess that's all that has to be changed
[17:35:54]
<Tio> maybe because the version of friendica is the same for the ynh package?
[17:35:54]
<Tio> i can commit to that from now on
[17:35:54]
<Tio> so it is something that should be addressed very fast so that not many are updating their friendica versions
[17:35:55]
<Guillaume Bouzige> for me if I cancel, go for 4h bike ride, come back start the job is ok
[17:35:55]
<axolotle> Yep, something like that should work:
```toml
# config_panel.toml
[panel]
[panel.section]
[panel.section.action_name]
ask = "Click me"
type = "button"
icon = "star"
style = "success"
```
```bash
# scripts/config
run__action_name() {
echo "ACTION!"
}
```
[17:35:55]
<Aleks (he/him/il/lui)> discussing with axolotle i think i'll end up making a PR to host the icons in the apps repo at some point ... the biggest question is wat do with the apps that have no icon/logo, we need some sort of placeholder system, maybe with the icon of the category or tags
[17:35:55]
<Aleks (he/him/il/lui)> eh D:
[17:35:55]
<Aleks (he/him/il/lui)> https://aria.im/_matrix/media/v1/download/matrix.org/pwKhCvoxKuQjQEXZBNTbgTkG
[17:35:55]
<Tio> I rely on Friendica so I want to make sure all works
[17:35:55]
<Tio> worked thanks
[17:35:55]
<Aleks (he/him/il/lui)> can you maybe make a pull request with the appropriate commit change ? :|
[17:36:18]
<Tio> we just need to fix this one 😄
[17:36:18]
<Tio> and btw I'd like to also test these on my server but without making a backup first since that will take ages and I have external backups. however i do not know how to trigger an update without backups `sudo NO_BACKUP_UPGRADE=1 yunohost app upgrade appname --force` used to work but seems not to anymore
[17:36:18]
<Tio> let me see
[17:36:18]
<Tio> because of wrong commit versions that were added to the ynh package
[17:36:23]
<axolotle> yep a generated sprite can be a good idea, a single file to fetch from apps
[17:36:23]
<tituspijean> > <@Alekswag:matrix.org> hmmm could be the webhook system being down ?
I noticed it not responding from time to time. In the webhooks logs on Github, its answers with a 500 error in those instances.
[17:36:23]
<Aleks (he/him/il/lui)> alrighty
[17:36:23]
<Aleks (he/him/il/lui)> i mean you'd usually go to e.g. https://github.com/friendica/friendica/releases/tag/2022.12, scroll down, use that archive
[17:36:23]
<Aleks (he/him/il/lui)> okay
[17:36:23]
<Aleks (he/him/il/lui)> well there's also some script in lib/ to auto push the level stuff every friday
[17:36:23]
<Aleks (he/him/il/lui)> though the real important thing is only 1-2 lines :D https://github.com/YunoHost/CI_package_check/blob/master/analyseCI.sh#L161-L162
[17:36:26]
<Aleks (he/him/il/lui)> hmnah !testme isnt supposed to care about that
[17:36:36]
<🅞🅤🅕🅜🅘🅛🅞> i have the same problem if i cancel a job in my CI, i must restart yunorunner.service
[17:36:38]
<Émy - OniriCorpe> le CI semble planté
`!!! Another analyseCI process is currently using the lock ./CI-1.lock !!!`
[17:36:39]
<Tio> if this is the case, it is the second time this is happening thus making it impossible for us to update to the latest friendica stable, and we were unable for the past months because of the same error. Second, the ones that have already updated wont be able to go back to stable since that will break their install
[17:36:41]
<Aleks (he/him/il/lui)> could be worth having a look at the yunorunner logs when this happens, maybe there's some clue about why/where it crashes exactly
[17:36:52]
<Tag> well it seems to have repaired on its own...
[17:36:52]
<Tag> et voilà $ dig ci-apps-dev.yunohost.org
ci-apps-dev.yunohost.org. 150 IN A 46.231.241.32
[17:36:52]
<Tag> #4619 ?
[17:36:53]
<Tag> que pour l'instant je paye totalement du coup
[17:36:53]
<Tag> I was in a Aquilenet Meeting and
[17:36:53]
<Tag> (on va éviter un allé-retour paris-bordeaux)
[17:36:54]
<Tag> (mais du coup divisé en deux, avec ma part)
[17:36:54]
<Tag> ah, yep, not that huge
[17:36:54]
<Tag> yup
[17:36:54]
<Tag> I'll restart failed jobs
[17:36:54]
<Tag> the analyseCI was still running, I have to kill a subprocess(?) and the script ends and release the lock
[17:36:54]
<Tag> /o\
[17:36:55]
<Émy - OniriCorpe> https://ci-apps-dev.yunohost.org/ci/job/4704
the job is stuck since one hour 😅
[17:36:55]
<Tag> aaaaaa
[17:36:55]
<Tag> then the CI kill the old container if it's still up when a new job is created
[17:36:56]
<Tag> Émy - OniriCorpe: all good now
[17:36:57]
<Tag> or maybe not that huge ?
[17:36:58]
<Tag> Pour l'instant ça va, mais on va mettre en place des mesures sur les équipements de la Mezzanine et on y verra plus clair par rapport à la conso électrique du DC assoc'
[17:36:58]
<Tag> Not sure that it's the "cancel" button only... I noticed this happen when a job is timed out after 180mins
[17:36:58]
<Tag> xD
[17:36:58]
<Tag> yup, that could need a massive refactor but CI_package_check is huge
[17:36:59]
<Tag> et pour l'instant qu'il était à la mezzanine aquilenet m'en demandait 35€/mois
[17:36:59]
<Tag> on avait prévu de compter 35€/mois une fois que tout serait à cogent pour "la moitié du serveur" et j'aurais pris l'autre à 35€/mois (1U à Cogent 70€/mois)
[17:37:00]
<Tag> le prix de l'électricité augmente
[17:54:50]
<Guillaume Bouzige> 🤞
[17:54:50]
<@rosbeefandino:3cmr.fr> If no logo at least it keeps the colored background
[17:54:50]
<Guillaume Bouzige> https://ci-apps-dev.yunohost.org/ci/job/4616
[17:54:51]
<@rosbeefandino:3cmr.fr> > <@Alekswag:matrix.org> hmmm not sure to understand the benefit of a sprite file, that sounds overly complex for just having a single file to fetch 🤔
👍 more i have only 5 apps on m'y ynh why to load all logos. And and who administrate the logo position in the sprite.
[17:54:51]
<Guillaume Bouzige> 🤩
[17:54:51]
<@rosbeefandino:3cmr.fr> Hooo, i'm so disconected, i'm packaging only what i need when it's broken. I will throw an eye 👁️🗨️
[17:54:51]
<@rosbeefandino:3cmr.fr> Just for my information what is the v2 app?
[17:54:51]
<@rosbeefandino:3cmr.fr> > <@Alekswag:matrix.org> ah yeah i was more thinking about integrating this in the webadmin first, not in user portal (for now)
😆one more task ;)
[17:54:56]
<@rosbeefandino:3cmr.fr> hooo sorry i read better and matrix discord bridge already do that with the domain value
[17:54:57]
<Guillaume Bouzige> > <@tag:lostpod.me> #4619 ?
https://ci-apps-dev.yunohost.org/ci/job/4619 soon to be timeout...?
[17:55:23]
<Guillaume Bouzige> yup
[17:55:35]
<Pierre (plabuse)> ok, c'est quoi la taille de tes icônes (140x140), j'aimerais bien faire un sprite css, genre https://w3bits.com/css-sprites/ pour utiliser ensuite dans le code.
[17:59:14]
<Tag> Ah, it's fully stuck
[17:59:15]
<Tag> ok CI-2 is not locked anymore, I restarted the failed builds
[17:59:15]
<Tag> Je vais changer l'IP de ci-apps-dev pour une un peu mieux routée
[17:59:19]
<Tag> J'ai fais le tour de https://github.com/search?o=desc&q=org%3AYunoHost-Apps+%21testme&s=created&type=Issues et relancé les tests que le webhook n'a pas pu catch lors de la coupure
[17:59:19]
<Tag> My guess is the container 1 has finally done with the jobs that wasn't killed properly
[17:59:19]
<Tag> > <@Alekswag:matrix.org> je sais plus combien on économise par rapport à scaleway mais moi ça me choque pas si YunoHost paye genre chépa 30 ou 50€/mois
faut juste qu'on fasse le claire chez aquilenet sur la facture d'élec, parce que pour l'instant elle regroupe la conso du local, des communs de l'immeuble et de l'hébergement ^^
[17:59:28]
<Tag> aussi j'ai relancé deux tests (pour nextcloud et plateau) qui ont fail probablement à cause de la coupure
[17:59:49]
<Tag> mais oui à première vue passer à 50€ c'est pas déconnant
[18:19:54]
<🅞🅤🅕🅜🅘🅛🅞> someone already got this error message ?
[18:20:44]
<🅞🅤🅕🅜🅘🅛🅞> 54966 INFO DEBUG - ++ wget --tries 3 --no-dns-cache --timeout 900 --no-verbose --output-document=/var/cache/yunohost/download/bookwyrm/app.tar.gz https://github.com/bookwyrm-social/bookwyrm/archive/refs/tags/v0.5.3.tar.gz
54966 INFO DEBUG - + out='failed: Network is unreachable.'
54966 INFO DEBUG - + ynh_die '--message=failed: Network is unreachable.'
[18:20:46]
<🅞🅤🅕🅜🅘🅛🅞> Complete log: https://yunorunner.travelpandas.eu/ci/logs/bookwyrm-testing_amd64_stable_complete.log
[18:24:51]
<Aleks (he/him/il/lui)> zblerg
[18:25:22]
<Aleks (he/him/il/lui)> yeah idk shit happens, sometimes network becomes unreachable for some reason, this can happens for a shitload of reason
[18:25:37]
<Aleks (he/him/il/lui)> anything special happened on your CI server while this happened ?
[18:29:33]
<🅞🅤🅕🅜🅘🅛🅞> nothing happened to my knowledge, after I really tweak I never package an application. I launch for the first time
[18:33:30]
<Aleks (he/him/il/lui)> then idk, naively just retry and see if that works or reproduce the same error
[18:42:24]
<🅞🅤🅕🅜🅘🅛🅞> same errors when i retry, i'm gonna to search and i'm back to you if i find a solution
[18:52:15]
<Aleks (he/him/il/lui)> could it be that you're in IPv6 only maybe
[19:12:37]
<Yunohost Git/Infra notifications> [package_linter] @jedie commented on issue #107 Github action "YunoHost apps package linter" is failing: Yes, see it also: https://github.com/YunoHost-Apps/django_example_ynh/actions/runs/3752086868/jobs/6373846938 Bugfix: J... https://github.com/YunoHost/package_linter/issues/107#issuecomment-1361952959
[19:14:08]
<Yunohost Git/Infra notifications> [package_linter] @jedie commented on issue #107 Github action "YunoHost apps package linter" is failing: Yes, see it also: https://github.com/YunoHost-Apps/django_example_ynh/actions/runs/3752086868/jobs/6373846938 Bugfix: J... https://github.com/YunoHost/package_linter/issues/107#issuecomment-1361952959
[19:29:27]
<Yunohost Git/Infra notifications> [package_linter] @alexAubin pushed 1 commit to master: Lazyload toml only when its actually needed to simplify boring dependencies issues https://github.com/YunoHost/package_linter/commit/bd6ad626d9a343b665bad1bc19e671fc54255c52
[19:29:39]
<Yunohost Git/Infra notifications> [package_linter] @alexAubin closed issue #107: Github action "YunoHost apps package linter" is failing https://github.com/YunoHost/package_linter/issues/107
[19:29:39]
<Yunohost Git/Infra notifications> [package_linter] @alexAubin commented on issue #107 Github action "YunoHost apps package linter" is failing: Somewhat fixed in https://github.com/YunoHost/package_linter/commit/bd6ad626d9a343b665bad1bc19e671fc54255c52 https://github.com/YunoHost/package_linter/issues/107#issuecomment-1361993830
[20:17:33]
<🅞🅤🅕🅜🅘🅛🅞> > <@Alekswag:matrix.org> could it be that you're in IPv6 only maybe
my server is connected in ipv4 and ipv6
[20:42:07]
<🅞🅤🅕🅜🅘🅛🅞> anyone available to merge this? --> https://github.com/YunoHost-Apps/peertube_ynh/pull/361