Monday, September 04, 2023
apps@conference.yunohost.org
September
Mon Tue Wed Thu Fri Sat Sun
        1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
 
             

[00:16:41] <Solrac> Pardon me, does resources.apt.packages not install the packages? I got an error saying its not installled 🤔
https://paste.yunohost.org/raw/ivaruturut
[00:16:55] <Solrac> Packages reference; https://github.com/SolAZDev/motioneye_ynh/blob/master/manifest.toml#L106
[00:22:16] <Nicole> > <@lsolrac:matrix.org> Packages reference; https://github.com/SolAZDev/motioneye_ynh/blob/master/manifest.toml#L106

Just going off what I see in other repos, you probably don't want the commas in there. https://github.com/YunoHost-Apps/mediawiki_ynh/blob/master/manifest.toml#L109
[00:23:02] <Solrac> oh 😲 alrighty, retrying
[00:31:55] <Solrac> also how can I reference the administation password for a config file?
[01:28:40] <Yunohost Git/Infra notifications> App headscale goes down from level 7 to 6 in job [#18280](https://ci-apps.yunohost.org/ci/job/18280)
[01:59:59] <Solrac> https://paste.yunohost.org/raw/ofizodecaj What can I do for conflicting packages? And I guess missing packages
[05:00:20] <Yunohost Git/Infra notifications> App pihole rises from level 3 to 8 in job [#18283](https://ci-apps.yunohost.org/ci/job/18283) !
[06:25:22] <Yunohost Git/Infra notifications> App fider stays at level 3 in job [#18284](https://ci-apps.yunohost.org/ci/job/18284)
[06:50:25] <lapineige> > <@lsolrac:matrix.org> also how can I reference the administation password for a config file?

I don't remember the syntax, but I'd suggest to look at an app manifest that set an admin password. I believe synapse is one of them
[07:37:34] <Yunohost Git/Infra notifications> App wetty goes down from level 8 to 3 in job [#18285](https://ci-apps.yunohost.org/ci/job/18285)
[08:02:32] <Thomas> > <selfhoster1312> Thomas, maybe the mail server disables PLAIN authentication? you could try with another mail client to be sure :)

Great idea, I just tried to login with fittrackee__2@mydomain.tld and the password of mail_pwd as password, I've authentification failed
[08:03:17] <Thomas> https://aria.im/_matrix/media/v1/download/matrix.org/gdZpikDSHcnLKqFUzBBdMitJ
[08:04:39] <Thomas> So it seems that the credentials are wrong? Althought there are determined during the install script...
[08:09:38] <lapineige> > <@thovi98:matrix.org> So it seems that the credentials are wrong? Althought there are determined during the install script...

I don't know how, but if you could check the logs on the web server (does the request reach the server ? (nginx) Does the mail server say anything ?) it could help to diagnose it
[08:11:20] <lapineige> You could also try to connect with another app/software, that would double check the credentials… yet I don't think that should be the first track to follow, maybe it's just a bad response from the server or something.
edit: oh, sorry I didn't read it all, it's already a new mail client
[08:17:52] <lapineige> I've got it running !!
Turns out that if I send an Ajax request to /somepage, handled by flask at /somepage endpoint, in nginx it should be proxyfied to… / and not /somepage.

So now I need to figure out why if I send simple URL parameters it works, but if I send more complexe one (bloc of text) it doesn't even 404…
[08:18:30] <Thomas> Yess with multiple clients. The only logs I find is:


postfix/qmgr[1580051]: 3F12823C8477: removed
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: connect from localhost[::1]
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: Anonymous TLS connection established from localhost[::1] to domain.tld: TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (3072 bits) server-digest SHA256
Sep 4 10:05:26 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL PLAIN authentication failed:
Sep 4 10:05:32 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL LOGIN authentication failed: UGFzc3dvcmQ6
[08:18:42] <lapineige> add \`\`\` before and after the code bloc 🙂
[08:19:26] <Thomas> Yess with multiple clients. The only logs I find is:

```
postfix/qmgr[1580051]: 3F12823C8477: removed
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: connect from localhost[::1]
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: Anonymous TLS connection established from localhost[::1] to domain.tld: TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (3072 bits) server-digest SHA256
Sep 4 10:05:26 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL PLAIN authentication failed:
Sep 4 10:05:32 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL LOGIN authentication failed: UGFzc3dvcmQ6
```
[08:19:37] <Thomas> Yess with multiple clients. The only logs I find is:

```
postfix/qmgr[1580051]: 3F12823C8477: removed
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: connect from localhost[::1]
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: Anonymous TLS connection established from localhost[::1] to domain.tld: TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (3072 bits) server-digest SHA256
Sep 4 10:05:26 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL PLAIN authentication failed:
Sep 4 10:05:32 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL LOGIN authentication failed: UGFzc3dvcmQ6
```
[08:19:58] <Thomas> Yess with multiple clients. The only logs I find is:

```
postfix/qmgr[1580051]: 3F12823C8477: removed
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: connect from localhost[::1]
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: Anonymous TLS connection established from localhost[::1] to domain.tld: TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (3072 bits) server-digest SHA256
Sep 4 10:05:26 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL PLAIN authentication failed:
Sep 4 10:05:32 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL LOGIN authentication failed: UGFzc3dvcmQ6

```
[08:21:40] <Thomas> > add \`\`\` before and after the code bloc 🙂

Mmh didn't work 😁
[08:22:00] <lapineige> remove the last line break ?
[08:22:32] <Thomas> Yess with multiple clients. The only logs I find is:

```
postfix/qmgr[1580051]: 3F12823C8477: removed
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: connect from localhost[::1]
Sep 4 10:05:24 next postfix/submission/smtpd[1983247]: Anonymous TLS connection established from localhost[::1] to domain.tld: TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (3072 bits) server-digest SHA256
Sep 4 10:05:26 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL PLAIN authentication failed:
Sep 4 10:05:32 next postfix/submission/smtpd[1983247]: warning: localhost[::1]: SASL LOGIN authentication failed: UGFzc3dvcmQ6
```
[08:22:43] <lapineige> Oh, doesn't fix it
[08:22:52] <lapineige> which app is it ?
[08:23:02] <Thomas> Element
[08:23:25] <Thomas> On Android
[08:34:38] <lapineige> wow, space
[08:35:01] <lapineige> can you try to write a new message between those \`\`\` ?
[08:35:51] <lapineige> oh no wait I tried, indeed there is no formatting. I don't know why
[08:36:18] <orhtej2> \```
test
\```
[08:36:23] <orhtej2> like that?
[08:36:28] <orhtej2> ```
test
```
[08:36:38] <orhtej2> because I saw that working on multiple clients
[08:36:51] <orhtej2> as it's standard markdown way of inserting code blocks
[08:37:44] <lapineige> Yeah but somehow Element on Android doesn't activate markdown by default or something like that
[08:46:51] <lapineige> I think I've found it : If I try the URL manually, removing the '+' that replaces those spaces (for some reason that change a thing…), I have a "Bad request, Request Line too large". A gunicorn issue apparently.
And no damn tool would have raised that anywhere 🙄
(it's a bit pointless to document it here as it's a volatile channel, but anyway…)
edit: no, doesn't change anything, nevermind
[08:58:49] <lapineige> > I've got it running !!
> Turns out that if I send an Ajax request to /somepage, handled by flask at /somepage endpoint, in nginx it should be proxyfied to… / and not /somepage.
>
> So now I need to figure out why if I send simple URL parameters it works, but if I send more complexe one (bloc of text) it doesn't even 404…

Well, no, I should have checked first: it gives me the html / page in a file name with .zip, not the actual zip I wanted 😆
[09:04:16] <orhtej2> Thomas: have you tried `EMAIL_URL=smtp://__APP__:__MAIL_PWD__@127.0.0.1/` ?
[09:04:56] <orhtej2> ah but of course you have
[09:12:13] <Thomas> Yes thank you, I've tried with ports 25, 465 and 587, as you can see here : https://github.com/SamR1/FitTrackee/issues/412
[09:56:58] <lapineige> Ok, found it, while trying some Nginx combinaison I inserted a typo in the flask code, hence I found the good solution but it wasn't pointing at the right page, so… It works !
Also it turned out that the `limit_request_line` parameter in gunicorn was indeed a limiting factor, with no log talking about it (until I activated debug level of logging, which is stupid, this is an error…). I shouldn't have removed my previous post, but now it's documented again 🙂
But… there still seems to be a line limit, even if I set it to unlimited, and there is no longer an error message.

So… Are you ready for the dirty hack of writing parameters to a file, send an empty XHR, then reading the file from python side ? 😂
(if anyone have a better idea…)
[10:12:12] <lapineige> The workaround worked 🙂
[10:13:25] <lapineige> So… It's alive, and online ! 🎉
https://yologen.lapineige.fr/
[10:16:44] <Thomas> Wow congratz 🤩
[10:18:54] <lapineige> It's ugly, it's probably a bit buggy because of the last minute dirty workarounds, but hey, it's alive 😃
[10:20:02] <lapineige> That's the moment I should promise myself not to touch that piece of code ever again 😄 ( At least I swear I will not touch that piece of nginx-gunicorn config myself alone…)
[10:22:58] <orhtej2> > It's ugly, it's probably a bit buggy because of the last minute dirty workarounds, but hey, it's alive 😃

Wheres the code?
[10:23:32] <lapineige> I need to clean it up a bit, and then I'll put it online.
But not today. I need a break 😂
[10:24:58] <orhtej2> > I need to clean it up a bit, and then I'll put it online.
> But not today. I need a break 😂

Sure, no hurry! :)
[11:05:48] <Aleks (he/him/il/lui)> > So… It's alive, and online ! 🎉
> https://yologen.lapineige.fr/

Zomg you published it \o/
[11:07:18] <Aleks (he/him/il/lui)> "Version tutoriel" vs "Version épurée" zomg :3
[11:18:16] <lapineige> why ?
[11:18:30] <lapineige> (and yeah sorry I made the French version first, for the sake of simplicity)
[11:25:42] <lapineige> > So… It's alive, and online ! 🎉
> https://yologen.lapineige.fr/

Oh yeah and by the way please consider this link to be a sort of closed beta, don't publish it in the wild for now please 🙂
[11:26:10] <lapineige> (oh, that triggered element 🎉 animation again, nice 😂 - so here you go for a third one, because why not 😉)
[11:29:46] <orhtej2> > (oh, that triggered element 🎉 animation again, nice 😂 - so here you go for a third one, because why not 😉)

sure
[11:30:04] <orhtej2> (because animation triggers on repies as well for some reason)
[11:33:54] <tituspijean> https://aria.im/_matrix/media/v1/download/pijean.ovh/94d199db84044447e4f8008c91b84d48aa654d220d9f652f149ec1c736b5d7cb
[11:34:02] <tituspijean> (╯°□°)╯︵ ┻━┻ fucking work firewall
[11:34:51] <tituspijean> I will have a look tonight 🎉
[11:40:15] <lapineige> > <@titus:pijean.ovh> sent an image.

🤣
I could be in the online game category yeah 😂
[11:40:32] <lapineige> (so my root domain is blacklisted somewhere 🤔)
[11:42:03] <tituspijean> nah it's a stupid filter I guess. if you are named Olga Meta, and your website is olGAMEta.org, it will get flagged
[11:42:33] <tituspijean> though in your domain name, I dunno... maybe the `yolo` ?
[11:43:11] <tituspijean> I have another word too, but I won't quote it here xD
[11:44:52] <lapineige> (well just try the root one 🙂)
[11:45:38] <tituspijean> yup, banned too xD
[13:24:08] <autra> Aleks (he/him/il/lui): hi! I face the exact same issue as this one: https://forum.yunohost.org/t/postfix-error-var-virtual-db-no-such-file-or-directory/25613
I understand that this is a problem with emailpoubelle, but how do you solve it? Remove emailpoubelle?
[13:24:39] <autra> (sorry, for some reason I cannot join the support room, so asking here)
[13:38:55] <Aleks (he/him/il/lui)> yes, remove emailpoubelle
[13:39:04] <Aleks (he/him/il/lui)> ideally somebody should volunteer to take care of it at some point
[13:39:34] <Aleks (he/him/il/lui)> i think the issue is that it adds a regenconf hook but some variable is never define and ends up being "empty string"
[13:41:04] <autra> thanks! 👍️
[13:54:50] <Salamandar> Fuck i hate mediawiki and their extension packaging
[13:55:14] <Salamandar> the 1.40 branch of ldapauthentication2 is stuck 6 months ago, and I *need* earlier, documented changes
[13:57:27] <Salamandar> so… I add this in the manifest.toml… so disappointed in mediawiki
[13:57:30] <Salamandar> WARNING THIS IS NOT IN THE 1\_40 BRANCH, AUTOMATIC JOBS WILL TRY TO DOWNGRADE THIS
[14:02:56] <Salamandar> *gasp* even the 1_39 branch is more recent…
[14:06:22] <Salamandar> eric_G, if you have any idea about how to handle this… whatever this is (i want to stay civilized)
[14:22:58] <lapineige> > <@titus:pijean.ovh> I will have a look tonight 🎉

I think we just revealed our *very secret* conspiracy to throw confettis all around the channel 😁

[14:23:09] <lapineige> (context for people not using Element+Matrix : everytime an 🎉 emoji is published, there is a wall of confettis shown above the UI 😄) *see, I did it again*
[14:24:36] <lapineige> > <@Salamandar:matrix.org> so… I add this in the manifest.toml… so disappointed in mediawiki

I'm confused about what's actually done and to fix what problem, can you explain again please ?
[14:25:42] <Salamandar> > I'm confused about what's actually done and to fix what problem, can you explain again please ?

Well that is a complex story.
[14:26:09] <Salamandar> Mediawiki comes with extensions. So we have a custom scripts that updates the URLs to the extensions archives (yeah because they have the habit of removing old tarballs, too)
[14:26:58] <lapineige> (To be clear about my intention : I'm trying to understand how we can help, if that's the case)
[14:27:03] <Salamandar> until then we used https://www.mediawiki.org/wiki/Special:ExtensionDistributor to retrieve the URL, but we had a lot of issues like that : https://github.com/YunoHost-Apps/mediawiki_ynh/issues/79
[14:27:19] <Salamandar> yeah but that is good general knowledge to share
[14:27:58] <Salamandar> soo, someone (Eric_g I think) discovered that we could download the same tarballs from Github artifacts (and git doesn’t delete commits, so they always are available)
[14:28:23] <Salamandar> Also extensions are versionned per Mediawiki version… So there is a branch for every MW version in the extensions repositories
[14:28:56] <Salamandar> So we had a nice idea : find the latest branch (but <= than our MW version of course), find the latest commit of the branch, and voilà
[14:29:20] <Salamandar> so that you can visualize :
[14:29:22] <Salamandar> https://aria.im/_matrix/media/v1/download/matrix.org/RoWTrduJGWRKjEheOjeZYbLf
[14:29:50] <Salamandar> And some weeks ago, Mediawiki 1.40 was released and i finally work on repairing ldap
[14:30:24] <Salamandar> but what can I see ? For Ldap extensions, the branch REL1_39 is up to date (some commits are 6 hours ago)
[14:30:31] <Salamandar> but the branch REL1_40 IS STUCK 6 MONTHS AGO
[14:31:23] <Salamandar> (and is not compatible with 1.40 because it’s missing some compatibility with PluggableAuth 6.x, 7.x)
[14:31:26] <Salamandar> so we can’t trust wikimedia to follow their own process
[14:33:04] <Salamandar> so now I’m writing a script that checks all "known" branches (hard coded list) and gets the most recent one…
[14:34:22] <Salamandar> and while i was writing this, my IP was rate-limited by github’s api ><’
[14:35:04] <lapineige> Uh yeah, sounds like troublesome work. Good luck with that 🤗
[14:36:03] <lapineige> > <@Salamandar:matrix.org> and while i was writing this, my IP was rate-limited by github’s api ><’

and I guess they block TOR network IPs ?
If not, launching the app with *torify* (`torify yourscript*`) could help, even if it'sslow
[14:38:44] <Salamandar> Nah, I have multiple IPs, thanks to work + phone + home
[14:53:28] <lapineige> I love the inclusion of work in this 😂
[14:58:32] <orhtej2> > ah but of course you have

ok, I have no clue. Perhaps it has something to do that the app is running on subdomain? I keep getting connections from subdomain.domain.tld
[15:02:29] <Salamandar> > <@Salamandar:matrix.org> eric_G, if you have any idea about how to handle this… whatever this is (i want to stay civilized)

Alright, I have a shitty script that does what I said, looks like it works + ldap will be fixed when I clean up my branch and merge to testing
[15:41:27] <Aleks (he/him/il/lui)> https://github.com/orgs/YunoHost-Apps/repositories only 6 repos left before the prophecy completes
[15:41:31] <Aleks (he/him/il/lui)> ah the number of repo is only displayed on https://github.com/YunoHost-Apps it seems /o\
[15:50:25] <Salamandar> 😄
[15:50:28] <Salamandar> oooof
[15:57:29] <lapineige> > <@Alekswag:matrix.org> ah the number of repo is only displayed on https://github.com/YunoHost-Apps it seems /o\

I can't find it ^^
[15:58:08] <Tag> https://aria.im/_matrix/media/v1/download/lostpod.me/OmXhDOJkFTRgSGXcPnpfRBRr
[15:58:10] <Tag> 👀
[16:01:11] <Émy - OniriCorpe> hey ! how often does the autoupdater (the new python script) run?
[16:01:51] <Tag> once a day IIRC ? there is a cron somewhere
[16:04:53] <Tag> Émy - OniriCorpe: Exactly "Everyday at 18:15"!
[16:07:27] <Solrac> > <@lsolrac:matrix.org> https://paste.yunohost.org/raw/ofizodecaj What can I do for conflicting packages? And I guess missing packages

This is a bit strange installation via yunohost failed but via apt it worked. Packages are installed
[16:08:35] <Émy - OniriCorpe> > <@tag:lostpod.me> Émy - OniriCorpe: Exactly "Everyday at 18:15"!

hm, so yesterday it missed this release https://github.com/Kareadita/Kavita/releases/tag/v0.7.8
and I think running it twice a day would be better
[16:09:15] <Tag> The update asset thing was wrong, I fixed it after
[16:09:30] <Tag> The script complains in #yunohost-dev:matrix.org https://matrix.to/#/!oUChrIGPjhUkpgjYCW:matrix.org/$9loTS5c1IUlN2VeUfSy8mV60zaq2XC20poY6g9pIV_g?via=matrix.org&via=aria-net.org&via=sans-nuage.fr
[16:09:58] <Tag> I should have pinged you :x
[16:10:21] <Tag> Here's my fix : https://github.com/YunoHost-Apps/kavita_ynh/commit/fdeecace825cf9550c051763af4cd3e12505c7fd
[16:15:00] <Émy - OniriCorpe> aaaaaaaah, so the doc is inadequate :/
https://yunohost.org/fr/packaging_apps_resources#regarding-autoupdate
[16:16:24] <Tag> Mmmh for `latest_github_release` there's two items : `autoupdate.asset` or `autoupdate.asset.$arch`
[16:17:00] <Tag> Mais ça pourrait être plus clair peut être
[16:17:46] <Émy - OniriCorpe> "you can"
it's not at all clear that it won't work without it
[16:17:59] <Émy - OniriCorpe> "you can"
it's not at all clear that it won't work without it (autoupdate.asset.$arch)
[16:18:31] <Tag> Well, if none provided, it will use the sources tar.gz IIRC
[16:19:39] <Émy - OniriCorpe> ah
[16:19:42] <Tag> I will confess, I usually read example_ynh rather than the doc :x
[16:21:30] <Émy - OniriCorpe> > <@tag:lostpod.me> I will confess, I usually read example_ynh rather than the doc :x

i did both 😬
[16:22:28] <Tag> ah, yes, the example_ynh uses tags and not releases /o\
[16:24:02] <Tag> anyway, the script should run fine at 20:15 CEST :)
[16:34:14] <lapineige> Which is somewhat confusing by the way, most of the time I find software using releases.
Maybe a comment could present an exemple of release, or even single commit archive, so that we are aware of how it works (without discovering github internal before even packaging something) ?
[16:34:34] <lapineige> (I'm *that* close to integrate it directly in the app generator 😁)
[16:35:15] <eric_G> If you want a +1 to reach the beast number, I packaged ownCloud 😶‍🌫️
[16:35:27] <lapineige> (Which reminds me of the need to discuss together of the packaging will be presented, and the final scripts generated, in the script. I cannot (_should not_) decide alone)
[16:37:08] <lapineige> > <@oniricorpe:im.emelyne.eu> hm, so yesterday it missed this release https://github.com/Kareadita/Kavita/releases/tag/v0.7.8
> and I think running it twice a day would be better

I would agree with this from a packager point of view, because it reduces the chances you see a new release, want to package it, but the bot didn't even started yet… while when it did, first there is less work to do, and most of the time the CI has finished… so sweeeet
[16:39:41] <Tag> so at 6am and 6pm ?
[16:40:54] <lapineige> for example
[16:42:05] <lapineige> it's only for (western) european times, but I'd go for 5am and 6pm, as it makes sure it's done before (most) wake up, or before people get back from work or whatever. (taking into account the CI time too)
[16:42:11] <lapineige> it's only for (western) european times, but I'd go for 5am and 5pm, as it makes sure it's done before (most) wake up, or before people get back from work or whatever. (taking into account the CI time too)
[16:42:14] <m606> > <@oniricorpe:im.emelyne.eu> "you can"
> it's not at all clear that it won't work without it (autoupdate.asset.$arch)

Indeed, I just notice I got trapped as well
[16:42:41] <lapineige> > it's only for (western) european times, but I'd go for 5am and 5pm, as it makes sure it's done before (most) wake up, or before people get back from work or whatever. (taking into account the CI time too)

But maybe they are other reasons for this (time and frequency), I can't tell
[16:44:02] <Tag> This is for purely arbitrary reason I think. So I updated the cron to run at 4am and 4pm UTC
[16:44:22] <lapineige> great, thank you ❤️
[16:44:26] <m606> > <@m606:matrix.org> Indeed, I just notice I got trapped as well

Hmm actually no... I had put a `autoupdate.asset`, but somehow it did not trigger an update
[16:44:38] <Tag> m606: on which app?
[16:44:46] <m606> > <@tag:lostpod.me> m606: on which app?

https://github.com/YunoHost-Apps/cyberchef_ynh/blob/master/manifest.toml
[16:45:37] <Tag> euh https://github.com/YunoHost-Apps/cyberchef_ynh/pull/4
[16:45:38] <Tag> :p
[16:46:30] <m606> ohhh and I should validate PR then ?
[16:47:11] <Tag> Yes, you'll have to review the test and validate the PR
[16:47:56] <m606> Ok thanks
[16:48:22] <Tag> And then you'll have to create a new PR from the testing branch to master branch
[16:49:36] <Tag> depending on the app/the complexity of the upgrade, you may ask users to try to upgrade to the testing branch (on the forum, etc...) and gather feedback before releasing the upgrade to the master branch
[16:49:51] <m606> that's a very easy one (basically static html+JS)
[16:50:45] <lapineige> > <@m606:matrix.org> ohhh and I should validate PR then ?

Basically, it only upgrade the version number and source. As you would (but it's easy to automate, computer-like work, so the computer does it). Then as if it were you making the changes, you need to review them, check if they don't break anything, and so on (the CI helps ;).
[16:52:22] <lapineige> > <@tag:lostpod.me> And then you'll have to create a new PR from the testing branch to master branch

this one is too easily forgotten: I would advice that you do it straight after that, and if not merging directly (it's good to have proper testing, at least for major changes, by real users), for my part I often launch the CI again (it's "useless" if there where no other change, yes), because it sends me an email notification that I can check later, no to forget to merge…
[16:53:01] <m606> Got it, i was looking forward to trying an upgrade to see how it works, but the developer doesn't push release too often. Also I think I havn't been warned by Github about the PR (not automatic when you have contributor status in the repo?)
[16:53:14] <lapineige> > <@m606:matrix.org> that's a very easy one (basically static html+JS)

In these cases, it's often safe to merge if CI is happy.
But if you use the app, I'd advice that you'd be the first to test it once CI is happy 🙂, just in case 😉
[16:54:01] <lapineige> > <@m606:matrix.org> Got it, i was looking forward to trying an upgrade to see how it works, but the developer doesn't push release too often. Also I think I havn't been warned by Github about the PR (not automatic when you have contributor status in the repo?)

you need to follow the repo, in the "watch" button on the top right, and check the corresponding option to be notified of each PR 🙂
[16:55:31] <m606> i was following when "getting mentioned and participating", not enough? I should set it to "follow all activites"?
[16:55:57] <Tag> > <@m606:matrix.org> i was following when "getting mentioned and participating", not enough? I should set it to "follow all activites"?

yup!
[16:56:22] <m606> oh I see there is the "custom settings" option - I can check "PR"
[16:57:32] <Tag> (issues, commits, PR...)
[16:57:44] <Tag> i suggest you to use "all activities" on repos you created, so you don't miss any activity originating from other users/bots
[16:58:17] <Tag> i suggest you to use "all activities" on repos you created/you care about, so you don't miss any activity originating from other users/bots
[17:11:47] <lapineige> > <@m606:matrix.org> i was following when "getting mentioned and participating", not enough? I should set it to "follow all activites"?

You're not mentioned, and not participating *in the PR*. That's the meaning 🙂

> I should set it to "follow all activites"?

This can be too much, in that case, select the custom option, and select the PRs. This can avoid being mentioned for every issue, for instance.
[17:12:30] <lapineige> (wow, took me 15min to post that one, feels like jetlag 😂)
[17:14:22] <Tag> (since I modified the cron task, i did manually trigger the source auto-upgrade script)
[17:17:23] <lapineige> > So… It's alive, and online ! 🎉
> https://yologen.lapineige.fr/

By the way, I took a look at my access logs during one more testing… Did you know the synapse matrix bot (I guess that's the one making the previews ?) interacted with it 6 times ? ^^
[17:26:09] <Émy - OniriCorpe> > <@m606:matrix.org> https://github.com/YunoHost-Apps/cyberchef_ynh/blob/master/manifest.toml

there's a typo
`"CyberChef_.*zip"` instead of `"CyberChef_*.zip"`
[17:26:51] <Tag> > <@oniricorpe:im.emelyne.eu> there's a typo
> `"CyberChef_.*zip"` instead of `"CyberChef_*.zip"`

nope nope, it's actually a regexp :x
[17:28:03] <Émy - OniriCorpe> > Basically, it only upgrade the version number and source. As you would (but it's easy to automate, computer-like work, so the computer does it). Then as if it were you making the changes, you need to review them, check if they don't break anything, and so on (the CI helps ;).

that's why on my GTS package I edited my auto upddate script (the old method) to add the changelog link to the PR, so I can check it easily and verify some things (like the config file changes for example)
[17:28:27] <Émy - OniriCorpe> it would be nice if the new auto update method could do that too
[17:29:13] <Émy - OniriCorpe> > <@tag:lostpod.me> nope nope, it's actually a regexp :x

oh my bad
[17:29:39] <Tag> > <@oniricorpe:im.emelyne.eu> that's why on my GTS package I edited my auto upddate script (the old method) to add the changelog link to the PR, so I can check it easily and verify some things (like the config file changes for example)

Yup, we discussed that in a meeting some time ago, i tried to implement this the day after and got distracted since then :/
[17:30:40] <Tag> It's not as easy as it was in the old method :/ (and it's python, not my best)
[17:31:35] <Tag> Ideally, it would include a changelog or at least a link to the release page
[17:31:39] <lapineige> > <@oniricorpe:im.emelyne.eu> that's why on my GTS package I edited my auto upddate script (the old method) to add the changelog link to the PR, so I can check it easily and verify some things (like the config file changes for example)

oh that's the thing !
It's brillant ! 🙂
[17:32:05] <lapineige> > <@oniricorpe:im.emelyne.eu> it would be nice if the new auto update method could do that too

Do you know where the code of that bot is ?
I might have a look at it, to see if it's simple enough to do it
[17:32:21] <Émy - OniriCorpe> > <@tag:lostpod.me> It's not as easy as it was in the old method :/ (and it's python, not my best)

i already checked and yeah it seems complex, especially if not "last_github_release"
[17:32:42] <lapineige> > <@tag:lostpod.me> Ideally, it would include a changelog or at least a link to the release page

I guess the changelog isn't always doable, depending on where is located, the formatting and so on ?
[17:32:49] <lapineige> > (and it's python, not my best)

I can help with that 🙂
[17:32:49] <Tag> Also we discussed the possibility to "extend" the script, so it run customs actions as update config files or dependencies
[17:33:05] <Émy - OniriCorpe> here it is https://github.com/YunoHost/apps/tree/master/tools/autoupdate_app_sources
[17:33:17] <lapineige> > <@tag:lostpod.me> Also we discussed the possibility to "extend" the script, so it run customs actions as update config files or dependencies

takemymoneyandshutup.gif
[17:34:28] <Tag> > <@tag:lostpod.me> Also we discussed the possibility to "extend" the script, so it run customs actions as update config files or dependencies

But that would need lot of work to run the script in a safe restricted environment, since it runs on our infra (and not github actions anymore...)
[17:35:05] <lapineige> Why would it be necessary ?
[17:36:10] <Tag> Imagine running a random script on your computer... now imagine 600 randoms scripts :p
[17:37:11] <lapineige> I don't get it, as its source is known and controlled, if it fails, it mostly die alone and that's it, right ? (no file to touch, nothing to install…)
[17:37:17] <Tag> By random script I mean, any packager could at some point edit the "extended script" from their package and make a mistake / include malicious code
[17:39:20] <Tag> Ah, I'm not really clear with "extend the script". Initially I though about a python script in the package repo, that would have been called by the autoupdater script.
[17:45:38] <lapineige> Oh ok, I was thinking about a common script for all use case, with different options/parameters to trigger it !
[18:35:59] <lapineige> I mean you said the current one run in your infrastructure ?
[18:36:49] <lapineige> I custom script for each repo can archive precise work, but that would be a burden for each contributor to adjust it if it wasn't good enough/working with the default.
All of that just to get a changelog. That won't happen, I believe.
[18:37:56] <lapineige> While we could have prepare let's say 3 defaults (one for tags, one for releases, one for another particular formating), and expose that with a single (simple ?) manifest option.
It won't fit 100% of the use case… but it would be easy to use, and maybe easier to maintain
[18:41:18] <Aleks (he/him/il/lui)> naively I'd think : if we're fetching stuff from a release, try to extract the release note ; otherwise packages could come with a simple `changelog` URL (ie in the `upstream` block) that would be mentionned in the auto-update PR body, and people can click on it, that would already be a good step
[18:43:04] <lapineige> yeah, clearly, and it's easy to implement (I mean the changelog thing, it's basically one more option in the manifest, then just an URL to copy-paste everytime)
[18:48:18] <wiredwhiskers> Hi hello, I come bearing questions and maybe fixes for LDAP issues for Baikal.
[18:48:28] <wiredwhiskers> and also maybe xBackbone
[18:50:15] <wiredwhiskers> Question: If I encounter a bug and then solve it myself, should I still open an issue on github or is there something else i should do? (sorry im such a noob at this lmao)
[18:51:47] <tituspijean> If the issue is cleary explainable and replicable, then yeah it's great to share it and its solution on github. It's less work for us 😅
[18:52:16] <eric_G> I wasn't joking after all (but it won't be a +1 as the repo already exist) https://ci-apps-dev.yunohost.org/ci/job/9309
[19:47:25] <Yunohost Git/Infra notifications> [apps] @ericgaspar pushed 1 commit to master: Update apps.toml ([aef9ada5](https://github.com/YunoHost/apps/commit/aef9ada5777b896f50e1e9386ba3f542a9362586))
[19:53:50] <lapineige> let's just make our own forkCloud so we can increase the number 😄
[19:54:30] <lapineige> Aleks (he/him/il/lui): if you wish a I a set of softwares I plan to package, I could create repositories (with empty content for now) to max out the number 😁
[19:56:07] <Aleks (he/him/il/lui)> But are you sure you want to unlock *the prophecy* tho
[19:57:17] <eric_G> the 9th gate...
[20:15:43] <lapineige> we should first enter into Bijû or even Baryon mode 😛
[20:19:35] <Yunohost Git/Infra notifications> [apps] @ericgaspar pushed 1 commit to master: Update apps.toml ([99d5c7de](https://github.com/YunoHost/apps/commit/99d5c7de59772df234de5eca9056d39d74fb2450))
[20:58:02] <Salamandar> Well. I spend an afternoon waiting for CI jobs at work so I had time to implement SSO support for Mediawiki ❤️
[20:58:03] <Salamandar> https://github.com/YunoHost-Apps/mediawiki_ynh/pull/106 and
[20:58:03] <Salamandar> eric_G: i think you’ll be interested
[20:58:03] <Salamandar> https://github.com/YunoHost-Apps/mediawiki_ynh/pull/107
[20:58:04] <Salamandar> (yeah shitty PR names but basically the first one fixes LDAP login, the second one is cleanup + SSO support + config_panel)
[20:58:07] <Salamandar> (TBH i post this here because i'd love some review ^^'')
[23:50:25] <m606> the part `.*` of the regex pattern matches `v10.5.2.` out of the string `https://github.com/gchq/CyberChef/releases/download/v10.5.2/CyberChef_v10.5.2.zip`. I guess the only issue was that I was not aware I had a PR to review :)