Gajim - 2024-05-07

How/where can I set the avatar image of a MUC in Gajim? `mod_vcard_muc.lua` is (should be) enabled.

After checking twice, I found it: Under "Manage"!

> gajim today suddendly broke the dark theme does anyone have an idea on what might be the cause?

also, does anyone know when and if gajim will ever support cross-client encrypted calls?

There's already a functional pr just hanging there nothing being done about it ✏

if it'll be merged into master, it'd be great

though I guess oss mainteinance is not that easy

In its current state, it cannot be merged into master.

(see the discussion at the merge request)

There's one that can

I see unresolved threads in both PRs…

chud, no, both cannot be merged into master (by discussion I meant to point to your MR)

İ thought you just decided not to. But it can be merged, it is working.

> chud, no, both cannot be merged into master (by discussion I meant to point to your MR) hasn't seen a response in a month... ↺

the pr is 1 month old

chud: just because it's working doesn't mean it can be merged. It needs to fit conceptually, it needs to be maintainable, and it needs to be reviewed.

taba: what are you implying?

Like the concept of working?

Maintainable like what gajim currently has right

> taba: what are you implying? how can it be unmergable if it hasn't seen a response in a month ↺

> Maintainable like what gajim currently has right We're working to make that better, not worse, right? ;) ↺

taba: please stop trolling

oh my goodness...

taba: last warning

i'm not trolling...

>> Maintainable like what gajim currently has right > We're working to make that better, not worse, right? ;) Working would seem like an improvement over broken. But that's just me İ guess ↺

I have much simpler and shorter PRs waiting for review for a lot more than that, so what? Volunteer maintainers set their priorities. Gajim is 20+ years old, it has a lot of technical debt, but AFAICT maintainers are working hard (unpaid!) to make sure every commit is the opportunity to improve the maintainability of the code base and set the bar high in terms of code quality and modern good practices.

exactly, I think we shouldn't demand or expect anything but instead be thankful to all of the devs working hard technically for free

> to improve the maintainability of the code base and set the bar high in terms of code quality and modern good practices. Functionality not being one of those?

> if it'll be merged into master, it'd be great maximus: fork it. will it get into a tagged version soon enough for it to matter to you? ↺

Believe me, I understand your frustration chud, but aren't calls broken ATM because of poor maintainability of the previous implementation? FWIW nothing prevents you from running your commits on top of master. That's what I do with my MRs, and eating my dogfood makes me fix/improve things. The unresolved threads of the MR actually lists stuff to do for the MR to be accepted, don't give up, I'd be happy to have calls in back in gajim too.

The last message from gajim maintainers says they'd go with the other branch, building up exactly on the previous implementation. İf you guys are actually willing to go with mine if the "code smell" of deep blocks is fixed, or anything of the kind, I'd be happy to oblige.

wrong chat

wrong chat

wrong chat

wrong chat

wrong chat

This person attempted to retract a previous message, but it's unsupported by your client.

https://draugr.de/upload/00f530bb49b4e1685114d28f8bb6fb44b2fcbc2f/bzMibRCRfS2c0PgHW8YmnOtXjoy4IKcAeeUNsu0L/1000127240.jpg

Lol

> egnun, what exactly makes gajim usable for server admins? lovetox, well because of all the admin and channel operator oriented features, Gajim supports. ✏

Hi! lovetox, I'm sorry! I completely forgot about that Weblate translation engine ) So, I just pushed the translation right in.

In fact I tried the service again and confirmation letter are not coming as always.

So, if there's no workaround, it looks like I'm completely stuck with that comrimation-letter-not-coming issue

slichtzzz, give me a moment

i wrote our administrator again, i hope we can fix this email issue soon

Is there some tutorial for writing plugins, or should I just figure it out by looking at existing ones?

you need to look at existing ones

ok

so, before I start, is it possible to write a plugin to support CSI?

no

Oh? Plugins can't send raw stanzas?

sorry i confused what CSI means

should be possible to send a something on the stream

client state indication lovetox ?

https://xmpp.org/extensions/xep-0352.html

IIRC

even though that's more a mobile client thing rather than Desktop's.

no its not, and i dont see it as really important for a desktop client

👍️

> > if it'll be merged into master, it'd be great > maximus: fork it. will it get into a tagged version soon enough for it to matter to you? not sure what you mean here

chud, as i wrote i see your MR as a proof of concept regarding the use of the libdatachannel. - The MR ignores that we have xmpp library - It provides no new framework for jingle related stuff - It does not integrate or tries to improve the current framework - it introduces new async concepts - in essence you dropped 5000 lines of code in separate modules which has no resemblence to the rest of the codebase still, i commend you for showing a working a/v solution with libdatachannel, and i will think about how we can salvage your work. Sadly the other contributor seems to be also not available anymore.

One thing it does do is work tho

As if the async stuff is a bad thing lol

İ mean of course we can wait another year(s?) for glib guys to maybe merge their implementation ✏

Theirs will be superior, but it only *will* and mine is there and now and can deliver.

Not sure what you mean by "ignores that we have xmpp library". Please tell me any specific issues that can at least in principle be fixed and I'll be happy to look into those. Like " doesn't display a user friendly message about a server not having tun/turn" yes İ can fix that. But "no semblance of the rest of the codebase" I'm not exactly sure what that means or how that is a breaking issue worth not having working calls.

"framework"... Why do we want one, let alone why is it a dealbreaker? There's a framework currently and it's broken so what good is it. A more complex structure will be more fragile. Do calls and file transfers even share that much logic in practice?

Also of those 5000 (is it 5000? Didn't count tbh, thought it'd be less) a good chunk are unit tests, which I'd argue will make it *more* maintainable.

you use async concepts in your whole MR, this does not resemble the Gajim codebase, you did this without asking or discussing if this is something that this projects wants at all. seems you are comfortable with it so you went ahead. But are you there for the rest of the way? Converting the xmpp library and the rest of Gajim to async? i highly doubt it. you have a whole module dedicated to parsing and building xml stanzas, thats what a xmpp library is for, we work since years to get everything that is xml out of Gajim. We have now some abstraction for jingle, and why would i throw this away. Not working does not mean that the abstraction is wrong or not working, its simply unmaintained code that somewhere breaks and nobody had time to fix it. At least this is a major design decision which should be discussed before. I would have expected some kind of thoughts about how others after you can integrate and extend your code.<

I dont see the vision of converting the codebase to async, i dont see any real benefit, there are sprinkled some points in the codebase where we need to request more than one IQ but thats it, thats for me not convincing enough to start adding async stuff.

lovetox MarsIronPI Maranda CSI might be useful even on a desktop client. E.g. when travelling with Deutsche Bahn or other train connections with poor internet connection, such as EuroStar or ... in fact most :-)

debacle, you need a trigger so gajim tells the server the state

on a phone this can be if the screen goes dark, or is locked

bad wlan is not something gajim can detect and tell the server

lovetox Yes, on a desktop maybe "minimized" or not?

yes or away

but its not evident how a user would know he should minimize gajim, because his wlan is bad

never heard of any application behaving that way before

CSI implies that you are not active on the application

because the server will not send you specific data anymore, like presence for example

lovetox Not likely, indeed. But at least if Gajim is minimized it would (maybe) eat less bandwidth or something.

yes, CSI is for saving data when you are *not* active on the application

its a "yeah why not" thing

we would probably implement with the away timer

Sound right. Probably priority 563 of your long list ;-)

yeah its definitly something i very much doubt a user would actively notice

i would invest the time rather in things the user profits more from, like reactions, retractions

file transfer, calls .. list is long

I fully agree!

before i implement , 5% data save

Only saying: Just because an app is for desktop, it doesn't mean it weren't mobile, too. ✏

Too many negations, I'm confused by myself.

slichtzzz, the admin of the server is on holiday until next week, so sorry, please be patient we will get it eventually

lovetox pushed 1 commits to branch gajim/master imprv: Dismiss invitation notification when MUC is joined - https://dev.gajim.org/gajim/gajim/-/commit/bd7b61ed43e8b6e1e33641f15b87deffc0b472ca

lovetox: for me, CSI would be useful for running Gajim on my pinephone

lovetox: so plugins can't send the raw XML to the server?

hm .. i think you can

Oh ok. Is there some existing plugin that does that I can look at?

lovetox?

> you use async concepts in your whole MR, this does not resemble the Gajim codebase, you did this without asking or discussing if this is something t https://conference.gajim.org:5281/pastebin/9e6a5b18-8ea4-45d2-a817-6f148742617b

My goodness

Guess İ should've started that with "Dear lovetox, İ hope this letter finds you well"

To sum up: I am more than willing to discuss and fix any issues with this, but please do give me feedback and specific issues you see so that I actually have something to reply to. Please let me know the condition of this being accepted into mainstream. I would love to devote more time to this, but I don't want it to be in vain.

Btw one of the things that I personally would like to do is give user the option to choose codecs (e.g. on a shit connection you may wish to sacrifice opus quality for the narrowband pcmu )

lovetox: I guess you already know but the fix for the slow startup worked well.

fwiw, I do hope this a/v MR gets merged

fjklp: which one?

I don't know the code of either and we having nothing working now so whichever has a chance :)

OK. I'm actually running one of them myself right now

MarsIronPI, im not sure any plugin does that

lovetox: is there any API doc?

no, look in the client.py module it sends stanzas

lovetox: OK, thanks!

there is a nbxmpp client object there you need to access that directly to send non-stanzas

OK

cal0pteryx: Damn, that new message highlighting is niiiiice. I see that we aren't scrolling now. Is this to prevent that bug of not getting the target message into view?

fjklp, I didn't touch scrolling, only how the animation is applied and the animation itself

What keys are required in the plugin-manifest.json?

cal0pteryx: Hmm. On my flatpak, when I click on a message search result, the main chat window briefly shows the "home" screen (gajim icon and "Start Chatting" button), then displays the chat history and double-blinks the highlight.

fjklp, no clue, never saw that.

are you using flatpak?

no, but that should not make a difference regarding message highlighting

I'm seeing what I just described when clicking message search results and when I click reply links it simply pauses on the current screen, then immediately transitions the the place where the target message is, no scrolling animation. This is with flatpak. ✏

I think it was always the case for me that flatpak did not show the highlight animation before these changes.

I don't know why this would be unique to my machine

I forgot that I set a custom theme. Flatpak is such a pain. I'll test it without the theme some other time.

but I'm pretty sure flatpak behaved differently with animation with adwaita default

wurstsalat pushed 1 commits to branch gajim/master refactor: Separate GroupChatInvitation widget from dialog - https://dev.gajim.org/gajim/gajim/-/commit/279227075bd641866b9a101a5ba9adbdcdd77dc6

wurstsalat pushed 1 commits to branch gajim/master cfix: GroupChatInvitation: Provide accepted/declined signals and center widget - https://dev.gajim.org/gajim/gajim/-/commit/417c2b7c56280b3d1ea4be005952e9d94a6d0735