Gajim - 2024-09-10

> fjklp, thanks for testing so thoroughly :) Thanks for noticing :D ↺

Honestly, my tendency to try to find bugs in gajim kind of seems like a psychological disorder

lol

Do you work in QA? You seem to have a knack for finding bugs

I do not, but thanks

lovetox: do you think the changing occupant-id problem is really to blame on the server or maybe the client should deal with it?

xiaoleiw20: contact your server admin

I'm guessing gajim used to have an issue tracker somewhere else and was migrated to the gitlab. There are big time gaps between the earliest issues. Not sure what accounts for that. ✏

I'm guessing gajim used to have an issue tracker somewhere else and was migrated to the gitlab? ✏

Yes, there was Trac before

Good day! Did anyone has experienced a non-working button "Next" with Ad-Hoc Commands?

This appears to happens with the new Data Forms view of Gajim.

https://xmpp.pimux.de/file_share/4abd4046-3ecf-4d32-b367-5d28f944181f/scrot-rofi_20240910-092021.png

Button "Next" is functional on previous versions of Gajim.

What's your version?

That was a bug we fixed in recent commits

Version 1.9.3

> That was a bug we fixed in recent commits Thank you ↺

Ah it's not merged yet

Earlier I was in discussion with someone about the keyboard shortcut to quote/reply the previous message, which is bound to [ Shift ] + [ Ctrl ] + [ ↑ ] on my system. My interlocutor brought up how it does the traditional quote-to-reply rather than the recently added `<reply xmlns="urn:xmpp:reply:0">` type. I was wondering: Should Gajim look into supporting both use cases (possibly with a setting option or a different keybind), or move to the `<reply>` tag based solution?

maybe an entry on how to use hats would be a good addition to the gajim faq ✏

Right. I'll add one later

asterix. cal0pteryx. lovetox. Did you receive an email message which starts with "This comment be deleted!"?

I intend to delete this message https://dev.gajim.org/gajim/gajim/-/issues/4501#note_215988

I did not receive anything

sch: please stop pinging everyone

cal0pteryx. I have pinged to the admins.

I will not do it again. Sorry.

lovetox. I will send it to you via PM

Developers, I encourage you to consider XEP-0277. https://dev.gajim.org/gajim/gajim/-/issues/4501#note_215987 *Implement microblogging XEP*

Only adding a parser for such content (Atom Over XMPP) will do wonders in XMPP. ✏

I code with Python and I am willing to assist!

Gajim is not a microblogging client, and in fact viewing atom was once possible, but has been removed

maybe it can be implemented by plugin?

> Gajim is not a microblogging client, and in fact viewing atom was once possible, but has been removed cal0pteryx. This is not about microblogging (who has invented that strange word micro-blog, anyway?), and I am not fond of the word microblogging. ↺

This is about conveying information in a different view, and Atom Over XMPP iis a good solutio n to it.

As I said, there once was an Atom viewer, but it has been removed

I did not know that. May you please refer me to the commit of which Atom support has been removed? I am willing to restore that feature.

cal0pteryx, suppose this scenario: https://conference.gajim.org:5281/pastebin/3a75b8ae-ff84-4ef5-b5bd-a0716e95fd11 ✏

I have also depicted an interface to switch chat view and journal view. See https://video.xmpp-it.net/w/cfozoUeVLFbBFMCCSCJ1Dn?start=4m22s (Blasta - An XMPP PubSub Annotation Management System)

https://xmpp.pimux.de/file_share/459e4849-17ae-49f1-acab-f79d482aa826/psi_pubsub_interface_mockup.png

https://xmpp.pimux.de/file_share/0b4c77be-a4e8-4b41-b99b-e72b8331d400/psi_pubsub_interface_mockup_details.png

📔️ Atom PubSub posts. ~🔖️ Annotations (of the newly proposed Blasta system)~ You can ignore the latter.

> wurstsalat pushed 1 commits to branch gajim/master > > fix: StatusIcon: Use get_application_id for Gajim icon - https://dev.gajim.org/gajim/gajim/-/commit/d442bab052a77db53de5b2fe207a23dcfd116e5a > > Also, the messages of bot can be delivered to a PEP node `urn:xmpp:microbloggin:0` instead of interrupting this groupchat. ↺

And the messsages can be also delivered as XHTML with all content, which would not require to open the link to read the rest of the content. ✏

The benefits are great. Please think of it. And *I will assist* to add this feature.

> Also, the messages of bot can be delivered to a PEP node `urn:xmpp:microbloggin:0` instead of interrupting this groupchat. I was referring to a PEP node of JID gajim@conference.gajim.org ↺

sch: I'd be worried about feature and scope creep if gajim had an atom feed reader. *maybe* some pubsub features could make sense, but I think the maintainer's direction for gajim has been to focus on chat and chat exclusively which sounds reasonable to me. maybe through a plugin?

nicoco. Yes. A plugin would be good.

nicoco. concerning to features, I would suppose that it is more related to a proper interface.

nicoco. Did you receive an email with the deleted message of ticket #4501? If not, then I will send to you the content of that message.

I don't receive any notifications for tickets I am not involved in :)

lovetox pushed 1 commits to branch gajim/master imprv: Dataforms: Improve read only form presentation - https://dev.gajim.org/gajim/gajim/-/commit/23358e1f9c4fdd936f446c86e36ad33566fe5290

lovetox pushed 1 commits to branch gajim/master imprv: Control: Improve query to load messages around a timestamp - https://dev.gajim.org/gajim/gajim/-/commit/fa3ba1a2e92b44e703c978df0c4a8bdf3b5574de

Nice

HELP! How can I stop Gajim from downloading archived messages?

Two bots of mine have constantly sent messages to each other, and I now receive the whole message history of these messages. ✏

you cant

simply wait

Hello. On Gajim 1.9.3, is there a server admin menu or tool ?

Yes adhoc commands

you can either try Accounts -> Execute Commands

if the server offers the admin action on your account

if not you can use the accounts -> Service discovery and type your domain

and then select the top entry, and click commands

wurstsalat, is my PR good to go now?

lovetox, after I tried Account -> Execute Commands, I get en error message "item-not-found". I need to investigate on my server.

seb, are you an admin ?

Yes

I have no error on server logs

are you setting the admin = block globally or per vhost ?

marmistrz: not sure, I was only checking your changes, not making sure it works properly

it is always best for prosody if you use it to set the admin globally so it will be applied to muc components as well

Yes, it is set globally.

https://upload.slidge.im:5281/file_share/nWw6_iF6_pXg3cIWbTkDyRf4/Screenshot_20240910_204003.png

is this new way of displaying typing notifications intended?

Some update on dev.gajim.org being slow for me: now it just won't work at all via git+ssh (web UI works just fine).

SSH debug logs show it's stuck at `debug1: Connecting to dev.gajim.org [2001:41d0:203:1ce9:100::1] port 22.`

> Yes, it is set globally. So, to end my admin ad-hoc problem: I was miss configured my Prosody server instance. I have enabled globally the module pubsub and defined a pubsub component. After removing pudsub from modules_enabled, it work. ↺

If I force IPv4 it looks like it works but I cannot read my fork apparently? Exemple: ``` $ git push -4 git@dev.gajim.org:nicoco/gajim.git:test ... The namespace you were looking for could not be found ... ```

nicoco, it's an idea. the typing indicator in the chat banner on top makes little sense, it should be close to the bottom, where new messages appear. also it should not block the view, and the chat should not jump

I see, it does block the view if you want to see who is typing in a group though :)

ideas welcome

in group chats without description (the second line below the name), it was especially furiating, because the chat would scroll up a bit once somebody was typing :D

Maybe, instead of needing to click on it to see who is typing, "open" it while the mouse hovers?

> in group chats without description (the second line below the name), it was especially furiating, because the chat would scroll up a bit once somebody was typing :D Indeed, I noticed that ↺

I haven't yet figured out how to work the mouse-enter/leave event on overlay children :P

Ha! Joys of GUI :P

Doesn't it look better on the right side maybe? It hides the participant avatar of the last person to have sent a msg (if there is no row merging). Or maybe with some transparency? ✏

I tried opacity, but that looks weird to me

right-hand side can be far away, depending on your monitor

Well for arabic speakers (and other LTR languages) it is far away already from the "start" of the input box.

Anyway, I can live with it, but I'm not 100% convinced it's an improvement over the banner area in the current state. I do agree with your objections about it being up there.

Is inside the input box (on the opposite side of the "start" of the text) an option maybe?

yeah why not simple a line above the input box?

if nobodies typing it will simply look like a bit of space between the last message and the input box

~that would make the view "jump" I guess, unless the space is always reserved~ ✏

yeah always reserve it ..

I'm not fond of the idea of losing screen real estate, but this is less an issue for the elite HiDPI monitor users ;)

its one line of text

`Fixes "next" button staying greyed out...` is good to merge for me

Just as info, it seems conference.gajim.org has an ipv6 record but doesn't work with it. Some federating servers might get problems if they don't ha dle that well. > IP: 2001:41d0:203:1ce9:100::1 > Connection: [Not OK] > dial tcp6 [2001:41d0:203:1ce9:100::1]:5269: i/o timeout > IP: 91.121.56.21 > Connection: [OK]

^ This is consistent with my issues accessing dev.gajim.org with the git CLI

Thanks for investigating

When I discover the services of my server with the menu Account -> Discover services, then select a gateway and click on the `Browse` button, does it send a disco item request ?

seb: you can check for yourself by looking at the debug console's xml log

Ctrl+Shift+X

I checked and yes, it send a disco items request

Thanks

lovetox pushed 1 commits to branch gajim/master fix: Dataforms: Fix completion of single list forms - https://dev.gajim.org/gajim/gajim/-/commit/a783fb10194845088c89fd24147cbcba75a333c2

Hello, I'm doing some experimentatoin with omemo and I'm trying to get gajim to change its device keys (which seem to be accoutn specific, at least I have different fingerpritns for different accounts). I deleted and re-added the account in Gajim but somehow I get the same keys back. I also checked in $HOME/.config/gajim but I found nothing relevant there, including in the settings.sqlite database

so the question is... where are keys stored?

can i ask why are you trying to do this?

I set up my own sever and I was playing around with conversations (the android app). From their drocumentation is not clear what happens when a user you trusted connects with a new device. I would hope/expect that messages are *not* encrypted for the new device, so I'm trying to test this out. ✏

I set up my own sever and I was playing around with conversations (the android app). From their documentation is not clear what happens when a user you trusted connects with a new device. I would hope/expect that messages are *not* encrypted for the new device, so I'm trying to test this out. ✏

looked up in the source, it's in MY_DATA/omemo_ID.db, where MY_DATA is $HOME/share/gajim

> I set up my own sever and I was playing around with conversations (the android app). From their documentation is not clear what happens when a user you trusted connects with a new device. I would hope/expect that messages are *not* encrypted for the new device, so I'm trying to test this out. Is that not explained by the "Blind Trust Before Verification" setting? ↺

But if you specifically mean whether old messages are decrypted by new devices, no they are not, since OMEMO has perfect forward secrecy

> Is that not explained by the "Blind Trust Before Verification" setting? That's what I read, but it does not explicitly say that so I was trying to make sure ↺

> But if you specifically mean whether old messages are decrypted by new devices, no they are not, since OMEMO has perfect forward secrecy I was asking whether new messages are decrypted by new devices that have not been trusted yet ↺

Works as expected, I get a nice "This message was encrypted with OMEMO, but not for your device." message :D

steven3k: a good read: https://gultsch.de/trust.html

> steven3k: a good read: > https://gultsch.de/trust.html Yeah, I read that ↺

wurstsalat pushed 1 commits to branch gajim/master fix: JingleFileTransfer: Fix message data argument - https://dev.gajim.org/gajim/gajim/-/commit/19443c303595d32572575244ac61f1b1d597716a