Gajim - 2023-07-04

Hey, does anyone here know why I can't see profile pictures and nicknames of other users who I haven't added as friends yet? Even if they've set their profile to "public"?

what do you mean by can't see nicknames?

as in, it only shows the username part of their JID (eg. "alex" from alex@denshi.org as opposed to the published nickname, "Denshi") until I add them as friends and they add me back. Same thing for avatars

Because we don't request them

lovetox, updated https://dev.gajim.org/gajim/gajim/-/merge_requests/965

> Because we don't request them Is that a design decision? I was looking through the advanced configuration editor to see if there was a setting somewhere, but there doesn't seem to be one. I understand the privacy implications of allowing it though.

re MR 965, I believe that end to end encryption isn't possible when using bridges. So it'd be cool if bridge usage could be somehow detected and gajim prompts the user to disable e2ee for those specific chats

> re MR 965, I believe that end to end encryption isn't possible when using bridges. So it'd be cool if bridge usage could be somehow detected and gajim prompts the user to disable e2ee for those specific chats Of course it is. ↺

What do you mean by bridge? Something like biboumi? Public groupchats like this one?

lissine, ^

I checked that even if encryption-by-default is enabled, public groupchats like this one will still not be encrypted, as expected

> What do you mean by bridge? Something like biboumi? Public groupchats like this one? biboumi, as well as slidge I'm not a slidge user myself, but I believe one to one conversations can be bridged

nicoco: what do you think about this? > re MR 965, I believe that end to end encryption isn't possible when using bridges. So it'd be cool if bridge usage could be somehow detected and gajim prompts the user to disable e2ee for those specific chats

I am not a client dev, but I don't think bridged contacts needs special handling. They just won't advertise that they can do OMEMO, and clients should not try to send OMEMO-encrypted messages to them?

exactly. they wouldn't have omemo keys published so omemo-encrypting a message to them wouldn't be even possible

That's why I said it's more than just selecting a default. We need sane error handling and comfortable user guidance.

If a groupchat doesn't support OMEMO (like this one), it didn't try to use OMEMO when I tested

cal0pteryx, ^

What if you cant fetch keys? What if the server prevents omemo from working?

Then you should get the standard errors as though you enabled it manually

that's what Conversations does too

lovetox, I found that “Forget this Group Chat” does the both. It will delete the bookmark, and leave the room too. Maybe you should change the text to “Leave and forget this Group Chat”?

Or give it the function that can join an room without adding the room into bookmarks automatically?

> Maybe you should change the text to “Leave and forget this Group Chat”? If you decide to do this, the Simplified Chinese translation is “离开并忘记此群聊”.

Philipp Hörist pushed 1 commit to branch _refs/heads/master_ of _gajim_ < https://dev.gajim.org/gajim/gajim >: *0cc9b18f* < https://dev.gajim.org/gajim/gajim/-/commit/0cc9b18f0d3b0621c85da8a237f099690283b4fb > feat: Add setting for default encryption

☭Mike Yellow, there is no reason to click "forget" if you dont want to leave and forget about it

i think the text is fine

Fair. Understandable.

Hi friends, I'm doing some dev with Gajim on a local version of Prosody, with self-signed certificates, and can't seem to find how to get Gajim to accept the certificate authority enough to upload files.

Incoming files work, though with no preview (and the HTTPS verification toggle doesn't seem to do anything).

But outgoing fail with "The signing certificate authority is not known". I seem to remember this working in the past, though -- perhaps something changed in config? Anything I can check?

Link Mauve: do you manage the AUR gajim package? omemo_dr is missing from its dependencies

cal0pteryx, it lists python-omemo-dr in the depends array.

Strange

deuill, no there is currently no way short of changing a few lines of code

httpupload.py -> _accept_certificate() needs to return True

Thanks lovetox, I'll just do that for now!

I suppose the function should respect the prior trust though, right? I might get to fixing that if you agree.

Hmmm even with `_accept_certificate` being bypassed, there's a follow-up error in `nbxmpp`: ``` nbxmpp.http Request(139667248114304): g-tls-error-quark: Unacceptable TLS certificate ```

Oh well

I'm trying to figure out how `get_emoji_data` works. It loads a gresource /org/gtk/libgtk/emoji/en.data, but I can't find any corresponding file in the /usr/share/gtk-3.0/emoji/ directory. Moreover, once I properly understood what the code is supposed to be doing, I see that the processor indeed works.

Where is the corresponding gresource actually stored? I can't find anything corresponding to the english file in my /usr/share/gtk-3.0

mjk, hannibal, ^ any secret knowledge? :)

I don't have /usr/share/gtk-3.0/emoji/en.blahblah but english data loads just fine, it's probably compiled in, or in another dir somewhere

as a last resort, you could probably find it or its sources at GTK's gitlab

marmistrz, gresource is a mechanism which appends data to an ELF.

That way you don’t have to distribute so many files with your application.

So the gresource is probably in the /usr/lib/libgtk-4.so file.

Link Mauve: libgtk-3 you mean :P

Err, yes.

Sorry, I don’t live in the past. :p

I figured as much ^^

deuill, ups, seems when we refactored the http code we forgot to actually issue this signal

so it does nothing currently

i dont see a easy way to circumvent that for you

Philipp Hörist pushed 1 commit to branch _refs/heads/master_ of _python-nbxmpp_ < https://dev.gajim.org/gajim/python-nbxmpp >: *eec01dea* < https://dev.gajim.org/gajim/python-nbxmpp/-/commit/eec01dea39ac47c45ebb39667e3fc3c2372c28e5 > fix: HTTP: Add set_accept_certificate()

deuill, if you use python-nbxmpp from master

you can suply a func with set_accept_certificate_func() which when returns true should work

> Philipp Hörist pushed > feat: Add setting for default encryption it's happening!!1

now some people could start enabling this and report if things get funny in corner cases

and we have a transition path towards omemo-by-default!

So cool to see people excited about the MR I filed :D

stopped working as a programmer, finally have time and energy for hobby programming xD

thanks for the contribution

I tested it and it seems to work

I guess it keeps the last encryption setting for a private chat even if the chat has been closed?

Hi. First I'd like to thank you for Gajim. I have some feedback about the UI. When I first started using Gajim, I found it confusing to log in/log out. There's an `Accounts` menu at the top but it doesn't let you log in/log out. What do you think about adding an entry there?

Second, in the conversations list on the left, for conversations that have unread messages, what do you think of making the last message bold?

I know there's the message counter, but it's easier if it's bold too

(The second idea is inspired from the conversations client, also the name of the contact/channel would be bold too)

lissine, login/logout is a concept of websites

they keep you logged in because of cookies, and you actively have to logout

in Gajim you can add an account, and it will connect to the server. If you dont want that anymore, you can deactivate the account

understood :)

but in any way this is a action someone will do on very very rare occasion, so having it in a menu on one click seems unnecessary

about your other idea, making the name bold, i would have to see how that looks, but im not against trying, maybe someone has the time to try it and post some screenshots

https://share.ellini.org:5281/file_share/bAS3JiNP2NOMOrFHsVLlNr-H/52ZQh_tzStOl9bto-Q-p-Q.jpg

this is what I mean, you can see that conversations that have unread messages have the name and last message in bold