Gajim - 2025-09-24

hi, I'm recently encountering a problem where whenever I receive voice messages (or audio files in general), gajim downloads the audio as it should but I can't play it at the same time, I have to close the chat and reopen it in order to get the playback UI (with play, speed etc.)

it only apperead recently as it worked fine before

it's almost like it gets stuck at 100% downloading

maximus: thanks, we currently rework this code, will take a look

keep up the great work!

Good Day

Are the devs aware that when sending a long text which contains code and gajim collapses the message in the middle of the codefence, then the highlighting is lost?

(it is applied once you unfold the message)

No, probably the highlight check is done after cutting the text

Please create a issue

lovetox: did you fix the issue with d&d pictures or did it fix itself because the latest gnome update changed something?

I didn't fix anything

well it works

:)

https://dev.gajim.org/gajim/gajim/-/issues/12475 ← screenshots are attached

Hello

> https://dev.gajim.org/gajim/gajim/-/issues/12475 ← screenshots are attached ❤️ ↺

On Windows, is there a location inside the the `UserData` folder to put your own hunspell dictionaries?

not in userdata

but somehwere in the other folders

check /share/hunspell

I know, I just wanted to know if it perhaps also searched for dictionaries somewhere inside UserData. OK, then I'll have to adjust my scoop manifest to keep `share/hunspell` in its persist directory. ✏

no Gajim does no search, this is automatically done by the spelling libs

and im not aware of any way to change that

Thx

there's not anything particularly useful in gajim's debug logs

it just says 'no route to host' when connecting to nuegia.net because it nerver tries the other SRV records

I'm only able to connect now because I turned on the tor proxy, which apparently has better ipv6 support

but this is not ideal

ideally it should try every possible SRV record both v4 and v6 until it successfully connects

maybe GLib has no support for trying multiple srv entries

anyway, you can specify the host yourself in the account connection settings, no need for tor

> maybe GLib has no support for trying multiple srv entries wait what? ↺

srv records require trying multiple until you successfully connect, that's like the basic design in the original RFC

i said maybe, i dont know and never checked, is this something you would expect from a network lib?

I've never seen a network lib that supports SRV at all, that's why I'm surprised

This call looks up the SRV record for service at domain for the “tcp” protocol. It then attempts to connect, in turn, to each of the hosts providing the service until either a connection succeeds or there are no hosts remaining.

reads like it does support it or?

tom: please provide an account on your server, so i can test this myself

> tom: please provide an account on your server, so i can test this myself I can do that, but is this something that cannot be troubleshooted with host and netcat? ↺

> This call looks up the SRV record for service at domain for the “tcp” protocol. It then attempts to connect, in turn, to each of the hosts providing the service until either a connection succeeds or there are no hosts remaining. "connection succeeds" how is this defined? If just "TCP accepted" it's an incorrect implementation ↺

i dont think that is relevant

as tom says there is no connection made at all

lovetox, check your DMs for the invite code

It only tries the IPv6 SRV record, gets a 'No Route to host' error and then never tries the fallback SRV which contains an IPv4 record

but ipv6/4 has nothing to do with srv records

but it's possible it has early-abort bugs triggered by certain responses I guess

Which glib function is this?

https://gitlab.gnome.org/GNOME/glib-networking

tom, i can connect fine

what version?

[2600:3c0a:e002:bb00:d250:99ff:fe3f:3832]:5222 [2600:3c0a::f03c:95ff:fed8:2321]:5222 172.234.232.197:5222

these ips are all tried and it succeeded then with ipv4

can you check your glib version in the about dialog?

yeah manually connecting works

but gajim doesn't read alternate SRV records on FreeBSD

>[2600:3c0a::f03c:95ff:fed8:2321]:5222 Where did you get this address? That should not work. I don't know who that is

it does resolve both srv entries for me

here a command to get more debug output

ok

set this environment var

G_MESSAGES_DEBUG=all

then start gajim

brb

and here my logs ``` (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.137: GSocketClient: Starting new address enumeration (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: GSocketClient: Address enumeration completed (out of addresses) (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: GSocketClient: Address enumeration failed: No DNS record of the requested type for “_xmpps-client._tcp.nuegia.net” (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: GSocketClient: Connection failed: No DNS record of the requested type for “_xmpps-client._tcp.nuegia.net” (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: g_socket_client_connect_async: starting new g_socket_client_connect_async() with GTask 0x8bfdd60 and GSocketClientAsyncConnectData 0x8d90fd0 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: GSocketClient: Starting new address enumeration (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: lookup_by_name_with_flags_async: starting new lookup for xmpp.nuegia.net with GTask 0x8d911c0, LookupData 0x8db5710 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.196: lookup_by_name_with_flags_async: starting new lookup for xmpp.nuegia.net with GTask 0x8bf8cb0, LookupData 0x8db85f0 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.227: GSocketClient: Address enumeration succeeded (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.227: g_socket_client_enumerator_callback: starting connection attempt 0x8db6a90 for GSocketClientAsyncConnectData 0x8d90fd0 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.227: GSocketClient: Starting TCP connection attempt (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.228: GSocketClient: Connection attempt failed: Network is unreachable (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.228: GSocketClient: Starting new address enumeration (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.231: IPv4 DNS error: Error resolving “xmpp.nuegia.net”: No address associated with hostname (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.231: lookup_by_name_with_flags_async: starting new lookup for r2.sea.us.nuegia.net with GTask 0x8dc1990, LookupData 0x8d82780 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.231: lookup_by_name_with_flags_async: starting new lookup for r2.sea.us.nuegia.net with GTask 0x752f1b0, LookupData 0x8db5710 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.265: GSocketClient: Address enumeration succeeded (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.265: g_socket_client_enumerator_callback: starting connection attempt 0x8d91af0 for GSocketClientAsyncConnectData 0x8d90fd0 (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.265: GSocketClient: Starting TCP connection attempt (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.502: GSocketClient: TCP connection successful (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.503: GSocketClient: Starting application layer connection (launch.py:252349): GLib-GIO-DEBUG: 23:59:41.503: GSocketClient: Connection successful! ```

what i see but probably not relevant, you have no xmpps-client records, this is often tried first by clients so it would be good if you have them

seconds it says `IPv4 DNS error: Error resolving “xmpp.nuegia.net”: No address associated with hostname`

which is probably nothing good

but you see in the logs it first tries ipv4/6 for xmpp.nuegia.net fails, then tries for r2.sea.us.nuegia.net and succeeds

and here your question who the ip belongs to

https://share.hoerist.com/philipp/message/qTfFS8MdIHXIwgIG/aa91965e-4e8d-428a-8e1b-b2af2716e81a.png

It may say connection succeeded but it doesn't connect

GUI says connecting still and theres no protocol traffic

then start gajim with -v at the same time so we get what gajim does also in the log

but i have to go to sleep now

I see it

Also with wireshark

I think somebody is tampering with the TCP connection after gajim sends its connect string

As I send that and the TCP connection stalls out

I send keyboard mashing I get the usual malformed XML error

Spooky

Its a 'stream error'

Something similar is happening with IMAP connections too with a fake TCP connect, but there's no responses even to keyboard mashing

What are some things I can do if there's a mitm attack?

In witeshark after the client sends its handshake string, the TCP just infinitely TCP retransmits

it's probably not a mitm attack, that's like the least plausible explanation and Occam's razor says don't jump to that one

What else could it be, looks like mitm to me!

plenty of things, bugs most likely

Whatever it is its not gajim's fault

I see gajim connecting

have you tried connecting from the same machine with say telnet ?

But something between us is hijacking the TCP session and preventing flow after the first handshake

> have you tried connecting from the same machine with say telnet ? Yes ive been testing this with netcat ↺

Its shady af

Some sensitive ports TCP connect too like ssh and IMAP and SMTP but do not send anything

what does netcat do

tom: I get it with some accounts, or even certain MUCs.

it could be as simple of a bug as sending data but not calling flush()

Not in netcat

that'll work fine sometimes, but not always

tom: only in Gajim tho. I also use updated Gajim on FreeBSD

I don't know what netcat impl you are using or when/if it flushes