Gajim - 2024-09-11


  1. fjklp

    > yeah why not simple a line above the input box? This is what I suggested in the issue that I can't find now

  2. fjklp

    I have to say, I dislike the current typing indicator. I can't come up with anything better than having a line above the input box

  3. fjklp

    there it is https://dev.gajim.org/gajim/gajim/-/issues/11913

  4. nicoco

    Hey, if you want to continue the merging spree, I think https://dev.gajim.org/gajim/gajim/-/merge_requests/998 is worth having a look, it's a +8 line diff. It takes advantage of the XEP-0172 `<nick>` element in message stanza from people not in roster, which is very useful for ugly JIDs, such as +123456789@quicksy.im or 12375678@transport.whatever. (for quicksy I'm not 100% sure the `<nick>` element is there, I don't have an account to test, but if it isn't it should be IMHO).

  5. fjklp

    I've got a few chats stuck in the joining phase but they receive messages from the muc

  6. cal0pteryx

    Is it the same server for these chats?

  7. fjklp

    same account different servers hosting the mucs

  8. fjklp

    same account, different servers hosting the mucs

  9. fjklp

    I've had this happen before

  10. fjklp

    I need to get better at debugging

  11. fjklp

    I think it's something going on with my account server. (there really ought to be an unambiguous term for this, like matrix uses 'homeserver'). Now one of the mucs is showing "connection failed" unknown posix error: timeout"

  12. fjklp

    but, it's weird that I'm apparently able to receive new messages in this one muc that is stuck connecting

  13. fjklp

    but, it's weird that I'm apparently able to receive new messages in this one muc that is stuck connecting. This I have seen before.

  14. Menel

    Seldom something like s2s can break for only one direction

  15. Menel

    Fun that if thinks is should be a posix error. And it is an xmpp error

  16. Mike Yellow

    An user came into a room, spoke something then left, then another user come with the same nickname, the client think they are the same person, so in history messages it changes the avatar of the previous user to the subsequent user's. Supporting XEP-0421 (occupant ID) would help to handle avatars correctly.

  17. Mike Yellow

    An user came into a room, spoke something then left, then another user come with the same nickname, the client thinks they are the same person, so in history messages it changes the avatar of the previous user to the subsequent user's. Supporting XEP-0421 (occupant ID) would help to handle avatars correctly.

  18. cal0pteryx

    Yes

  19. Mike Yellow

    Also it would be possible to display the same avatar for history messages with different nickname but same occupant ID, it reduces a little anonymity, though.

  20. debacle[m]

    Question about optimal workflow for a use case, I have very often: In chat "A" (no matter if one-to-one or MUC) somebody sends a cat picture. I want to pass that picture in chat "B" (again, one-to-one or MUC does not matter). At the moment I save the picture to `/tmp/` and then upload it again. Somehow this feels clumsy, like I miss an obvious shortcut. Ideally one, that does not involve using the pointer device (mouse). Is there any better way to do that?

  21. lovetox

    Right click copy url

  22. debacle[m]

    There is "Copy Link", OK. And then?

  23. lovetox

    You paste it into another chat

  24. debacle[m]

    > You paste it into another chat But then there is just an URL. It's not like HTTP upload.

  25. debacle[m]

    I.e. clients such as Gajim do now show the cat inline.

  26. lovetox

    Gajims does when you activate it

  27. cal0pteryx

    Ideally, we'd have a "Share" or "Forward" button, yes

  28. lovetox

    yes but probably at a point where there is more to share than the link

  29. nicoco

    I think even just a forwarding a link + the `<oob>` element that all clients expect would already be useful

  30. nicoco

    Ideally, some indication of the source of the forwarded message would be nice, but that's more complex with the current way attachments are implemented in most (all?) clients (one attachment=one message and no associated text)

  31. it20

    hello admin, i gave message here before about an issue.. after 1.7.1 in linux version of gajim i am facing some weird problem.. after adding 15 id next id go down below and there is no scroll button to find it.. and at the same time my writing window go down as well for all chats.. U told that it will be fixed but still i am facing the issue in 1.9.3 (flatpak version).. I think there is a problem in minimize and maximize button also.. there problems are co related i guess.. how can i fix it? or is there anyway u told devs to solve it asap.. i am facing a very hard time.. thanks

  32. fjklp

    >after adding 15 what does this mean?

  33. bot

    lovetox pushed 1 commits to branch gajim/master feat: Add setting to prevent sync of group chats - https://dev.gajim.org/gajim/gajim/-/commit/72340e4303d33926205e34f8f052d8097032608f

    🎉 1
  34. lissine

    >>after adding 15 > what does this mean? I think they meant 15 jid's

  35. cal0pteryx

    it20: I think I remember. It's about accounts being added to gajim. Once there are too many, they push down within the account sidebar. I fixed this issue in https://dev.gajim.org/gajim/gajim/-/merge_requests/1010 But this MR does a lot more than that. That's why it's still a draft. But that's what I wanted to work on anyway.

  36. cal0pteryx

    nicoco: thanks for all your MRs :)

    💚 1
  37. steven3k

    Trying to remove an account where the server is no longer reachable (actualy, it is reachable but no longer serves the same domain) causes Gajim 1.9.3 to get suck on "Trying to remove account"

  38. cal0pteryx

    steven3k: yes, there is an open issue about this.

  39. steven3k

    cal0pteryx, found it thanks. My first search found a chat log that claimed it was fixed in 1.8.4

  40. steven3k

    any workaround?

  41. bot

    lovetox pushed 1 commits to branch gajim/master revert: StatusIcon: Use get_application_id for Gajim icon - https://dev.gajim.org/gajim/gajim/-/commit/d3c1bdc37884a49d2ce037178465cec8b03668d7

  42. cal0pteryx

    steven3k: not that I'm aware of, sorry. There have been multiple bugs around removing an account. This one persists up to now

  43. steven3k

    cal0pteryx, removing the account while it is in the "Offline" status seem to work. But changing status to offline after trying to remove the account does not. So restart Gajim -> set account to offline -> remove worked.

  44. cal0pteryx

    steven3k: thanks

  45. Adrian Kieß

    Hello everyone. I wish you all good success, in your further doings. Adrian from Germany

  46. bot

    lovetox pushed 1 commits to branch gajim/master other: Flatpak: Don’t rename icon for nightly - https://dev.gajim.org/gajim/gajim/-/commit/b5f75e3df45b38276e4731e9a37b905dc5c1de6d

  47. lovetox

    Adrian Kieß, thanks !

  48. Adrian Kieß

    lovetox, You are welcome! Gajim 1.9 series is absolutely great piece of software! I use it under Debian Linux. I hope a newer series of Gajim. will be available for NetBSD pkgsrc, too.

  49. lovetox

    nicoco, i struggle to reproduce the nick change error on a server

  50. lovetox

    i think Gajim enforces stringprep on the nickname, so all nicknames should valid to a server

  51. nicoco

    lovetox, oh I linked the wrong issue

  52. nicoco

    https://dev.gajim.org/gajim/gajim/-/issues/11930

  53. nicoco

    this has an example of a nick that is accepted by gajim but rejected by prosody stable

  54. nicoco

    nicoco 👔 🔫

  55. lovetox

    ok thanks

  56. nicoco

    https://upload.slidge.im:5281/file_share/TIlIPg0_0uwrk99IaKMYuYAD/Screenshot_20240911_202140.png

  57. lovetox

    yeah its completely random which emojis are allowed and which not

  58. nicoco

    hehe I'm running with my patch so I see the error :D

  59. nicoco

    > yeah its completely random which emojis are allowed and which not believe me I know

  60. moparisthebest

    Stringprep with which version of IDNA and which version of Unicode? Also I think prosody now requires local nicks to be valid with stringprep *and* PRECIS (which I guess is your problem)

  61. nicoco

    it has been a PITA for me with my transports, not only emojis but various stuff, and AFAIU it's because slixmpp does not use the latest hot stuff to check what's allowed, and I already a lot more about that than I wish I had to (but not enough to fix things in slixmpp though)

  62. nicoco

    it has been a PITA for me with my transports, not only emojis but various stuff, and AFAIU it's because slixmpp does not use the latest hot stuff to check what's allowed, now I know a lot more about that than I wish I had to (but not enough to fix things in slixmpp though)

  63. moparisthebest

    PRECIS is another one of those sad mistakes where it'd maybe be nice to move to but it's not backwards compatible and there's no way to migrate to it :'(

  64. nicoco

    How come it's so hard to just have a reference JID validation lib with bindings for all popular languages? It's one of the things I don't understand, but again, I haven't dug deep enough and I'm sure I'm underestimating the amount of work it would be to make it. Also, I got bored so fast when I tried to investigate it than I guess I have a beginning of an explanation as to why: because it's damn boring.

  65. bot

    lovetox pushed 1 commits to branch python-nbxmpp/master other: Add tests for resource validation - https://dev.gajim.org/gajim/python-nbxmpp/-/commit/b44690dafeef45a0dcb9dd4561e65f20f94daf49

  66. lovetox

    nicoco, jid validation is trivial in high level languages like python

  67. lovetox

    the problem i had to solve now, which no lib will help you is giving you an emoji codepoint that is allowed

  68. hannibal

    > lovetox pushed 1 commits to branch gajim/master > > other: Flatpak: Don’t rename icon for nightly - https://dev.gajim.org/gajim/gajim/-/commit/b5f75e3df45b38276e4731e9a37b905dc5c1de6d That will most likely break nightly (though I cannot test right now). There needs to be an icon whose name matches the gapplication id.

  69. hannibal

    That icon will be shown in the application overview

  70. bot

    lovetox pushed 1 commits to branch gajim/master cq: Application: Silence pygobject warning - https://dev.gajim.org/gajim/gajim/-/commit/57750385c36f32515d3dbe681729fed729f841de

  71. lovetox

    hannibal, there is hannibal

  72. lovetox

    icons/scalable has the devel icon

  73. hannibal

    But that is not installed

  74. lovetox

    you mean in install_metadata.py?

  75. hannibal

    The line https://dev.gajim.org/gajim/gajim/-/blob/master/flatpak/org.gajim.Gajim.Devel.yaml?ref_type=heads#L421 also needs to be removed, when the Devel icon is installed

  76. hannibal

    > you mean in install_metadata.py? Yes

  77. lovetox

    and how would rename-icon change that?

  78. lovetox

    if its renamed before installing, it will be not installed

  79. hannibal

    https://dev.gajim.org/gajim/gajim/-/blob/master/flatpak/org.gajim.Gajim.Devel.yaml?ref_type=heads#L421 renames the devel file, than it is installed by your script and after that flatpak-builder renames it to the new name

  80. hannibal

    When rename-icon is set

  81. lovetox

    it renames all files named org.gajim.Gajim -> org.gajim.Gajim.Devel

  82. lovetox

    how can install_metadata.py then find org.gajim.Gajim? its not there anymore

  83. lovetox

    hence can not be installed

  84. lovetox

    what does `mv gajim/data/icons/hicolor/scalable/apps/org.gajim.Gajim{.Devel,}.svg` do?

  85. lovetox

    looks to me like moving the devel icon and the gajim icon, to the current working dir

  86. lovetox

    ok no that renames the devel file

  87. lovetox

    then my scripts installs it, and then you say flatpak renames it? but how would flatpak builder know where i installed the icon to

  88. lovetox

    haha whole stuff is another rabbit hole

  89. lovetox

    1. we are expected to install icons and desktop files into /app/... 2. flatpak exports files from there to ./local/share/flatpak if the files fit a specific convention

  90. lovetox

    1. we are expected to install icons and desktop files into /app/... 2. flatpak exports files from there to ~/.local/share/flatpak if the files fit a specific convention

  91. lovetox

    3. system adds ~/.local/share/flatpak to XDG_DATA_DIR

  92. lovetox

    so we need to export all icons, which we need other applications to access (e.g. libindicator)

  93. lovetox

    so the real problem here seems to be that we use an icon name for libindicator, that changes depending on install

  94. lovetox

    or we depend on the application id just on flatpak env

  95. lovetox

    and assume that nobody changes the app id on flatpak installs

  96. lovetox

    otherwise i dont see this as possible

  97. lovetox

    now its also clear why the pending message icon did not work on flatpak

  98. lovetox

    hmm i think the only way is to build in detection if we run as nightly

  99. lovetox

    hannibal, currently we place a file "run-as-flatpak" to discover this

  100. lovetox

    i think it would be nicer if we modify the source code, and change some variable

  101. lovetox

    or is there some kind of code checksum stuff that disallows us to change code while builing the flatpak?

  102. hannibal

    pre-build code changes are no problem, that's what the command section https://dev.gajim.org/gajim/gajim/-/blob/master/flatpak/org.gajim.Gajim.Devel.yaml?ref_type=heads#L417 does

  103. hannibal

    pre-build code changes are no problem, that's what the `commands` section https://dev.gajim.org/gajim/gajim/-/blob/master/flatpak/org.gajim.Gajim.Devel.yaml?ref_type=heads#L417 does

  104. lovetox

    ok, i think i identified multiple other problems that come with changing the application id for nightly

  105. lovetox

    did we do this only so we can run nightly side by side with the normal Gajim installation?

  106. lovetox

    probably flatpak does not allow us the same id as another package

  107. hannibal

    > did we do this only so we can run nightly side by side with the normal Gajim installation? I think that was the main reason (and of course the different config/data dir)

  108. hannibal

    flatpak apps can have the same id, but branch needs to be different if you want to install them at the same time

  109. hannibal

    To set the branch flatpak-builder can be called with `--default-branch=BRANCH-NAME`

  110. bot

    wurstsalat pushed 1 commits to branch gajim/master imprv: Unify usage of 'synchronise'/'synchronize' strings - https://dev.gajim.org/gajim/gajim/-/commit/8a1106a7289d5f8fe325a338e6131ae8db101822

  111. hannibal

    Though not sure if plugins can be told to only work with a specific branch

  112. bot

    wurstsalat pushed 1 commits to branch gajim/master refactor: GajimPopover: Make menu optional - https://dev.gajim.org/gajim/gajim/-/commit/2badeff8e4f5d5f81b70110a2b74b20e6258adfe

  113. mike

    Is there a way to verify OMEMO fingerprints in 1.7.3?

  114. lovetox

    yes, click the shild icon

  115. steven3k

    mike, there is a button on the top right with an "i", look in the OMEMO tab

  116. lovetox

    i doubt that existed in 1.7.3

  117. lovetox

    simply click shield icon in chat

  118. steven3k

    lovetox, oops sorry

  119. lovetox

    hannibal, could you please review https://dev.gajim.org/gajim/gajim/-/merge_requests/1076 and test the nightly build

  120. mike

    Oh yeah duh, tapping the shield worked, thanks!

  121. mike

    I was looking under Manage accounts, where it does show for me in 1.9.3