Gajim - 2018-03-22

  1. mrDoctorWho so, what about message input box bug?
  2. spicewiesel good morning all
  3. spicewiesel Asterix, lovetox: that's what is shown in the XML console when connecting the irc chat:
  4. marmistrz Hi! The latex plugin doesn't render anything for me.
  5. marmistrz The only thing in the logs is:
  6. marmistrz ``` 22.03.2018 08:33:06 (E) gajim.plugin_system.omemo python-qrcode or dependencies of it are not available 22.03.2018 08:33:06 (W) gajim.plugin_system Plugin plugin_installer not loaded, plugin incompatible with current version of gajim: 1.0.0 > 0.16.9 ```
  7. mrDoctorWho tabs are not highlighted when user's composing a message/paying attention to the conversation, etc
  8. spicewiesel mrDoctorWho, yes... and the indication for new message is nearly not notable (for me)
  9. 0xAFFE it sometimes depends on the font and channel name, at least under windows, on my linux work setup the tab strip has a good readability
  10. lovetox yeah spicewiesel, this seems to be a problem with disco the muc, i look into fixing this
  11. spicewiesel lovetox, cool, thx
  12. Lenny Hi guys, is there any possibility how to save roter windows position under gajim 1.0.0? Every time I closed it, it's reopened on different screen, which is really uncomfortable with multihead desktops
  13. Asterix Lenny: yes it's fixed for next version
  14. Lenny good
  15. Lenny already in repository? will get master branch
  16. Asterix marmistrz: it works for me. Plugin enabled? All is ok when you click the test button in pref window of the plugin?
  17. Asterix Lenny: which repos? Debian?
  18. marmistrz Asterix, plugin enabled, all self-checks are OK
  19. Asterix marmistrz: the plugin makes latex things clickable. You click on it and image is generated
  20. marmistrz Asterix, oh, I thought it would auto-render, thanks!
  21. marmistrz that was not intuitive
  22. Asterix We don't auto-render for security / perf reasons
  23. Asterix Lenny: it's fixed since yesterday only.
  24. Wiktor Hello, is there a plan to migrate to "final" HTTP Upload namespace (`urn:xmpp:http:upload:0`)? I noticed that Gajim still uses the old one (`urn:xmpp:http:upload`):
  25. lovetox Wiktor, i would have to check if all servers support that
  26. lovetox because i dont want to maintain 2 versions
  27. lovetox and there is no gain for the user
  28. Wiktor I was just digging through Prosody modules and both http_upload and http_upload_external support the new namespace. But maybe some older servers do not? hmm
  29. Wiktor Yeah, actually I'm writing http upload component and wondering why are they using two namespaces. Conversations uses :0 and Gajim uses bare.
  30. lovetox i write it to the list of things for 1.1
  31. Wiktor 👍 great!
  32. Wiktor Hmm... now that I think about it :0 may not be so easy to implement, there is added support for specifying headers for PUT request (that was absent before) like `Authorization`...
  33. Ge0rG Hi! I have an instance of Gajim 0.16.6 that seems to do GC1.0 joins. Is that a bug or is the client just desync?
  34. Wiktor FTW I filed an issue for `:0`
  35. lovetox im not aware what a GC1.0 joi is
  36. lovetox and im not aware what it means that a client is desync
  37. Ge0rG lovetox: GC1.0 join is a directed presence to a MUC without an <x/> element. Being desync in that context means the client thinks it's joined and sends a directed presence without <x/>, but the MUC thinks the client isn't there any more (i.e due to s2s errors), and sees the new presence as a join instead.
  38. lovetox so we should always add x?
  39. lovetox when we want to jon
  40. lovetox but never add it if we are joined?
  41. Ge0rG lovetox: exactly
  42. Ge0rG lovetox: are you?
  43. Ge0rG lovetox: a precise answer to this question is important, because I want to make joining-without-x illegal
  44. Ge0rG lovetox: rationale is at
  45. lovetox i never looked at it, and im on work and cant
  46. lovetox but i can give you an answer in the evening
  47. lovetox if it is not that way i can change this of course in the next version of gajim
  48. Ge0rG lovetox: I'd appreciate that very much. Also feel free to read through that mail, it might give insight into syncing MUCs
  49. lovetox k thanks
  50. Ge0rG Though I think that sending <presence unavailalbe> right in front of a join isn't required any more.
  51. Darlan lovetox:
  52. lovetox yeah seems obvious whats the problem here
  53. lovetox bash: flatpak-builder: command not found
  54. lovetox install flatpak-builder
  55. Darlan oops!
  56. Darlan It seems that there is no such package at SBo
  57. zuglufttier I'm using gajim 1.0.0: If I click on the arrow to hide the muc participants, the muc roster is hidden. But if I change the tab to another tab, it will be open again if I go back to that tab.
  58. lovetox please open a bug report about it
  59. Darlan lovetox: me?
  60. lovetox no
  61. zuglufttier Me :D
  62. zuglufttier Will do!
  63. lovetox i cant help if your system doesnt provide flatpak-builder
  64. lovetox though i find it very weird that it would provide flatpak, but not flatpak-builder
  65. Maranda lovetox gajim doesn't support xep-96 transfers?
  66. Maranda
  67. Maranda lovetox, sorcery used by iMessage
  68. debacle If I select multiple contacts in the roster and send a group message, it seems that Gajim uses full JIDs for users who are currently online, but bare JIDs for users who are currently offline. I had expected, that bare JIDs have to be used in both cases. Can somebody enlighten me?
  69. Ge0rG sending messages to full JIDs is frowned upon, indeed.
  70. debacle Ge0rG, now we can frown together upon Gajim :~)
  71. Ge0rG debacle: open a ticket? ;)
  72. debacle yes, I will
  73. debacle I checked with the XML console, that Gajim uses full JID indeed
  74. debacle However, it uses it also for offline users in some cases, for whatever reason
  75. Ge0rG A full JID for offline users?
  76. debacle Yes. Maybe the "last known" client?
  77. Ge0rG Probably
  78. lovetox debacle its not a bug to adress a message to a full jid
  79. lovetox for online users, offline is another story
  80. debacle lovetox, but it is useful to address the full JID?
  81. debacle even for online users?
  82. lovetox its infact written down in the RFC that a client SHOULD do this
  83. debacle oh, wowo
  84. debacle oh, wow
  85. Ge0rG lovetox: the RFC isn't up-to-date.
  86. lovetox im not arguing if this is the best way or not, but it is certainly no bug
  87. lovetox and if you cant describe a concrete problem while gajim does this
  88. lovetox then its not worth to even look at it
  89. Ge0rG lovetox: sending messages to full JIDs causes issues when the receiver goes offline
  90. Ge0rG lovetox: ejabberd is violating the RFC routing rules because users complained that messages sent by Gajim were lost.
  91. lovetox you mean on theoretical server installation without carbons
  92. lovetox and mam
  93. lovetox otherwise i could not imagine how a message got lost
  94. lovetox but even without this, how is the message exactly lost? if gajim does not know that a client is offline, then the server doesnt know that it is offline, if i send a message and the server cant route it, would he not have to put it into offline message storage?
  95. Ge0rG lovetox: there are race conditions between gajim sending the message and noticing that the receiver is offline
  96. Ge0rG lovetox: message loss will hopefully only happen with type=normal, not with type=chat.
  97. lovetox why whats the difference with chat?
  98. Ge0rG lovetox: the RFC
  99. Ge0rG lovetox: page 8
  100. lovetox there is a issue open about removing resource locking
  101. Daniel Asterix, is the gitlab bot down?
  102. lovetox but i have to check all implications in gajim code
  103. Ge0rG lovetox: the consensus from last Summit is: keep resource locking for IQs, but remove it for messages
  104. Ge0rG lovetox: also don't depend on client advertised features for things like 0184 acks and LMC, because it will end up in MAM/Carbons anyway.
  105. lovetox we do not
  106. lovetox lmc is always used
  107. lovetox and 0184 also
  108. Ge0rG lovetox: that's great
  109. debacle Ge0rg: What's broken in "208-01-30.00" and how can fix it? :~)
  110. Ge0rG debacle: :P
  111. Asterix Daniel: why?
  112. lovetox Asterix it does not seem to work
  113. lovetox i just closed mutliple issues and there is no message
  114. lovetox Ge0rG, so the x should only be in join presences is this correct?
  115. lovetox this is already done by Gajim 1.0.0
  116. Asterix A problem with the upgrade then. API maybe changed ...
  117. Asterix We never did GC1.0
  118. lovetox we maybe do on reconnect ..
  119. lovetox i have to check what happens on reconnect
  120. Ge0rG lovetox: yes, <x/> only on join.
  121. Ge0rG lovetox: I'm seing periodic GC1.0 joins from 0.16.6, and also some from 1.0.0, but they might well be false positives due to desync
  122. lovetox but on desync x should also be added or not
  123. lovetox i can certainly say that on join we use x
  124. Ge0rG lovetox: when you think you are not joined right now, send <x>
  125. Ge0rG lovetox: but sometimes the server will think you are not joined when you just send a presence update
  126. lovetox so i checked, on disconnect we rejoin on coming online
  127. lovetox on sm resume we do nothing
  128. lovetox or should we do something on sm resume Ge0rG
  129. SaltyBones Ge0rG: lovetox, wanted to know why he should remove resource locking.
  130. SaltyBones I added your links to the issue but they are a bit general. ;)
  131. lovetox we already talked about it today SaltyBones
  132. lovetox :)
  133. SaltyBones Kk :)
  134. SaltyBones Timestamp? :D
  135. lovetox 17:40
  136. Ge0rG lovetox: you could send a self-ping on resume, but that's not so important. You probably should send periodic self-pings anyway
  137. Asterix indeed there is a 500 error in gitlab when I test the bot ...
  138. Asterix that really soudn a gitlab bug ... The hook doesn't seem to be called at all ...
  139. bot Yann Leboulanger pushed 3 commits to branch _refs/heads/master_ of _gajim_ < >:
  140. Asterix hello I'm back !
  141. Asterix For those interested:
  142. pep. Asterix, interesting, thanks
  143. bot Philipp Hörist closed an issue in _gajim_ < >: #8984: < Hidden muc roster is shown if a previously minized tab is opened >
  144. bot Philipp Hörist pushed 1 commit to branch _refs/heads/master_ of _gajim_ < >: