Gajim - 2018-01-19

  1. bot André proposed a new merge request for _gajim/master_ <>: location_listener: location update notifier requires three parameters
  2. bot André updated a merge request for _gajim/master_ <>: location_listener: location update notifier requires three parameters
  3. David I just upgraded from Debian Jessie to Stretch (MX16 to MX17). I now can't add python-farstream to allow voice chatting! Help?
  4. David Anyone using Gajim on Debian Stretch?
  5. ‎David‎ +1
  6. bot Philipp Hörist pushed 2 commits to branch _refs/heads/master_ of _gajim_ <>:
  7. bot Philipp Hörist closed an issue in _gajim_ <>: #8856: <_on_location_plugin_update() error>
  8. bot Philipp Hörist merged a merge request for _gajim/master_ <>: location_listener: location update notifier requires three parameters
  9. Maranda Ohhh synchronise history works too :O
  10. Maranda 🤣 🤣 🤣
  11. zuglufttier What exactly does "synchronize history" do? Is it using MAM?
  12. debacle Can I see my account passwords somehow?
  13. debacle (in 1.0.0-alpha2)
  14. bot mt created an issue in _gajim_ <>: #8857: <Gajim crashes when opened inside an rdp session>
  15. Link Mauve Wasn’t this already fixed by not using Gdk.Screen’s physical size?
  16. mt has anybody encountered "hanging" characters in the message window when scrolling through the roster ?
  17. mt whoops
  18. mt wrong window
  19. mt nevermind
  20. nico will sync history work with prosody 0.10 mam ?
  21. pitchum debacle, I can find my own password in cleartext in ~/.config/gajim/config
  22. Link Mauve debacle, it depends whether you’re using a libsecret provider or not.
  23. Link Mauve If you have e.g. gnome-keyring or ksecretservice, query them; otherwise where pitchum said.
  24. debacle pitchum, Link Mauve: Cool, some passwords are in the config, others are in gnomekeyring :~)
  25. debacle Thanks!
  26. Maranda Okay and there goes #publish-options
  27. Maranda I dunno if I want to implement xep-363
  28. bot samweisgamdschie created an issue in _gajim_ <>: #8858: <portable 1.0.0a2 crash after config>
  29. Maranda didn't bot have some commands like ping?
  30. Maranda zuglufttier, and yes it uses MAM
  31. lovetox Link Mauve, yes fixed, but the issue creator uses not HEAD
  32. bot Growl created an issue in _gajim-plugins_ <>: #260: <PGP plugin - Own messages is not shown>
  33. bot Growl modified an issue in _gajim-plugins_ <>: #260: <PGP plugin - Own messages are not shown>
  34. lovetox Sync History is when you have 10.000 messages in your MAM history and you want to download all of them
  35. Maranda hmm why does OMEMO needs #publish-options?
  36. Maranda I don't get it
  37. Maranda (not from the xep at least)
  38. lovetox where does it say it needs that?
  39. Holger Maranda: Conversations uses publish options to make the OMEMO nodes world-readable so that people can access them even if they're not on my roster.
  40. Maranda I can't find, but both Gajim and Conversation want #publish-options
  41. Maranda Ohhh ok
  42. lovetox gajim does not want it for omemo
  43. lovetox gajim wants it to publish bookmarks
  44. lovetox and make them private
  45. Maranda XD
  46. Maranda Ok fine, well I suppose that's good for both
  47. Holger In the bookmarks case that's what the XEP says.
  48. lovetox the idea of publish options is that i can publish a item to pubsub and configure the node in one flow
  49. Maranda lovetox, will Gajim use the whitelist access model to make 'em private right?
  50. lovetox no gajim will use publish options
  51. Maranda I stand my case
  52. Maranda *will it use the whitelist access model right*? 😉
  53. lovetox the server should set then the whitelist model to whatever was specified in the publish options
  54. Maranda (you specify it through publish-options)
  55. lovetox yes compared to configuring the node manually
  56. Maranda Ok
  57. lovetox i could also pull the config, then change node settings
  58. lovetox so publish options is a kind of shortcut to that
  59. lovetox does your fork have persitent pep?
  60. Maranda cause whitelist in Metronome is a bit of a hack, *but still compliant*, it should work.
  61. Maranda yes
  62. lovetox thats great
  63. lovetox :)
  64. Holger The shortcut makes especially sense because it would be a little ugly to re-configure the node on each and every publish.
  65. bot Growl created an issue in _gajim_ <>: #8859: <Unusable popups>
  66. Holger ... while you do want to make sure the access model is as desired on each and every publish.
  67. Maranda Holger, that I'm not sure if I currently support, XEP-60 is a bit hideous on that and what I understood is that if the node exists the server by default should check if the publish-option matches the node configuration as "pre-condition"
  68. Maranda else refuse the publish, or can behave differently but it's not really well written.
  69. Holger Maranda: If you think it's hideous today you should re-read the previous revision a few months ago :-)
  70. lovetox Maranda, i think overwrite is the best behaviour
  71. Holger IMO publish options were completely underspecified and therefore unusable back then. Which rendered XEPs such as the bookmarks one broken.
  72. Maranda Holger, there isn't a part describing *how* you retrieve publish-options, I had to guess it by looking at schemas
  73. lovetox publish options should have been added to some xmpp registry where you can look them up but nobody ever did that
  74. Holger Maranda: But yes today it works the way you describe it. The behavior is so far only specified for the access_model option, and that's defined to be a PRECONDITION.
  75. Holger Maranda: "Retrieve publish-options"?
  76. Holger lovetox: The registry exists in 0060. And it only has the access_model option.
  77. Maranda How do you know which publish-options the server supports, it would need to retrieve it from the server like the node configuration form tbh Holger.
  78. Holger lovetox: There's been some recent discussion whether to add more options or whether to instead just specify that all node config options should work as PRECONDITIONS.
  79. Maranda the publish-options form I mean
  80. lovetox Maranda thats not important
  81. Holger Maranda: Ah, yes there's no way to do that.
  82. lovetox if a publish option is not supported by the server it should reject the publish
  83. Maranda Holger, well according to schemas there's
  84. Holger Maranda: You use the publish option and the server will reject the publish if it doesn't understand it.
  85. Holger Maranda: Huh.
  86. lovetox A pubsub service advertising support for publishing options MUST reject publications with unknown fields.
  87. Holger Trial'n'error still sucks sometimes.
  88. Maranda You send a publish-option payload like configuration on pubsub#owner
  89. bot Growl created an issue in _gajim_ <>: #8860: <Lock icon on encrypted messages>
  90. Maranda according to schemas
  91. Maranda or I got it wrong
  92. Maranda but seems sensible
  93. Holger "Dear user, please enter your blog post here:" ... then when pressing the "Send" button "Oh, dear user, sorry, your server refuses your long text."
  94. Maranda lol
  95. Maranda well maybe it's uncovered
  96. Maranda but what I did seemed to have a sense tbh
  97. Holger Maranda: Sounds sensible yes. But when we discussed this last time nobody was aware the Schema suggests that.
  98. Maranda Holger, prolly it doesn't, i think it specifies the form format, but it could be viewed both ways
  99. Maranda Or at least at the very moment it seemed logic 🤣
  100. Holger Whatever, somebody[tm] should sit down and fix publish-options. They have been fixed to address the immediate use case of OMEMO/bookmarks, but it's still crap.
  101. Maranda Holger, right now I implemented it so that it checks if the publish_model is either open, or you have "publish" access else it'll refuse to hand out the form model.
  102. Holger Ah.
  103. Holger Well I don't think the client will get any other server to return that form under any circumstances :-)
  104. Maranda Holger, it'll work anyways
  105. Maranda XD
  106. Holger I fully agree it makes sense and it should go into the spec.
  107. Holger Or, well.
  108. Holger Another sane idea is IMO to just define what I mentioned above: all node config options can be specified as publish options and are handled as PRECONDITIONS.
  109. Holger Ditch OVERRIDE and METADATA.
  110. Maranda I mean I looked through the spec up and down, and I was "how the hell a client would know which fields does a pep service supports" 🤣
  111. Holger Or if you need that, invent separate syntax for that.
  112. Holger Then you no longer need a separate publish options form.
  113. Holger
  114. Holger That's what I'd do.
  115. Maranda Not that I ever seen a client implementation ever retrieving form fields tbh
  116. Maranda hmm 🤔
  117. Holger Yes the forms idea was mostly to throw it at the user without having the client understand the semantics, of course.
  118. Holger I.e. like Psi/Gajim/whatever do for MUC config.
  119. Holger But that's not how stuff usually works these days.
  120. Maranda well I agree on the 1:1 mapping tbh
  121. Holger Yes anything else seems really weirdo to me.
  122. Maranda in Metronome's publish-options I implemented at least max_items, persist_items, access_model and publish_model. Tbh following the example just access_model seems silly.
  123. Holger But strictly speaking you must reject anything else :-)
  124. bot Philipp Hörist modified an issue in _gajim_ <>: #8860: <Lock icon on encrypted messages>
  125. Holger Maranda: Tomorrow, persist_items might be registered as an OVERRIDE (hah!) and tomorrow's client can then depend on you either handling it that way or rejecting it :-)
  126. Holger Maranda: (In other words, I agree the idea is nonsense.)
  127. lovetox so you did not add access_model?
  128. Maranda Holger, I think atm Metronome just ignores the fields it doesn't understand so I'm probably not very compliant there. But with clients sending just the crap they like without checking it was a required COMPAT measure.
  129. lovetox its the most important option
  130. Maranda lovetox, err of course I did
  131. lovetox k
  132. Maranda I just added the other 3 too
  133. lovetox yeah good, i think that enough for now, im not aware that anyone uses other options
  134. Holger Maranda: Just ignoring is slightly hairy (if you announce publish-options support) I think. Imagine just ignoring access_model.
  135. Holger But not sure there's other options with similar security implications, and there's probably nothing else used in practice right now.
  136. Maranda lol, that's supported, but still in the past i had pubsub stuff just sending configuration forms filled with like 1 or 2 fields unsupported while the rest was and was just rejecting the configuration stanza breaking those clients
  137. Maranda had to change it so it just ignore what it didn't understand for that reason
  138. Holger But did you have clients sending publish-options?
  139. Maranda Nope
  140. Maranda it was just about node configuration, and on create configuration
  141. bot Philipp Hörist closed an issue in _gajim_ <>: #8857: <Gajim crashes when opened inside an rdp session>
  142. Maranda Holger, but still I support all that's require *for now* I think.
  143. Holger Yes if you support access_model, Conversations and Gajim are happy, and if those two are happy, everyone is happy.
  144. Maranda haha
  145. lovetox maranda do you have a server running with metronom?
  146. lovetox where i can register an account
  147. Maranda
  148. lovetox k, interested to test the publish options, the code was forever in gajim, but no sever supported it for a long time
  149. Maranda I think it's one of the few Metronome "Highlanders", although the others won't be uptodate with the changes I recently made so they're useless.
  150. Maranda lovetox, report any bug plx :P
  151. andy Hello, Anyone know why Gajim not detect correctly certificate of my domain connected to a says that is ok.
  152. andy I've got message: "Certificate for my_awesome_domain.tld" issued for
  153. bot andy created an issue in _gajim_ <>: #8861: <Gajim detect wrong certificate for domain hosted on server>
  154. lovetox i also got this lately
  155. lovetox but im not exactly sure if gajim does something wrong here
  156. lovetox Holger, do you have maybe an idea?
  157. andy detect certs of "my server" corectly
  158. lovetox maybe a problem with an intermediate certificate
  159. lovetox i have to look into it when im home
  160. bot andy modified an issue in _gajim_ <>: #8861: <Gajim detect wrong certificate for domain hosted on server>
  161. andy lovetox, thx.
  162. andy Another strange thing is a issued date of a certificate.
  163. andy 2018 01 18
  164. bot andy modified an issue in _gajim_ <>: #8861: <Gajim detect wrong certificate for domain hosted on server>