Gajim - 2017-02-19

  1. de-facto .
  2. lynxis i've the problem that gajim isn't reconnecting automatic after a disconnect (e.g. because of suspend). I don't use the network manager, so how can I notify gajim via d-bus or other api of the network become available again?
  3. lovetox i think you cant lynxis
  4. thomas Good morning! I want to use mod_mam (with prosody) in gajim but cant find any option to activate.
  5. lovetox there is no activation
  6. lovetox your server announces its capabilities at the start
  7. lovetox and gajim acts on it
  8. lovetox gajim doesnt support MAM for MUC though
  9. lovetox only single chat
  10. thomas oh, ok in this case, it seems mam isnt working on my prosody
  11. Link Mauve thomas, note that the mod_mam available in Prosody 0.10 is for version :1, while Gajim still implements only version :0.
  12. thomas ok, so what can I do?
  13. Link Mauve There was a community module in prosody-modules, which you could use until someone fixes Gajim.
  14. Link Mauve You could also just fix Gajim.
  15. Link Mauve Note that Conversations is planning on updating to version :1 soon as well.
  16. thomas I'm not a programmer, so fixing gajim is not an option...
  17. thomas and I run prosody 0.9.12-1~jessie1
  18. lovetox then maybe you use the wrong module, maybe ask in the prosody channel which module you have to install for mam:0 to work
  19. thomas thank you!
  20. lovetox generally you dont have to configure anything in gajim regarding that
  21. lovetox but you can look at the xml console in gajim
  22. lovetox at the start
  23. lovetox then you see if your server announces support and if gajim querys mam and what the result is
  24. thomas ok
  25. SaltyBones test
  26. thomas Hello, I cant select a few menu entries in gajim "actions" menu
  27. thomas "discover services" for example
  28. thomas nothing happens
  29. lovetox are you on ubuntu?
  30. lovetox you have to start the applications with UBUNTU_MENUPROXY=0
  31. lovetox to make menus work
  32. lovetox thomas
  33. Link Mauve lovetox, why don’t you just set this environment variable in the launcher?
  34. lovetox because i know nothing about how linux applications are started
  35. Link Mauve And you didn’t think of asking a Linux user about it?
  36. Link Mauve Just read the file.
  37. Link Mauve lovetox, does it still need that on master?
  38. lovetox no master doesnt have that problem
  39. Link Mauve Ok, then I don’t care. ^^
  40. lovetox but is not just executed on the bash
  41. lovetox would not every command i give there effect the whole system?
  42. lovetox i think this would probably better be placed into gajim.desktop
  43. Link Mauve But then anyone starting Gajim from the command line wouldn’t have it set.
  44. Link Mauve Another option would be at the very beginning of
  45. lovetox yeah also what would the command do on non ubuntu systems
  46. Link Mauve It’s not a command, it’s an environment variable.
  47. Link Mauve And it would do nothing.
  48. Link Mauve Just clutter the environment.
  49. lovetox ah ok so i just edit the env from python
  50. lovetox i will try that
  51. lovetox hm i think i found a major bug
  52. lovetox in the gtk3 branch it seems we have a endless loop in drawing the roster
  53. lovetox and its not clear if this is a gtk bug
  54. lovetox calling the append method on a treeview with GLib idle_add causes a endless loop in gtk
  55. lovetox ok not a gtk bug
  56. bot Philipp Hörist pushed 2 commits to branch _refs/heads/master_ of _gajim_ <>: *dd3a2f3a* <> Fix endless loop in various CellRenderers context.get_background_color() first calls the CellRenderer to render the cell before giving back the color. When we call this inside a CellRenderer, this is an endless loop *18df7163* <> Merge branch 'roster' into 'master' Fix endless loop in various CellRenderers See merge request !59
  57. lovetox 3 days for one commit
  58. zak I am so happy, that you do not really love tox, lovetox :-)
  59. lovetox haha :)
  60. zak There will not be any new-feature version before Gajim 0.17 which will be with GTK3, correct?
  61. lovetox i think we call it 1.0
  62. lovetox but yeah, i want to release the gtk3 soon
  63. zak "soon"? :-D
  64. lovetox im working now on plugin installer and the most important plugins
  65. lovetox so that people can switch without losing too much
  66. lovetox i dont know really maybe one month if all goes good
  67. zak cool
  68. zak don't rush things. a stable version is more important than an early version
  69. arune lovetox: 👌👌👌
  70. Floatingterror Hi. Can someone explain to me how the google translator plugin works?
  71. andrey.g Are there real bugs, gtk3 branch resolves, which is not possible with gtk2?
  72. Link Mauve andrey.g, the aforementioned issue of not having menus working on Ubuntu, for example.
  73. Link Mauve Also, gtk2 didn’t support Wayland.
  74. Link Mauve It also required an X11 server on OS X.
  75. andrey.g Link Mauve, menus sounds like something must be working on gtk2, if not just a gtk2 bug? Wayland - ok. Did gtk3 switch to something else than X11 on OS X?
  76. Link Mauve No, it’s an Ubuntu bug.
  77. Link Mauve andrey.g, yes, on OS X they use the native way to display windows.
  78. Link Mauve Just like on Wayland.
  79. andrey.g I've missed that with OS X, thanks. Is this a well-known (long standing) bug on Ubuntu with a report?
  80. rougerabbit Mir soon
  81. arune andrey.g:
  82. andrey.g arune, thanks. Hmm, submenu is something usual, what many applications have.
  83. arune andrey.g: but they are probably static, ie generated once at startup
  84. arune Whereas gajims menus change depending on accounts
  85. andrey.g arune‎, thanks for pointing this out, that's really makes difference.
  86. andrey.g Does somebody chat in languages, requiring more than one keyboard layout? It is luckily possible to pin different spell checkers to different tabs (it would be even better if Gajim would remember this!), but how to pin different keyboard layouts?
  87. Link Mauve andrey.g, when I used to use this feature, it did remember.
  88. Link Mauve As for the layout, I think that’s impossible.
  89. andrey.g Link Mauve, after Gajim close/open?
  90. andrey.g Same unresolved issue: "Different keyboard layouts for different IM tabs and windows"
  91. Link Mauve andrey.g, yes, but I stopped using it around maybe 2010, so I have no idea if this feature hasn’t been broken at some point.
  92. Link Mauve andrey.g, maybe report an issue.
  93. lovetox the problem on ubuntu is not ubuntus fault, is how gajim builds his menus, that does not work with global menus anymore in ubuntu
  94. Link Mauve andrey.g, this feature request at Pidgin’s bugtracker should be closed, since it is impossible AFAIK.
  95. andrey.g Link Mauve, ok.
  96. zak lol... "Opened 10 years ago, last modified 6 years ago"
  97. zak Pidgin development is probably almost dead it seems.
  98. andrey.g I think, it should be just forwarded to mate/gnome/... and/or gtk, qt,..
  99. andrey.g It looks really sad with Pidgin.
  100. zak What is really sad is, that so many people still use it. :(
  101. zak I just queried the 12 blogs linked in the news section of Pidgin. None of them has a recent update within the last 2-3 years. Some of them are even not existent any more.
  102. andrey.g zak, I treated Gajim as an in-between solution due to high memory consumption, but as it looks, I've sticked with Gajim, looking forward for something more lightweight (not necessary jabber)...
  103. zak Not Jabber? How can you even think of something else!!! ;)
  104. andrey.g I really like an idea to have a protocol neutral GUI with arbitrary protocol implementations as plugins with respective extensions.
  105. andrey.g zak, once I've figured out, that Jabber is about instant messaging not necessary reliable.
  106. zak When is jabber not reliable today?
  107. andrey.g lovetox‎, you mean it could be Unity bug / "feature"?
  108. lovetox no gajim just builds menus in away that is not supported by global menus
  109. andrey.g zak, on not persistent connections.
  110. lovetox andrey.g nobody would use conversations if it would be not reliable on non persistent connections
  111. lovetox its as reliable as any other protocol
  112. lovetox even more reliable with the right extensions
  113. lovetox or any other mobile jabber client
  114. zak andrey.g‎: Read here
  115. andrey.g lovetox‎, gtk allows to implement something not supported by what?
  116. lovetox of course you can write code with gtk that doesnt work, why shouldnt it be possible?
  117. lovetox we build our menus when you click on them, with the click event
  118. lovetox there is no click event sent from a global menu on system layer
  119. lovetox so there is no menu build
  120. andrey.g lovetox‎, store-and-forward looks reliable because as long as you haven't yet got an ack, you suppose the message is not sent, once you've got it, you're 100% sure, the other side is now responsible for the message. KISS. How the message transmission is done in Jabber with all the extensions, I don't look through and I haven't seen any diagram, showing a way of the message with at least only typical errors. The other algorithm: synchronization of message history also looks obviously reliable.
  121. andrey.g zak‎, thanks, I'll look into, but also no diagrams.
  122. andrey.g lovetox‎, it's possible to emit events programmatically.
  123. andrey.g System layer?
  124. lovetox why are we talking about this?
  125. lovetox gtk2 has no support for global menus
  126. lovetox ubuntu just takes the menu of the programm so it looks like other linux unity programms
  127. lovetox but doesnt emit the same events as when the menu is on gajim
  128. lovetox you can either rewritte all the menus to not be build on click event
  129. lovetox or you just add one line to your .desktop file
  130. lovetox and be fine
  131. lovetox and about reliable of the protocol
  132. lovetox mobile jabber clients would not have such a upcoming if messages would be lost
  133. lovetox you can browse through 2000 issues on conversatiosn github
  134. lovetox you will not find one that is about losing messages
  135. lovetox and mobile connection is the worst case
  136. lovetox so please hold back with bold statements like "jabber is not reliable"
  137. andrey.g I thought, Unity handles all things transparently to programs, so no changes are needed and what does work on not-Unity, must work also on Unity, otherwise it is a Unity bug. So ok, done with it.
  138. andrey.g lovetox‎, I'm not bold categorical (see added word "necessary") ;)
  139. andrey.g zak, look at the following example: I've sent a message with Gajim, but this time it has not reached the server yet, and I have to close Gajim now. After I start it again, what happens?
  140. zak If you close any messaging application (or even e-mail) before the sending-process is completed, I guess no protocol can ensure anything.
  141. andrey.g Clear. In an e-mail application I see the process of sending, but in Gajim atm. I don't see when my message reached my server.
  142. zak If you close the application cleanly (i.e. no kill -9 or anything like that), it should finish the process. Again, the same problem applies to any other messaging application.
  143. andrey.g write() finished successfully, but OS couldn't really yet transmit this away. So application is closed cleanly.
  144. zak It is TCP, so I think an ACK belongs to the process.
  145. zak What you mean would be UDP
  146. andrey.g No, it is really TCP: write() doesn't guarantee that the recipient has received anything, since it is buffered. Also TCP has buffers.
  147. andrey.g So "stream resumption" seems to me not to apply to the case application closing during connection loss (which something you cannot easily observe, otherwise you could take this into account).
  148. andrey.g These ACKs, you're speaking about, are exactly application level ACKs that are mandatory even using TCP. TCP acks are rather for OS.
  149. zak Exactly. So if you close the app, the TCP connection is closed cleanly, and no data is lost.
  150. zak Everything else would not be compliant behaviour.
  151. Link Mauve “20:06:51 andrey.g> Clear. In an e-mail application I see the process of sending, but in Gajim atm. I don't see when my message reached my server.”, Gajim knows it though, because it asks the server for an ack whenever it sends any stanza.
  152. andrey.g While the recipient is still not reachable, the TCP connection and application are closed. What happened to the buffered packets?
  153. Link Mauve That’s so it can resume the stream in case it notices it’s been disconnected.
  154. Link Mauve It’s not used in-between restarts though.
  155. Link Mauve What could be done though is waiting for all of the pending acks to arrive, before closing.
  156. Link Mauve andrey.g, recipient doesn’t matter with MAM.
  157. Link Mauve Note that I’m supposing you use a correctly-configured server, with MAM, Stream Management, Carbons enabled.
  158. andrey.g Link Mauve‎, thanks! This deserves a bug report: closing Gajim in case missing acks for last message(s) could then show a Window "Closing: waiting for server response...", which could be interrupted, then user would be aware of his responsibility, that some messages could be not yet sent.
  159. andrey.g Link Mauve, > recipient doesn’t matter with MAM Sorry, I don't get this.
  160. Link Mauve andrey.g, it doesn’t matter whether your recipient has zero, one or more resources connected, as long as all of their clients are using MAM and their server is correctly configured.
  161. Link Mauve (Which, to my knowledge, counts only Gajim and Conversations currently.)
  162. lovetox also you are talking about 2 different things here
  163. lovetox 1. the protocol itself lacks something, that makes messages go lost -> i think this is not the case
  164. lovetox 2. gajim does not transparently inform the user about everything that happens with the message
  165. lovetox 2. is certainly for some cases a valid statement
  166. lovetox now we could talk about how likely your cases are and how often they happen, and how much work it is to cover those
  167. andrey.g Link Mauve, thanks. Now I've noticed, that I used word "recipient", while I meant actually my server as recipient (TCP connection between Gajim and server).
  168. Link Mauve andrey.g, ah, then Gajim always knows whether a particular message has been received by your server or not.
  169. Link Mauve (As long as your server supports Stream Management, and you didn’t disable that in Gajim.)
  170. andrey.g lovetox‎, maybe I'll open the bug report about this.
  171. andrey.g 0198 talks also of group of messages, so I thought, not every message must be acked. But if Gajim explicitly asks for every message not making use of grouping, then good.
  172. Link Mauve andrey.g, check your XML console to be sure, because a client could indeed chose to only ask for acks after some timeout.
  173. lovetox that would be not good :)
  174. lovetox like that you end up with double sended messages
  175. andrey.g OK, will check this. In general there are for each "Out" an increased "h" value. But sm:2 means that 6 year old sm:3 is still not implemented? Another one "rare case, not deserves time spending"?
  176. andrey.g lovetox‎, why?
  177. lovetox sm:3 is implemented in nbxmpp
  178. lovetox and also when it was not, there were no complaints about lost messages
  179. Link Mauve lovetox, why is Gajim still using :2 then?
  180. lovetox i dont think gajim uses nbxmpp
  181. lovetox if you have an old nbxmpp you will have 2
  182. lovetox *-dont
  183. Link Mauve andrey.g, are you not using the latest nbxmpp?
  184. Link Mauve 0.5.5 that is?
  185. andrey.g On Debian gajim depends on python-nbxmpp. And it is of version: 0.5.4
  186. lovetox yeah nothing wrong with that, we didnt increase the min nbxmpp version for this
  187. andrey.g It has been packaged on Thu, 13 Oct 2016, so fresh, not 5+ years old.
  188. lovetox actually sm3 is not in 0.5.5 , its only in nightly
  189. lovetox or its partly in 0.5.5
  190. andrey.g Aha, so it is bleeding edge for nxxmpp.
  191. andrey.g Aha, so it is bleeding edge for nbxmpp.
  192. lovetox tis basically bleeding edge for everyone, i dont think many clients support sm3
  193. lovetox or to be clear, the namespace bump to sm3 has happend longer ago
  194. lovetox but recent changes to the xep are not even a year old
  195. andrey.g BTW, I've heard message correction is not supported in MUCs. Am I wrong? (I've corrected my previous message.)
  196. lovetox it is supported in gajim
  197. andrey.g So it doesn't depend on 1-1 or MUC?
  198. lovetox no
  199. andrey.g ok.
  200. lovetox its just another message, that has a tag with the message id for which it should be a replacement
  201. lovetox if a client honors that or how he displays it
  202. lovetox is up to the clinet
  203. arune andrey.g: conversations does not correct messages from users not in your roster
  204. arune So in anonymous muc they aren't corrected (unless your the room manager)
  205. andrey.g arune‎, does it drop these corrected messages?
  206. arune andrey.g: is displayes both
  207. arune
  208. arune andrey.g: ⬆️
  209. andrey.g Hm, it squashes them in the same message.
  210. arune andrey.g: no, it does that for messages sent under a small timeframe
  211. arune See lovetox messages at the bottom
  212. andrey.g good to know (in this sample he didn't use correction there)
  213. gandogar Hello, I want to use omemo-e2e in muc. Which conditions must be satisfied?
  214. lovetox