Gajim - 2018-03-15

  1. Link Mauve “22:37:28 lovetox> never had this, i dont know why gtk just crashes on such errors can it not just print the thing and go on”, it depends on how you compiled it, rom1dep probably chose to make asserts fatal (so they are easier to debug), while most distributions opt for just printing them.
  2. rom1dep Link Mauve: I didn't do anything special, it's fedora's default. I don't care enough about Gtk to compile my own :)
  3. Link Mauve Check what Fedora’s default build options are, then.
  4. Link Mauve It’s surprising they would do this, though.
  5. Link Mauve rom1dep, also make sure you don’t have G_DEBUG=fatal-critical in your environment.
  6. rom1dep Link Mauve: this is the package , build logs with configure options: , nothing G_DEBUG in my env.
  7. marmistrz Asterix, may it be that this issue with corrections is due to some prosody module missing
  8. marmistrz ?
  9. Link Mauve Hmm, nothing seems wrong here.
  10. Link Mauve So maybe another issue.
  11. lovetox marmistrz, no corretion has nothing to do with the server
  12. lovetox what are you saying? it still does not work for you?
  13. lovetox where are you trying this?
  14. lovetox in a MUC, in single chat?
  15. UsL hello peoples. How do I make Gajim follow the system clocks 24h format? As it is now it only goes to 12 and have AM/PM after it. I have no idea what that means.
  16. Link Mauve UsL, probably by changing LC_TIME to a locale which does 24h.
  17. UsL I think that is set system wide. Gajim is the only program that does not follow it
  18. UsL is it a setting in Gajim somewhere?
  19. Link Mauve That’s probably more of a GLib thing than a Gajim thing.
  20. UsL hm..
  21. lovetox i think we can format it in ACE
  22. lovetox but dont know for sure
  23. Link Mauve Oh?
  24. UsL I am on 0.16.6 btw.
  25. lovetox i cant look right prefernes -> advanced -> advanced config editor
  26. lovetox maybe search for "times"
  27. lovetox value in config is called
  28. lovetox time_stamp
  29. UsL thanks! I'll look
  30. lovetox what you probably want there is [%H:%M]
  31. UsL yep
  32. lovetox Link Mauve, we use normally [%X] for this value
  33. UsL ah, I guess I need to log off to make it take
  34. lovetox so somewhere your time locale is wrong on your system
  35. lovetox UsL no just close the window
  36. Link Mauve lovetox, that’s the preferred time in the current locale, right?
  37. lovetox yes
  38. lovetox but we do this with python
  39. Link Mauve Makes sense.
  40. lovetox so there is no GLib involbed
  41. Link Mauve Ok.
  42. lovetox *involved
  43. UsL hm, I have the one window for all layout so I guess I need a restart
  44. lovetox yeah
  45. UsL oh I also wondered about pgdwn/pgup. It doesn't scroll the backlog
  46. UsL I'll look in that settings window, btw maybe it is there. This is my first time using Gajim..
  47. lovetox no you can set shortcuts yourself
  48. lovetox *cant
  49. UsL so there is no way to make those keys work?
  50. UsL that is a bummer being a keyboard guy..
  51. UsL are there other keys that scroll the backlog? So that I didn't misunderstand.
  52. Asterix There is a shortcut wiki page. Look there
  53. UsL ah, shift. Okay, great.
  54. 0xAFFE I got this trace, the plugin has Version 2.1.6, I do not know how this happened:
  55. lovetox thanks 0xAFFE
  56. louiz > why would this id not be incremented if 2 messages are received in the same second > not much of an id if its no unique lovetox: it's unique, juste not necessarily incrementing (if you delete a record and restart biboumi for example)
  57. lovetox why do you not just use a autoincrenting row index as mam id
  58. lovetox louiz
  59. lovetox what you have seems error prone
  60. lovetox you should only have to ever sort your request results after the row id
  61. lovetox there is just no need to bring timestamps into this and then deal with such issues as not enough resolution on the timestamp like this commit
  62. Link Mauve lovetox, it shouldn’t be a normal operation to delete a MAM record.
  63. Link Mauve The only two deletion-related operations are: deleting from the origin to some point (which can be all of it), or replacing a message with a tombstone.
  64. lovetox im not getting why this has something to do with deleting records
  65. lovetox or does the db fill wholes in the row id index
  66. lovetox *holes
  67. Link Mauve lovetox, he said that’s the only possibility for having a duplicated id.
  68. lovetox ok maybe im missing something here, every db has a unique id feature
  69. lovetox thats really basic
  70. lovetox it seems only natural to use this field as mam-id
  71. lovetox there can never be 2 entrys with the same unique id, as the db would not allow to add such records
  72. louiz I'm not talking about uniqueness, but ordering
  73. louiz and you can't rely on the id for this
  74. lovetox why not?
  75. louiz because it doesn't work
  76. louiz proof: I had my archive in the wrong order
  77. lovetox because it seems you have rows with the same ID
  78. louiz no
  79. lovetox is the order you put in data into the db not the order you received the messages?
  80. Holger FWIW the reason ejabberd re-uses the timestamp as the ID is that you need to store the timestamp anyway and I see no advantage in having a separate counter column. (But I don't understand the ordering issue with an autoincrenting ID either.)
  81. louiz no, lovetox
  82. lovetox Holger, what resolution do you have on the timestamp?
  83. lovetox i remeber it beeing a long number
  84. lovetox so i guess 1.000000
  85. Holger lovetox: Yeah, microseconds. (There was also a uniqueness guarantee in the past but that's dropped now.)
  86. lovetox because its not possible in practice i guess only theoretical
  87. lovetox louiz, so you should tackle that problem, that your messages are not written in order to the db
  88. Holger Right. The guarantee is somewhat expensive so not worth it.
  89. apollo13 BTW ordering by auto incrementing IDs goes to shit if you ever use more than one database
  90. lovetox im sure there are DB solutions that can cluster and still guarantee uniqunes
  91. lovetox and order
  92. lovetox but i think there is no xmpp server that has that many users that it would have the need to spread the load over multiple dbs
  93. lovetox simple solution i just came up with would be to not let the DB generate the row id, instead pull it from a source that guarantees you only to give the same id once and in order
  94. Link Mauve You want to destroy any benefit of having multiple locations, basically.
  95. Link Mauve Also, I call bullshit on them all being too small.
  96. Link Mauve You just don’t see them in your daily life.
  97. louiz lovetox: they are written in order in the db
  98. lovetox so then i still dont understand your problem, just have a autoincrementing row index
  99. lovetox then sort after that
  100. apollo13 lovetox: sure you can cluster and guarantee uniqueness -- but generally you'd try to avoid that if at all possible
  101. apollo13 Especially for insert heavy scenarios
  102. bot shtrom created an issue in _gajim_ < >: #8948: < Incorrect focus for new windows/dialogs >
  103. bot shtrom created an issue in _gajim_ < >: #8949: < Show file transfers/thumbnails within conversation windows >
  104. louiz It's already autoincrement SERIAL, in postgresql), and it doesn't work
  105. louiz And I don't have a problem, it works properly as it is now
  106. louiz Also, if I decide to import old data from an other database, it should be sorted before the more recent ones. Only the dztestamp can provide that
  107. bot Lonnie created an issue in _gajim_ < >: #8950: < Cannot get group chats to work >
  108. bot Philipp Hörist pushed 1 commit to branch _refs/heads/master_ of _gajim_ < >: *8f2bc1e1* < > Handle connection lost notification click Fixes #8823
  109. bot Philipp Hörist closed an issue in _gajim_ < >: #8823: < "A programming error has been detected" >
  110. bot Philipp Hörist closed an issue in _gajim_ < >: #8381: < gajim shows error messages after closing the window >
  111. bot Philipp Hörist closed an issue in _gajim_ < >: #8155: < Gajim don't want to start >
  112. bot Philipp Hörist closed an issue in _gajim_ < >: #8799: < A programming error has been detected >
  113. bot Philipp Hörist closed an issue in _gajim_ < >: #8931: < unable to reconnect manually >
  114. bot Yann Leboulanger closed an issue in _gajim_ < >: #8949: < Show file transfers/thumbnails within conversation windows >
  115. lovetox Daniel,
  116. lovetox
  117. lovetox sadly we have no hi res icons for mood and activity
  118. lovetox i want to switch them to noto but mapping the moods to the emoticons is tiresome
  119. lovetox i also added tooltips now