Gajim - 2016-04-18


  1. bot RSS: Feeds for Gajim • Changeset [16019:b19bd9be4a87]: Use the timestamp of the forwarded message in mam processing if … Use the[…] https://trac.gajim.org/changeset/b19bd9be4a8729c1b929ebae282b627c4a04a718 • Ticket #8277 (Random old messages appear in open chat windows on reconnect (without …) closed fixed: In 16019:b19bd9be4a87: Use the […] https://trac.gajim.org/ticket/8277#comment:15
  2. tmolitor asterix: I tested PIL with your new windows binary and it is working as expected :))
  3. tmolitor asterix: sadly omemo isn't working...I'll talk to kalkin about that...
  4. bot RSS: Feeds for Gajim Plugins • Changeset [769:f0d3759795e1]: Better handling of thumbnail creation and some preparations for py3… Better handling of thumbnail creation and some preparations for py3… https://trac-plugins.gajim.org/changeset/f0d3759795e167f649748526591f046b7bfe754b
  5. lovetox tmolitor: does the image preview plugin also decode incoming encrypted omemo pictures?
  6. lovetox i only get it decrypted when i am copying the incoming link and send it back
  7. tmolitor lovetox: yes, I added this functionality 2 month ago...
  8. tmolitor lovetox: that's strange...i shouldn't matter if the link is incoming or outgoing...
  9. lovetox tmolitor: and do you see any chance that the decryption process is faster in the future?
  10. lovetox why can conversation decrypt the pictures so much faster?
  11. tmolitor lovetox: that's because there is no native C implementation of aes_gcm for python...
  12. lovetox picture decrypt works as expected now, sorry, it seems i didnt wait long enough, 1 MiB pictures needs more then a minute here to decrypted and im on a high end cpu
  13. tmolitor the gcm math has to be done in pure python code at the moment...
  14. tmolitor lovetox: I know...you can even feel the delay for longer text messages ;)
  15. tmolitor but the next version of pycrypto will support aes_gcm as far as I know...
  16. lovetox ok nice, thanks anyway for coding that, 1 minute is not too bad versus never :D
  17. lovetox tmolitor: the version that supports gcm (pycrypto) seems to be in alpha state for 2 years
  18. lovetox what about that: http://pycryptodome.readthedocs.org/en/latest/src/introduction.html
  19. lovetox it is a fork and provides gcm as it seems
  20. tmolitor lovetox: yes, I (and kalkin, the author of the omemo plugin) know...but the problem is, that this fork is not api compatible but it uses the same module name....so you cannot install both in parallel...
  21. tmolitor lovetox, right...I know...I think this is the reason for the cryptodome fork...
  22. tmolitor but there is so much software out there which uses the original pycrypto library that it is unsure if pycryptodome would ever take pycrypto's place...
  23. tmolitor lovetox: in my opinion the pycryptodome developers took a very bad decision when they did not rename the module to something new so that both libraries could be installed together...
  24. lovetox tmolitor: ah ok, from there wiki page they decribe a way to install it seperatly
  25. lovetox pip install pycryptodomex In this case, all modules are installed under the Cryptodome package. PyCrypto and PyCryptodome can coexist.
  26. tmolitor lovetox, oh okay...that sounds good...
  27. tmolitor nice to hear that I'm wrong (I (or: kalkin, to be correct) did my research a few month back)
  28. tmolitor maybe they changed this since then or we did miss this feature...
  29. lovetox but it seems there is another problem at least for windows
  30. lovetox https://github.com/Legrandin/pycryptodome/issues/14
  31. lovetox i believe gajim uses cx_freeze?
  32. tmolitor I don't know exactly...
  33. tmolitor asterix is packaging the windows installer...
  34. tmolitor kalkin recently did some optimisations of his aes code, too...I'll ask him if there is anything I can copy to the url_image_preview plugin...
  35. tmolitor the current roadmap for the py3 version of gajim is a generic crypto api for all plugins, so that plugins can provide crypto functions and other plugins or the gajim core can use this to send/receive encrypted files and so on...
  36. tmolitor lovetox: so, someday this will certainly be improved :)
  37. lovetox nice to hear
  38. tmolitor lovetox: yes :) but this will take some time, since I'm very busy with my own small company and university :/
  39. lovetox yeah yeah i know, there is so much to code, sometimes i want to dive in myself
  40. lovetox but i also have a job which takes most of my time
  41. lovetox sometimes i think i just should pay someone to fix the things i want to have :D
  42. tmolitor lovetox: I had the same idea, too :D
  43. tmolitor that's why I put a $50 bounty on implementing cloud notification in chatsecure (a ios xmpp app)...but sadly it didn't help...
  44. lovetox what site did you use for the bounty?
  45. lovetox i wondered myself how i would do this best
  46. lovetox i read these days the whole problem with GPL and axolotl on the apple app store
  47. tmolitor bountysource, I think...
  48. lovetox so basically there will be no omemo for a long time
  49. lovetox on iphone
  50. lovetox was really sad to read that :/
  51. tmolitor yes...if the license isn't changed...but the chatsecure people seem to try to negotiate a license change...
  52. tmolitor well..that's apple...unfortunately...
  53. tmolitor https://github.com/ChatSecure/ChatSecure-iOS/issues/367
  54. lovetox i think these talks are over :/
  55. lovetox https://github.com/anurodhp/Monal/issues/9
  56. lovetox He (Moxie) also mentioned there are no plans to relicense this code to others (except to WhatsApp, I guess)
  57. tmolitor well...that's sad...but maybe there is somebody who codes his own version under a better license :)
  58. bot RSS: Feeds for Gajim • Ticket #8329 (Pressing Clear Devices for OMEMO cause this fault) created Bug description Traceback (most recent call last): File "C:\Users\XXX\AppData?\Roaming\Gajim\Plugins\omemo\ui.py", line 61, in on_click self.plugin.clear_device_list(self.contact) File "src\plugins\helpers.py", line 107, in wrapper File "C:\Users\XXX\AppData?\Roaming\Gajim\Plugins\omemo\init.py", line 435, in clear_device_list iq = DeviceListAnnou[…] https://trac.gajim.org/ticket/8329
  59. bot RSS: Feeds for Gajim • Ticket #8330 (History Manager fault) created Bug description When I just open and close history manager (via Close cross in the upper right corner), doing nothing inside, the fault always happens: Microsoft Visual C++ Runtime Library Runtime Error! Program: C:\Program Files (x86)\Gajim\bin\history_manager.exe The application has requested the Runtime to terminate it in an unusual way. Please contact the applicatio[…] https://trac.gajim.org/ticket/8330
  60. Asterix nice tmolitor for the bug fixed. but what about this one: diff -r 54b0608a5886 src/common/connection_handlers_events.py --- a/src/common/connection_handlers_events.py Tue Apr 12 22:53:41 2016 +0200 +++ b/src/common/connection_handlers_events.py Mon Apr 18 21:49:49 2016 +0200 @@ -1032,10 +1032,13 @@ delay = self.stanza.getTag('delay', namespace=nbxmpp.NS_DELAY2) if not delay: return + self.msg_ = self.stanza.getTag('message') + delay2 = self.msg_.getTag('delay', namespace=nbxmpp.NS_DELAY2) + if delay2: + delay = delay2 tim = delay.getAttr('stamp') tim = helpers.datetime_tuple(tim) self.tim = localtime(timegm(tim)) - self.msg_ = self.stanza.getTag('message') to_ = self.msg_.getAttr('to') if to_: to_ = gajim.get_jid_without_resource(to_)
  61. Asterix at least this is done only once: tim = delay.getAttr('stamp') tim = helpers.datetime_tuple(tim) self.tim = localtime(timegm(tim))
  62. bot RSS: Feeds for Gajim • Ticket #8329 (Pressing Clear Devices for OMEMO cause this fault) closed invalid: please fill it in ​trac-plugins bug tracker https://trac.gajim.org/ticket/8329#comment:1
  63. bot RSS: Feeds for Gajim • Changeset [16020:00fb9f8cfb30]: call gpg with LANG=C to prevent decoding issues call gpg with LANG=C to prevent decoding issues https://trac.gajim.org/changeset/00fb9f8cfb307129b525ddecfc5ac20ef67d569b
  64. Link Mauve Speaking about gpg, I recently implemented the new OX XEP in slixmpp, and pygpgme is much better than gnupg.py.
  65. Asterix good to hear, if one day I don't know what to do :) But gnupg works not that bad ...
  66. Link Mauve The latest version was horribly broken here on gpg 2.1.11.
  67. Link Mauve For example, if I clicked cancel in the pinentry window, it would internally traceback on a thread but return success with some random bytes.
  68. Link Mauve And the non-latest commit didn’t have support for many changes having happened since gnupg 1.4.
  69. bot RSS: Feeds for Gajim • Changeset [16027:e89d125d0505]: call gpg with LANG=C to prevent[] • Ticket #8315 (FT error) updated[] • Ticket #8314 (Gajim[…] https://trac.gajim.org/ticket/8314#comment:3 • Ticket #7978 (IndexError: list [] • Ticket #3939 (Clic[…] https://trac.gajim.org/ticket/3939#comment:11 • [] • Ticket #8314 (Gajim window size[] • Ticket #7978 (IndexError: list []
  70. Asterix I don't understand your last sentence
  71. Link Mauve The gnupg interface changed a lot between 1.4 and 2.0, so you need the very latest revision of gnupg.py to fix most tracebacks.
  72. Link Mauve gpgme would have the same issue if it wasn’t developped by the same people as gnupg.
  73. Asterix yes I'll import them ASAP