Gajim - 2016-12-11


  1. lovetox_ arune
  2. lovetox_ your dream comes true, i just made a windows vault password storage work :D
  3. Link Mauve lovetox_, wouldn’t it be a better idea to implement a libsecret provider service?
  4. Link Mauve That way every single GLib program would be able to use it.
  5. lovetox_ seems complicated :)
  6. lovetox_ i dont even know what you are talking about :D
  7. Link Mauve I’m pretty sure it’d be simpler than to add that to every present and future program.
  8. lovetox_ libsecret doesnt support windows
  9. Link Mauve lovetox_, it doesn’t have to, Windows or a service have to support it.
  10. lovetox_ so what do i have to do?
  11. lovetox_ python keyring makes it pretty simple to store passwords in windows
  12. lovetox_ like 30 lines of code
  13. Link Mauve I would say to write a program in any language you want, which uses said Windows API and exposes the libsecret path as a dbus service.
  14. lovetox_ thats a bit overkill :)
  15. Link Mauve Why?
  16. Link Mauve That’s how every other platform does it.
  17. lovetox_ because it involves writing another programm, and using dbus on windows
  18. lovetox_ all that only so libsecret doesnt have to do the same that keyring already does?!
  19. Link Mauve I don’t understand what you just said.
  20. Link Mauve libsecret doesn’t do anything by itself, it’s just an API to access encrypted storage on different platforms.
  21. lovetox_ but why should someone do this?
  22. lovetox_ only that you can have one libsecret for every platform?
  23. lovetox_ why using dbus? does this have any value on windows?
  24. Link Mauve Err, yes, so that you don’t need to add any platform-specific code to every single program.
  25. lovetox_ but we would have to ship a additional programm?
  26. lovetox_ but we can say we have no windows code in our password.py in gajim
  27. Link Mauve That would just be launched automatically the first time you try to access a password.
  28. lovetox_ i take the 30 lines of platfrom dependend code
  29. lovetox_ something different
  30. lovetox_ can you maybe tell me why would one inherit a class if all methods get overwritten ?
  31. lovetox_ https://dev.gajim.org/gajim/gajim/blob/master/src/common/passwords.py
  32. lovetox_ like here
  33. lovetox_ the PasswordStorage class
  34. lovetox_ what gain do we have inheriting it?
  35. Link Mauve lovetox_, that way you can check isinstance(this_object, PasswordStorage) and know both of those method are implemented.
  36. lovetox_ ok thanks
  37. Link Mauve And because Python doesn’t have any concept of pure virtual method.
  38. arune lovetox_: wow! Wonderful!
  39. linus Link Mauve: but that goes against Python's conventions, duck typing is the way!
  40. linus Is nbxmpp not on gitlab yet?
  41. linus Nvm I just can't navigate gitlab properly
  42. linus hm. When I have OMEMO active on a chat and then disable the plugin, the chat is still displayed as secure even though it's not
  43. lovetox hm yeah, forgot to set this icon on deactivate
  44. lovetox thanks for reporting :)
  45. linus Suggestion: have the OMEMO button show up in every chat even if it's unavailable (greyed out, with a tooltip indicating why). Makes it less confusing as to why it might not be there.
  46. linus Also, maybe hg.gajim.org should redirect to dev.gajim.org now?
  47. linus kalkin, do you maintain the github mirror? The link is out of date and still refers to mercurial https://github.com/gajim/gajim
  48. linus Also, in longer-term things: I don't think I mentioned this before, but I think it would be very nice to avoid importing plugin modules if we're not actually enabling them, for performance and safety. In the current state, a broken plugin can break gajim even if it isn't activated.
  49. lovetox all good points
  50. lovetox definitly we should show the omemo icon as we do already in groupchats
  51. lovetox with the plugins i think its a bit harder
  52. lovetox because it would need to change completly the way we identify plugins
  53. lovetox but i think it also would be a good improvement
  54. lovetox the github mirror is maintained by Johannes i think
  55. linus not really, we've got the beginnings of it with the manifests
  56. arune lovetox: wow! Wonderful!
  57. lovetox then go linus :D
  58. linus and for the github mirror, kalkin is (the only person to be) listed as a member of the github org which is why I asked him
  59. kalkin linus: no i dont maintain it. Need to delete it and just leave a readme
  60. linus ah ok
  61. kalkin Also the omemo org needs to be removed now i think
  62. kalkin lovetox: can i close the issur tracker @ github?
  63. lovetox hm
  64. lovetox valid issues
  65. lovetox is there a way to import them to gitlab?
  66. lovetox i dont think so, should i open issues there and just copy the text?
  67. lovetox i think i will do this, i tell you when im finished
  68. linus Surely there's a script for it floating about the web
  69. lovetox yeah to import whole projects
  70. lovetox but only issues, into a already exsiting project
  71. lovetox hm i will ask asterix :)
  72. linus lovetox: https://gitlab.com/gitlab-org/gitlab-recipes/blob/master/import/github/import_all.rb
  73. lovetox k thanks, i will tell asterix to try to import it
  74. linus wait, it's only one issue?
  75. linus Then yeah importing it manually makes more sense I think
  76. lovetox its 4 open issues
  77. lovetox and 97 closed
  78. lovetox does it make sense to import closed ones..
  79. lovetox i dont know
  80. linus I think we're looking at different places
  81. lovetox https://github.com/omemo/gajim-omemo/issues
  82. lovetox i was talking about omemo
  83. linus oh right
  84. lovetox not the github repo for gajim
  85. lovetox :D
  86. kalkin Is it possible to use some thirdparty credentials like github/Facebook/whatever to use gajim gitlab?
  87. lovetox yes
  88. lovetox github
  89. lovetox there are icons under the login to show you whats possible
  90. lovetox google is also possible
  91. lovetox if you wish for some specific other provider just tell asterix
  92. linus lovetox, I'm guessing I should fork the repo to contribute changes and not push to new branches in the existing repo?
  93. lovetox correc
  94. lovetox t
  95. lovetox it wouldnt be possible anyway to push something to the main repo
  96. linus why not?
  97. lovetox you have dev status, you can make merge requests
  98. linus I mean, I see it isn't allowed, but it would certainly be possible
  99. lovetox and also accept your own merge requests
  100. lovetox i meant it is not allowed
  101. lovetox not that it wouldnt be possible
  102. lovetox the problem is gitlabs permissions system
  103. lovetox there is only do everything
  104. lovetox or nothing
  105. lovetox so we protected the main repo
  106. lovetox that no one can push to it directly
  107. lovetox but devs can create and accept their own merge requests
  108. linus Well you could protect the master, gtk3 and gajim_0.15 branches and let people push to other, new ones
  109. linus Ah well, forking works just as well
  110. lovetox i like it that way, i mean i dont want to think about every push because its final
  111. linus What's our lowest supported version of gtk?
  112. linus For the gtk3 version
  113. linus Seems to be 3.12, at least based on the ui files in the main gajim repo
  114. lovetox yeah i think we dont need to go lower than 3.12
  115. lovetox thats pretty old, people can use 0.16 if they have such a old system
  116. linus Why does url_image_preview try to use PIL anyway, when we have GdkPixbuf?
  117. lovetox hmm#
  118. Link Mauve IIRC it was because someone wanted to use some obscure format that wasn’t yet supported by GdkPixbuf.
  119. Link Mauve But it’s a bad idea imo.
  120. linus +1
  121. Link Mauve Contributing upstream is almost always a better solution.
  122. lovetox ah yes
  123. lovetox pixbuf doesnt support GIF i think
  124. lovetox its not obscure
  125. lovetox :)
  126. Link Mauve It most certainly does.
  127. lovetox not with resizing
  128. lovetox or something like that
  129. lovetox because we have to resize and show thumbnails or smaller pictures
  130. linus That sounds like nonsense to me
  131. linus (no offence)
  132. lovetox i cant remember but it was something like that
  133. linus and lovetox's commit adding PIL doesn't mention it at all in the commit message https://dev.gajim.org/gajim/gajim-plugins/commit/c287fdcb79779eb2771513d71bdf9bf5ab4f8b33
  134. Link Mauve So many unrelated changes in a single commit…
  135. lovetox haha that was one of my first commits
  136. lovetox it changes mostly formatting
  137. lovetox or much of it is
  138. lovetox that this is bad to read afterwards i learned later :)
  139. lovetox hm so use pixbuf then, i cant really remember anymore, but i know there was a reason to keep it, i think we will see :)
  140. lovetox was it maybe because of bmp?
  141. linus why on earth would you ever want to use bmp as an image transfer format?
  142. Link Mauve linus, AFAIK this is receiving, you can’t control what the sender will send you using this terrible transfer method.
  143. linus of course, but you could just ignore it :p
  144. Link Mauve Btw, Tobias has been working on a better way to advertise such file transfers in a conversation, which I quite like.
  145. Link Mauve See here for a preview: http://ayena.de/files/xeps/xep-async-file-sharing.xhtml
  146. lovetox yeah great :)
  147. lovetox thats the plan, image previewer should in the end really only preview random internet links if the user wishes that, and the rest like httpupload filetransfers should move into gajim core
  148. lovetox but it would be good if there would be a xep like that that tells us how and what data to send with such a async filetransfer
  149. arune lovetox: wasn't the problem reading webp from Conversations?
  150. lovetox hm i dont think so
  151. lovetox i think it has something to do with scaling, gtk 2 pixbuf could load and display all kinds of images, but scaling was another issue