Gajim - 2014-10-17


  1. Greg Hello everyone, I have a general question that hopefully someone can answer.
  2. Greg I'm packaging Gajim for a distro and I've built the python-nbxmpp dep and also python-openssl which was optional. When I run Gajim I can't connect to Jabber servers. When I run the program in verbose mode I see reference to libasyncns-python, which I've also downloaded and tried building, however libasyncns-python relies on libasyncns to build, since it's only the python bindings. I can't for the life of me find where libasyncns is currently hosted. I do see a lot of old references to it from years ago, which makes me believe it was simply no longer maintained. However I think this odd since Gajim just had a release and is still using that python module.
  3. Link Mauve Greg, do you have another error?
  4. Link Mauve I don’t have that library on my system, and Gajim connects without issue.
  5. Greg Yes, there are some read timeout errors for DNS, which made me believe that the libasyncns error held more weight.
  6. Greg Let me get you the full error I have
  7. Greg $ gajim -v Logger gajim level set to 1 Logger nbxmpp level set to 1 ** (gajim:1237): WARNING **: Trying to register gtype 'GMountMountFlags' as enum when in fact it is of type 'GFlags' ** (gajim:1237): WARNING **: Trying to register gtype 'GDriveStartFlags' as enum when in fact it is of type 'GFlags' ** (gajim:1237): WARNING **: Trying to register gtype 'GSocketMsgFlags' as enum when in fact it is of type 'GFlags' 10/16/2014 15:27:53 (E) gajim.gtkgui_helpers Unable to load icon gtk-spell-check: Icon 'gtk-spell-check' not present in theme 10/16/2014 15:27:54 (I) gajim.gajim Encodings: d:utf-8, fs:UTF-8, p:UTF-8 10/16/2014 15:27:54 (D) gajim.c.resolver Import of libasyncns-python failed, getaddrinfo will block Traceback (most recent call last): File "/usr/share/gajim/src/common/resolver.py", line 44, in <module> import libasyncns File "/usr/share/gajim/src/common/demandimport.py", line 95, in _demandimport return _import(name, globals, locals, fromlist, level) ImportError: No module named libasyncns 10/16/2014 15:27:54 (I) gajim.c.resolver Using NSLookupResolver /usr/lib/python2.7/site-packages/gobject/__init__.py:115: Warning: Attempt to add property cell_renderer_image+CellRendererImage::image after class was initialised type_register(cls, namespace.get('__gtype_name__')) /usr/share/gajim/src/roster_window.py:5145: Warning: g_object_set_qdata: assertion 'G_IS_OBJECT (object)' failed if not model[titer][data]: /usr/share/gajim/src/roster_window.py:5208: Warning: g_object_set_qdata: assertion 'G_IS_OBJECT (object)' failed if type_ == 'account' and model[titer][C_PADLOCK_PIXBUF]: /usr/share/gajim/src/roster_window.py:6540: Warning: g_object_set_qdata: assertion 'G_IS_OBJECT (object)' failed self.window.show_all() 10/16/2014 15:27:55 (D) gajim.plugin_system Attributes processing started 10/16/2014 15:27:55 (D) gajim.plugin_system PluginInstaller : <class 'plugin_installer.plugin_installer.PluginInstaller'> 10/16/2014 15:27:55 (D) gajim.plugin_system is subclass of GajimPlugin 10/16/2014 15:27:55 (D) gajim.plugin_system plugin_installer : <module 'plugin_installer.plugin_installer' from '/usr/share/gajim/plugins/plugin_installer/plugin_installer.py'> 10/16/2014 15:27:55 (D) gajim.plugin_system GajimPlugin.__init__() <entered> 10/16/2014 15:27:55 (D) gajim.plugin_system GajimPlugin.load_config() <entered> 10/16/2014 15:27:55 (D) gajim.plugin_system GajimPlugin.load_config() <left> 10/16/2014 15:27:55 (D) gajim.plugin_system PluginInstallerPlugin.init() <entered> 10/16/2014 15:27:55 (D) gajim.plugin_system GajimPlugin.local_file_path() <entered> 10/16/2014 15:27:55 (D) gajim.plugin_system GajimPlugin.local_file_path() <left> 10/16/2014 15:27:55 (D) gajim.plugin_system PluginInstallerPlugin.init() <left> 10/16/2014 15:27:55 (D) gajim.plugin_system GajimPlugin.__init__() <left> 10/16/2014 15:27:55 (D) gajim.plugin_system PluginInstallerPlugin.activate() <entered> 10/16/2014 15:27:55 (D) gajim.plugin_system PluginInstallerPlugin.activate() <left> 10/16/2014 15:27:55 (D) gajim.c.ged roster-received Args: (<common.connection_handlers_events.RosterReceivedEvent object at 0x7f7ddf6cefd0>,) 10/16/2014 15:27:55 (D) gajim.c.ged before-change-show Args: (<common.connection_handlers_events.BeforeChangeShowEvent object at 0x7f7dd31d43d0>,) 10/16/2014 15:27:55 (D) gajim.c.resolver resolve _xmpp-client._tcp.gryman.com type=srv 10/16/2014 15:27:55 (D) gajim.c.resolver Starting to resolve _xmpp-client._tcp.gryman.com using <common.resolver.NSLookupResolver instance at 0x7f7dd32f1cf8> 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout set for fd 11 on 20 seconds sh: 1: nslookup: not found 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout removed for fd 11 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout removed for fd 11 10/16/2014 15:27:55 (D) gajim.c.resolver Resolving result for _xmpp-client._tcp.gryman.com: [] 10/16/2014 15:27:55 (D) gajim.c.connection Connection to next host 10/16/2014 15:27:55 (I) gajim.c.connection >>>>>> Connecting to gryman.com [gryman.com:5222], type = tls 10/16/2014 15:27:55 (I) nbxmpp.plugin Plugging <nbxmpp.transports_nb.NonBlockingTCP instance at 0x7f7de2e9d488> __INTO__ <nbxmpp.client_nb.NonBlockingClient instance at 0x7f7de2e9d440> 10/16/2014 15:27:55 (I) nbxmpp.transports_nb NonBlockingTCP Connect :: About to connect to 107.170.229.244:5222 10/16/2014 15:27:55 (I) nbxmpp.transports_nb Plugging fd 11, W:True, R:False 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout set for fd 11 on 30 seconds 10/16/2014 15:27:55 (I) nbxmpp.transports_nb After NB connect() of 140178654614664. "Operation now in progress" raised => CONNECTING 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout removed for fd 11 10/16/2014 15:27:55 (I) nbxmpp.transports_nb pollend called, state == CONNECTING 10/16/2014 15:27:55 (I) nbxmpp.transports_nb Error while disconnecting socket: Transport endpoint is not connected 10/16/2014 15:27:55 (D) nbxmpp.client_nb While looping over DNS A records: Error during connect to 107.170.229.244:5222 10/16/2014 15:27:55 (I) nbxmpp.client_nb Disconnecting NBClient: Run out of hosts for name gryman.com:5222. Error for last IP: Error during connect to 107.170.229.244:5222 10/16/2014 15:27:55 (I) nbxmpp.plugin Plugging <nbxmpp.transports_nb.NonBlockingTCP instance at 0x7f7de2e9d488> __OUT__ of <nbxmpp.client_nb.NonBlockingClient instance at 0x7f7de2e9d440>. 10/16/2014 15:27:55 (D) nbxmpp.client_nb Client disconnected.. 10/16/2014 15:27:55 (D) nbxmpp.client_nb calling on_connect_failure cb 10/16/2014 15:27:55 (I) gajim.c.connection >>>>>> Connecting to gryman.com [gryman.com:5223], type = ssl 10/16/2014 15:27:55 (I) nbxmpp.plugin Plugging <nbxmpp.transports_nb.NonBlockingTCP instance at 0x7f7de2e9dd88> __INTO__ <nbxmpp.client_nb.NonBlockingClient instance at 0x7f7dc71d5878> 10/16/2014 15:27:55 (I) nbxmpp.transports_nb NonBlockingTCP Connect :: About to connect to 107.170.229.244:5223 10/16/2014 15:27:55 (I) nbxmpp.transports_nb Plugging fd 13, W:True, R:False 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout set for fd 13 on 30 seconds 10/16/2014 15:27:55 (I) nbxmpp.transports_nb After NB connect() of 140178654616968. "Operation now in progress" raised => CONNECTING 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout removed for fd 13 10/16/2014 15:27:55 (I) nbxmpp.transports_nb pollend called, state == CONNECTING 10/16/2014 15:27:55 (I) nbxmpp.transports_nb Error while disconnecting socket: Transport endpoint is not connected 10/16/2014 15:27:55 (D) nbxmpp.client_nb While looping over DNS A records: Error during connect to 107.170.229.244:5223 10/16/2014 15:27:55 (I) nbxmpp.client_nb Disconnecting NBClient: Run out of hosts for name gryman.com:5223. Error for last IP: Error during connect to 107.170.229.244:5223 10/16/2014 15:27:55 (I) nbxmpp.plugin Plugging <nbxmpp.transports_nb.NonBlockingTCP instance at 0x7f7de2e9dd88> __OUT__ of <nbxmpp.client_nb.NonBlockingClient instance at 0x7f7dc71d5878>. 10/16/2014 15:27:55 (D) nbxmpp.client_nb Client disconnected.. 10/16/2014 15:27:55 (D) nbxmpp.client_nb calling on_connect_failure cb 10/16/2014 15:27:55 (I) gajim.c.connection >>>>>> Connecting to gryman.com [gryman.com:5222], type = plain 10/16/2014 15:27:55 (I) nbxmpp.plugin Plugging <nbxmpp.transports_nb.NonBlockingTCP instance at 0x7f7de2e9def0> __INTO__ <nbxmpp.client_nb.NonBlockingClient instance at 0x7f7de2e9df80> 10/16/2014 15:27:55 (I) nbxmpp.transports_nb NonBlockingTCP Connect :: About to connect to 107.170.229.244:5222 10/16/2014 15:27:55 (I) nbxmpp.transports_nb Plugging fd 14, W:True, R:False 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout set for fd 14 on 30 seconds 10/16/2014 15:27:55 (I) nbxmpp.transports_nb After NB connect() of 140178654617328. "Operation now in progress" raised => CONNECTING 10/16/2014 15:27:55 (I) nbxmpp.idlequeue read timeout removed for fd 14 10/16/2014 15:27:55 (I) nbxmpp.transports_nb pollend called, state == CONNECTING 10/16/2014 15:27:55 (I) nbxmpp.transports_nb Error while disconnecting socket: Transport endpoint is not connected 10/16/2014 15:27:55 (D) nbxmpp.client_nb While looping over DNS A records: Error during connect to 107.170.229.244:5222 10/16/2014 15:27:55 (I) nbxmpp.client_nb Disconnecting NBClient: Run out of hosts for name gryman.com:5222. Error for last IP: Error during connect to 107.170.229.244:5222 10/16/2014 15:27:55 (I) nbxmpp.plugin Plugging <nbxmpp.transports_nb.NonBlockingTCP instance at 0x7f7de2e9def0> __OUT__ of <nbxmpp.client_nb.NonBlockingClient instance at 0x7f7de2e9df80>. 10/16/2014 15:27:55 (D) nbxmpp.client_nb Client disconnected.. 10/16/2014 15:27:55 (D) nbxmpp.client_nb calling on_connect_failure cb 10/16/2014 15:27:55 (D) gajim.c.connection Connection to next host 10/16/2014 15:27:55 (D) gajim.c.connection Out of hosts, giving up connecting to gryman.com 10/16/2014 15:27:55 (D) gajim.c.ged our-show Args: (<common.connection_handlers_events.OurShowEvent object at 0x7f7dd323cb50>,) 10/16/2014 15:27:55 (D) gajim.c.ged connection-lost Args: (<common.connection_handlers_events.ConnectionLostEvent object at 0x7f7dd320ec90>,) gajim.py:538: Warning: g_object_set_qdata: assertion 'G_IS_OBJECT (object)' failed gtk.main() ^CKeyboardInterrupt
  8. Link Mauve 10/16/2014 15:27:55 (I) nbxmpp.client_nb Disconnecting NBClient: Run out of hosts for name gryman.com:5222. Error for last IP: Error during connect to 107.170.229.244:5222
  9. Link Mauve Trying 107.170.229.244... telnet: Unable to connect to remote host: Connection refused
  10. Link Mauve When I telnet to it.
  11. Greg I'm connected to it now via Pidgin
  12. Greg Let me get on that server
  13. Greg Oh, I think I see what happened
  14. Greg It wasn't able to resolve the _xmpp-client._tcp.gryman.com srv record
  15. Greg So it just tried to make the connection to gryman.com
  16. Greg Which doesn't host the jabber service
  17. Link Mauve But there is no such SRV recorg.
  18. Link Mauve But there is no such SRV record.
  19. Link Mauve At least according to dig.
  20. Link Mauve Oh, there is.
  21. Link Mauve I just made a stupid typo.
  22. Link Mauve Yes, mail.gryman.com. works fine.
  23. Link Mauve Oh, so I *do* have libasyncns, which comes from http://0pointer.de/lennart/projects/libasyncns
  24. Greg That site is down
  25. Link Mauve Thanks Lennart.
  26. Greg hah
  27. Greg I emailed him, but no response yet
  28. Greg From the logs it looks like as a fallback it tried to use nslookup
  29. Greg The Jabber server I run has been up for a while now and it works just fine, including the DNS entries. So that's not the issue. I think the libasyncns-python is the issue from what I see anyway.
  30. Greg It also seems like that python module should be listed as an optional dependency
  31. Greg Or maybe not so optional in my case. :)
  32. Link Mauve Indeed. ^^
  33. Link Mauve Which distribution are you using?
  34. Link Mauve I see it’s missing on ArchLinux as well.
  35. Greg Evolve OS
  36. Greg It's very alpha, which is why I'm packaging. :)
  37. Greg I prefer Gajim over Pidgin, personally. ;)
  38. Link Mauve :)
  39. Greg Do you happen to know if Arch is patching Gajim to get around not having libasyncns?
  40. Greg Looks like their dep list has ldns
  41. Greg Ok, so they did patch Gajim to use drill instead of nslookup
  42. Greg sigh...
  43. Link Mauve libasyncns is packaged in ArchLinux, just not set as a dependency for Gajim.
  44. Link Mauve Anyway, 4am is too late for me, see you Greg. \o_
  45. Greg Alright, bye
  46. Greg Thanks for the help
  47. spicewiesel Is there any option to join bookmarked group chats minimized? I cannot find it...
  48. Link Mauve spicewiesel, in the bookmarks manager, there is a checkbox per-room called “Start minimized”.
  49. mcepl any comments on https://trac.gajim.org/ticket/7855
  50. mcepl ?
  51. spicewiesel Link Mauve, ahhhh how could I be that blind. Thanks!
  52. Link Mauve mcepl, lgtm.
  53. Link Mauve spicewiesel, I think a global option could be useful as well.
  54. mcepl so how does the process of actually merging patches work? Does Asterix have to approve it?
  55. Link Mauve Having to set that option once per room is no fun when you have a lot of rooms.
  56. spicewiesel yes, I think that would be nice. Would be ok as a changable default in the pref editor
  57. Link Mauve mcepl, yes, Asterix or someone else with push rights to the repository (dicson?).
  58. Link Mauve spicewiesel, maybe open a ticket about that.
  59. mcepl just has to say that gajim should switch IMHO to git ... hg sucks. Sorry, I have just to release it from my chest.
  60. Link Mauve mcepl, err, why?
  61. Link Mauve If you are using git-remote-hg just fine, why would the entire project change?
  62. mcepl pull requests, cheap branches, etc. ... anyway ... is there at least a blessed mirror on Bibbucket?
  63. Link Mauve I for one prefers hg a lot more compared to git.
  64. mcepl also, huge number of people using it ... you don'ŧ want to persuade your potential contributors to do something they don't want to do.
  65. mcepl anyway, I have lift it off my chest, and now I will keep shut up
  66. Link Mauve Pull requests are a platform feature, not a hg/git feature; what is not cheap in hg’s branches or bookmarks (AFAIK bookmarks are the equivalent of git branchs); and I have no idea about bitbucket, I never tried it myself.
  67. mcepl that's obvious you have never tried it
  68. Link Mauve I have tried git, btw.
  69. Link Mauve Many times, and I hate its cluttered UI.
  70. mcepl that's BS ... nobody forces you to use it.
  71. Link Mauve I know that, hence my usage of hg-git.
  72. mcepl I am happy to have git rev-parse, but you don't have to know about it
  73. mcepl anyway, enough of that
  74. Link Mauve I don’t get your point, sure no one is forcing me to use git, but then no one is forcing you to use hg either.
  75. Link Mauve Just use the one you like the most, and contribute as you want to the projects using either.
  76. mcepl it is obvious that projects still using non-git will rather die than switch (and perhaps they will). .... you at least don't use bzr ;)
  77. Link Mauve git doesn’t inherently use PRs, not bitbucket or github, those are projects choices.
  78. mcepl you do force me to use hg ... I have to deal with multiple heads and junk like that. Because this crap just happens.
  79. Link Mauve For example Mesa uses git and relies on a patches-on-a-ML workflow, which I find much better than some web platform.
  80. Link Mauve Err, just use bookmarks if you don’t want to have unnamed heads.
  81. Link Mauve Or git-remote-hg like you told me yesterday.
  82. mcepl well, theoretically I agree, but reality is that in the end for example for vimoutliner I use mostly github issue tracker and pull requests even though we have email list set up.
  83. mcepl yes, so I have to learn hg to be able to work with you
  84. bot RSS: Feeds for Gajim • Ticket #7855 (Use explicit set_use_underline on MenuItem) created There are many warnings Using positional arguments with the GObject constructor has been deprecated. cluttering stderr needlessly. https://trac.gajim.org/ticket/7855 • 0001-Use-explicit-set_use_underline-on-MenuItem.patch attached to Ticket #7855 Suggested patch https://trac.gajim.org/attachment/ticket/7855/0001-Use-explicit-set_use_underline-on-MenuItem.patch
  85. Link Mauve Oh, doesn’t git-remote-hg allows you to use git?
  86. mcepl poorly
  87. Link Mauve I thought it was some equivalent of git-hg, like you told me.
  88. mcepl multiple heads and crap like that happens
  89. Link Mauve Well then, just use your “semi-official git clone” on github, what is the issue with that?
  90. Link Mauve The workflow for Gajim is patches-on-trac, that would work with exactly any VCS, even none and plain diff -uN.
  91. mcepl yes, exactly ... so I see patches rotting in the trac for long time
  92. mcepl Added 8 years ago by anonymous .... https://trac.gajim.org/ticket/1564
  93. Link Mauve :D
  94. Link Mauve That’s only because nobody (i.e. not Asterix nor dicson) pushed that upstream.
  95. mcepl and I am not able to add keyword patch to my ticket, so this will CERTAINLY rot forever.
  96. mcepl "Changed 5 years ago by johnny" ... "thanks for the work, i didn't even know a patch was here. adding a few keywords, and updating the summary." That's five years ago.
  97. mcepl https://trac.gajim.org/ticket/3724
  98. mcepl (patch is seven years old)
  99. mcepl no, of course, problem is not in tools, but in humans ... but at least GitHub and pull request make it more obvious, and it is easier to make fork and pull all PRs to new repo.
  100. Link Mauve Well, you can open a ticket asking Asterix to move all of his infrastructure to github, if you really think it’s that helpful.
  101. mcepl as a potential contributor I go to look for the activity of the patch queue and when I see what I see in Gajim's trac, I would go away,
  102. mcepl and I know what he will answer me
  103. Link Mauve If you asked that to one of my projects though, I wouldn’t do it.
  104. mcepl as I said, projects who not use git now, will rather die than switch. And perhaps they will.
  105. Link Mauve ^^'
  106. Link Mauve I can’t wait for the huge wave of people contributing to Gajim because you made a clone on github!
  107. Link Mauve If it’s such a silver bullet, I’m sure thousands of people will contribute overnight!
  108. mcepl of course they won't, because nobody knows about it
  109. mcepl well, I have tried it wirh couple of projects I do maintain and the truth is that suddenly more cooperation (a little bit more) happened.
  110. spicewiesel mcepl, ".. I would go away," Same is with users. I checked some desktop clients as I was looking for a new one or the one with the features I need. And to be honest: It seemed all xmpp desktop clients are passed away.
  111. mcepl see Mozilla how they hardly maintain they hg infrastructure, because everybody wants to use git
  112. mcepl spicewiesell: which is the reason I am still here ... gajim is one of the best XMPP clients.
  113. mcepl but if I didn't know it already
  114. spicewiesel yep...
  115. Link Mauve mcepl, btw, your clone isn’t up to date, there have been a lot of commits since May.
  116. mcepl and especially because you don't have many contributors it is way better to make contributing difficult
  117. mcepl Link Mauve: which branch?
  118. spicewiesel I really like Psi+, too. But there were nearly 5 commits in the last 12 months, so... yeah... I do not hope to see xep-280 there in the near future...
  119. Link Mauve mcepl, both?
  120. mcepl oh of course, hg cannot delete old merged branches ... I knwo
  121. Link Mauve I have never found contributing patches difficult, I find Gajim’s source code much harder to get around.
  122. Link Mauve Once the patch is written, it’s trivial to send it on the trac.
  123. mcepl yes, seven years and eight years
  124. Link Mauve mcepl, you have only two branches on your repository, devel and master.
  125. mcepl no, I mean which hg branch is more recent ... I thought that default is the requiavalent of git master?
  126. spicewiesel (we migrate our internal stuff to git(lab) right now, as every colleague is ranting on trac for years now)
  127. mcepl of course, my devel is on the top of my master
  128. spicewiesel but I am no developer, so I cannot contribute own experiences .)
  129. Link Mauve mcepl, gajim_0.16 seems to be the most recent, default is the future 0.17 one.
  130. mcepl so that should be the most recent one, right?
  131. Link Mauve Yes.
  132. mcepl oh you cannot rebase, so you have mess in your branches, right
  133. Link Mauve I don’t see any mess in Gajim’s branches, but whatever.
  134. mcepl yes? so the development is spread somewhere between 0.16 branch and default, right?
  135. mcepl I told you which should stop this discussion long time ago and you still drag it on.
  136. Link Mauve Alright, let’s stop then.
  137. Link Mauve But if you need me I’m still here.
  138. mcepl well, I would need to explain why the development is going on stable branch (which I believe 0.16 is) and not on default.
  139. Link Mauve It wasn’t development AFAIK, merely stabilisation.
  140. Link Mauve Asterix really didn’t have time this past year, and so concentrated his efforts on that.
  141. mcepl butin the end one of the branches (default or 0.16) will be thrown away, right?
  142. mcepl (and yes, it doesn't matter whether you use hg or git)
  143. Link Mauve Why would it be thrown away?
  144. Link Mauve gajim_0.16 changes will be merged into default, and development will continue towards gajim_0.17.
  145. mcepl so, yes, you are developing on stable branch. Lovely.
  146. mcepl OK, let me switch and note that I have probably just wasted two days of work, because the development documentation forgot to tell me.
  147. Link Mauve “you” is solely Asterix.
  148. mcepl it doesn'ŧ matter ... it is confusing to anybody who might still be willing to contribute
  149. Link Mauve mcepl, I doubt it, gtk3 is that different from gtk2 your patches are probably needed anyway.
  150. mcepl and yes of course, if he don't want contrbutors, let him just say it.
  151. Link Mauve If you have criticisms about his ways, please tell him, in a ticket, by email or otherwise.
  152. mcepl yes, but all my changes will be thrown away, because I doubt you will rebase, and I won't have all fixes which are on stable branch.
  153. mcepl perhaps
  154. Link Mauve Why rebase when you can merge?
  155. mcepl when I cool down a bit
  156. mcepl anyway,
  157. mcepl actually you know what? I have tried to make some sense of the HG repo (yes, I have installed hgview just beacause of gajim) and I cannot make head and tail of it. It is not about hg v. git ... this repo is such mess, that I have to go back to bitlbee + hexchat.
  158. Link Mauve mcepl, what do you find messy in there?
  159. Link Mauve Sorry, this is a genuine question, I never thought about the repository itself or the way it was managed.
  160. Holger mcepl: I prefer Git myself, but I totally fail to follow your "it will be lost 'cause you cannot rebase" stance. Usually you wouldn't rebase either of these public branches with Git either, and of course nothing gets lost by merging.
  161. Holger And yes, Git (especially GitHub) helps with attracting contributors these days, but I don't think this fact justifies bitching at maintainers who prefer to work with other tools.
  162. Link Mauve I’ve seen some communities getting completely destroyed from a (hostile) switch to github/google groups.
  163. Link Mauve Pyjamas, for example.
  164. Holger "hostile" sounds bad.
  165. Link Mauve The main contributor stopped working on it due to that, and then the project died.
  166. Holger Wonderful.
  167. Link Mauve Well, maybe it got a few more contributions from github.
  168. Holger All other things being equal you usually will, these days. But when the maintainer has a preference, things aren't equal.
  169. mcepl actually you know what? I have tried to make some sense of the HG repo (yes, I have installed hgview just beacause of gajim) and I cannot make head and tail of it. It is not about hg v. git ... this repo is such mess, that I have to go back to bitlbee + hexchat.
  170. Link Mauve I’ve started closing feature branches that got merged long ago btw.
  171. Link Mauve It will make hg heads a bit shorter.
  172. Link Mauve I’ll wait for Asterix to see if we should also close older release branches.
  173. mcepl Holger: so please tell me what should I do. Hg repo has 43 active heads and 17 named branches (are one subset of the other?).
  174. mcepl One would think that default is the right place to develop, but LinkMauve tells that it iactually gajim_0.16 (that's catchy name)? Or perhaps gtk3 branch?
  175. mcepl who knows?
  176. mcepl this is absolutely not about technology (although I would really prefer git), but about the overall disorganization of whole project.
  177. mcepl only information (labelled "General info about current Gajim hg branches", notice the word "current") https://trac.gajim.org/wiki/GajimHgBranches has been last updated 3 years ago.
  178. mcepl Link Mauve: and BTW, I don't know much about pyjamas but they fallen apart from the internal civil wars more than from any technology, didn't they?
  179. Link Mauve mcepl, gtk3 branch will be closed right now, it got merged into default some months ago.
  180. Link Mauve mcepl, internal civil wars due to disagreements that self-hosted was better than centralized platforms.
  181. mcepl Link Mauve: I don'ŧ understand ... internal civil wars?
  182. mcepl and why is still open gajim_0.6, 0.8, 0.8.2, 0.10, etc.? How many versions you really really support (you meaning "gajim porject")?
  183. Link Mauve I doubt those are still supported, but I will wait for Asterix’ opinion about what to do about them.
  184. Link Mauve In other news, I closed all the feature branches that got merged to trunk at some point of the history.
  185. Link Mauve http://hg.linkmauve.fr/gajim
  186. Link Mauve There are still a few unmerged branches, and those version branches.
  187. Link Mauve unmerged feature branches*
  188. Link Mauve So 21 remaining ones.
  189. mcepl so there is still one more questoin ... what's the difference between default and 0.16 and what it is there.
  190. mcepl stable version is meant to me (as a Red Hat employee working on RHEL among other things) something which should get ONLY bugfixes. Development on the stable branch seems to me a synonymum for bad engineering.
  191. Link Mauve default is the future 0.17, with lots of changes wrt libraries, including python3, gtk3, gst 1.x, gobject-introspection, and a few more probably; 0.16 is the latest released version.
  192. Link Mauve No, 0.16 got released a few days ago, before that it wasn’t considered stable yet.
  193. mcepl and all changes on 0.16 branches relevant to default have been obviously cherry-picked to default, right?
  194. Link Mauve No, because nobody did that already.
  195. mcepl OK, just that phrase "released before it was considered stable" ....
  196. Link Mauve I mean, the gajim_0.16 branch got created long before 0.16 got released, in order to stabilise it.
  197. Link Mauve And nobody took the time to merge it into default because 0.16 got released less than one week ago.
  198. mcepl OK, so it is a bloody mess ... I ignore everything else than default and if anyboy eventually fixes it, good for me.
  199. Link Mauve How else would you do it?
  200. bot RSS: Feeds for Gajim • Ticket #7856 (group chats - default option for "minimized on autojoin") created Hi there, bookmarked groupchats are by default not "minimized on autojoin". this could be annoying if you have many chats. It would be fine to be able to set a default value for it, maybe in the advanced configuration editor. Thanks! https://trac.gajim.org/ticket/7856
  201. mcepl Link Mauve‎: do you agree that these branches are merged and could be removed (closed, whatever is hg terminology)?
  202. Link Mauve The 0.6 and so branches? Sure.
  203. mcepl no, it is more ... look at the list
  204. mcepl http://paste.fedoraproject.org/142831/13553013/
  205. mcepl (sorry, I forgot to add URL)
  206. Link Mauve Pull my repository, I closed most of those.
  207. Link Mauve Only the ones which never got merged, I didn’t close.
  208. Link Mauve Actually, after removing the obsolete release branches, it went down to 8 branches.
  209. Link Mauve There, down to three open branches.
  210. Link Mauve gajim_0.16, default, and splash_screen.
  211. Link Mauve I’m not sure the last one is still wanted though.
  212. Link Mauve It seems to be bad design to me.
  213. Link Mauve mcepl, there: % hg heads changeset: 15502:444afa661d55 branch: gajim_0.16 user: Yann Leboulanger <asterix@lagaule.org> date: Sun Oct 12 15:16:17 2014 +0200 summary: fix sending zeroconf messages changeset: 15467:097ac847b4b7 user: Yann Leboulanger <asterix@lagaule.org> date: Thu May 29 23:28:25 2014 +0200 summary: add a not to not translate an example string
  214. Link Mauve I also closed splash_screen.
  215. Link Mauve Thanks for bringing that up, I don’t know why no branch ever got closed, I guess because nobody bothered.
  216. mcepl I have still http://paste.fedoraproject.org/142835/53786141/ .... are all those plugins somewhere in a separate repo?
  217. Link Mauve Have you pulled right now?
  218. Link Mauve My repository I mean.
  219. mcepl no, I didn't ...
  220. mcepl I don't want to remove it completely blindly without knowing what I do
  221. Link Mauve Nothing got removed, you can still see closed branches by running hg heads --closed.
  222. Link Mauve I just pushed some closing commits.
  223. mcepl BTW, if you are into hatred for proprietary silos (and there is a lot to it) http://luther.ceplovi.cz/git/gajim.git/
  224. spicewiesel I have one user where I get the little red cross (msg maybe not received) after every message, but he got my messages. Could that be caused by a xep the other client does not provide?
  225. Link Mauve I think this is XEP-184, it should be enabled only when your contact advertise it in its disco#info.
  226. Link Mauve Is your contact connected from multiple clients?
  227. spicewiesel is there any way to check this manually?
  228. spicewiesel nope, only psi+
  229. spicewiesel (means: I only see this ressource, usually I should see all imho)
  230. Link Mauve Hmm, no graphical way in Gajim, AFAIK.
  231. spicewiesel so, I think it's only this one
  232. spicewiesel okay
  233. Link Mauve You could send a manual disco#info, are you familiar with XMPP’s protocol?
  234. spicewiesel not that much :(
  235. Link Mauve Ok, so open the XML console (Actions > Advanced > XML Console).
  236. Link Mauve Then, send this to him, replacing the JID with his own: <iq type='get' to='your-contact@server.tld' id='info'> <query xmlns='http://jabber.org/protocol/disco#info'/> </iq>
  237. Link Mauve If there is <feature var='urn:xmpp:receipts'/> in the received stanza, then he supports it.
  238. spicewiesel ok
  239. Link Mauve Oh, your-contact@server.tld/resource, don’t forget the resource.
  240. spicewiesel there it is: <feature var='urn:xmpp:receipts'/>
  241. spicewiesel so, in that cause it's weird that I see all that messages?
  242. Link Mauve So his client properly supports message receipts, I guess it’s just a bug on his end then, which prevents his client from sending the receipts.
  243. Link Mauve Maybe ask him to restart his client or something.
  244. Link Mauve I don’t know Psi+ at all, sorry.
  245. spicewiesel okay, I asked him to restart, we will see :)
  246. Holger Either way it would probably be sane to never ever show a red cross but only the green OK icon ...
  247. Holger As some other clients do it.
  248. spicewiesel yep
  249. mcepl Holger: that seems like a bad idea ... it is supposed to be a security function is n't? So, if we don't get the receipt we should make it obvious that our message got lost.
  250. mcepl If it is broken, we should fix it (psi is FLOSS, so they should fix it as well).
  251. Holger mcepl: It is not supposed to be a "security function". Or whatever it is supposed to be, it doesn't work that way in practice. If you get an acknowledgement, you know the message was received. If you don't get one, you know nothing, and hence you shouldn't draw a red cross.
  252. Holger mcepl: And this has nothing to do with broken clients. Another client might have received it via Carbon Copies and whatnot.
  253. Holger mcepl: Gajim should draw a red cross or whatever if it receives an error bounce.
  254. mcepl OK, so what's broken? gajim, psi, or XEP? we should fix whatever is broken, not avoid it.
  255. spicewiesel fix all TEH things! :)
  256. spicewiesel I think that if the other clients tells that it is supporting receipts it's ok to show a red cross if no receipt was received from that ressource, not?
  257. Holger mcepl: Either the XEP or nothing. I'm not sure whether you want to just not use it, or whether it might have some value to be able to draw green OKs in some cases.
  258. Holger spicewiesel: No, because other resources might have received it, and users don't understand that.
  259. spicewiesel ok
  260. mcepl well, it could be optional, but I can imagine people who would be really interested in knowing for sure that the messge went through (remember, they use XMPP on Wall Street)
  261. Holger spicewiesel: This happens every day in practice for me.
  262. spicewiesel Holger, indeed
  263. arune spicewiesel: issue 7856, just what I was about to report some weeks ago!
  264. arune Thanks!
  265. spicewiesel I think that's exactely the reason why inputmice refused all that receipt stuff with conversations. It just tells you nothing and in some cases the wrong thing.
  266. Holger spicewiesel: Though you can now enable the green OKs in the expert-setting-please-be-careful area.
  267. Holger He's getting old :-)
  268. spicewiesel hehe ;)
  269. spicewiesel If users see the red tick they get used to it, and if I use a client without that they think something went wrong if there is no tick... b
  270. Holger Yeah and once they learn that the red cross may or may not be correct, they learn to ignore that crap feature.
  271. Holger I think the green OK can be useful once users understood the semantics. But most don't, of course.
  272. arune When using openfire we did often get red cross in gajim, but never after switching to prosody
  273. Holger Heh ...
  274. arune (Messages was always delivered though)
  275. spicewiesel arune, ... :)
  276. Holger No matter whether the client is gone or not!
  277. Holger Prosody delivers.
  278. Holger :-)
  279. spicewiesel hehe
  280. spicewiesel "The Dude abides."
  281. spicewiesel after psi restart I got no red cross on the first message, but on the second, and no one for 3 and 4... reliable issus, I love em.