> On June 25, 2013, 8:17 a.m., Pali Rohár wrote:
> > Hello, nice work :-)
> > 
> > Some my comments:
> > * please make all coding style changes in separate git commit (it will be 
> > easier to review and read your patch)
> > * also if you can move crash during key generation to separate git commit
> > * and move also patch for unencrypted messages to separate git commit
> > * this patch dropping libotr 3.x support, but lot of linux distributions 
> > does not have libotr 4.x - can you update patch to support both versions? I 
> > do not want to see that linux distributions drop otr support in kopete 
> > because they do not have needed libraries...
> > 
> > For adding new strings, you need to contact kde-i18n-doc mailinglist and 
> > ask for it.
> 
> Michael Zanetti wrote:
>     Now read the description again. 
>     
>     > * please make all coding style changes in separate git commit (it will 
> be easier to review and read your patch)
>     
>     As the code was really badly formatted (no - we didn't have reviewboard 
> the time I wrote the plugin initially) and QtCreator automatically trims 
> spaces at the end of lines, I'm not going to manually add whitespaces at the 
> end of the lines to break that again.
>     
>     > * also if you can move crash during key generation to separate git 
> commit
>     I don't think it makes sense to split the fix of the crash into a 
> separate commit. Especially that code changes during the plugin porting 
> anyways.
>     
>     > * and move also patch for unencrypted messages to separate git commit
>     There is no patch for unecrypted messages. That was a change from libotr 
> 3.2 to libotr 4.0.
>     
>     > * this patch dropping libotr 3.x support,
>     No it is not. If you would have read at least the description you would 
> see that it is tested and works with Adium, which does only support OTR 
> protocol version v2. Heck, it even supports OTR protocol v1.
>     
>     > For adding new strings, you need to contact kde-i18n-doc mailinglist 
> and ask for it.
>     You mean to ask for an exception for getting it inot 4.11?
> 
> Pali Rohár wrote:
>     Spliting patch into more smaller (whitespace, crash fix, ...) is just 
> suggestion, now when we have git (but is it not required).
>     In your patch, you increased in cmake minimal libotr version to 4.0. So 
> you dropped support for libotr 3.2 library at compile time.
>     And yes, for adding new strings you need exception.
> 
> Michael Zanetti wrote:
>     > In your patch, you increased in cmake minimal libotr version to 4.0. So 
> you dropped support for libotr 3.2 library at compile time.
>     
>     Oh... you meant compile time... Yes. Its true. This requires 4.0.0 to 
> compile. However, I guess the reason for most distributions not shipping 
> libotr 4.0.0 yet is exactly because KDE has a hard dependency to 3.2.0 and 
> does not compile with 4.0.0 yet. I'm really not sure if its worth the effort 
> of having both versions supported as it would require #ifdef'ing all the 
> changes from this diff, sometimes in quite complex order. I think by the time 
> we would release this, libotr 4.0.0 has been out the door for more than half 
> a year. That should be enough for distributions to update. If they don't 
> update such a simple lib as libotr in half a year, they are not going to 
> upgrade their KDE version anyways. Once they realize KDE supports 4.0.0 they 
> will most likely upgrade their libotr and the 3.2.0 path just adds more 
> complexity without being used at all.
>     
>     Also, there is another problem... If we would support both versions at 
> compile time, then someone compiles it with 3.2.0, a later upgrade to libotr 
> 4.0.0 would lead to crashes as the API changes quite a bit. So it would not 
> only add maintainance efforts on our side, but add more complexeity for 
> distros to package.
>     
>     So I don't think that would be good. Do you think I should do that 
> nevertheless?
>     
>     > And yes, for adding new strings you need exception.
>     Is that the only exception needed to be able to push this to 4.11 or is 
> there something else needed?
> 
> Pali Rohár wrote:
>     > Also, there is another problem... If we would support both versions at 
> compile time, then someone compiles it with 3.2.0, a later upgrade to libotr 
> 4.0.0 would lead to crashes as the API changes quite a bit. So it would not 
> only add maintainance efforts on our side, but add more complexeity for 
> distros to package.
>     
>     This will not happen, because libraries have soname. And kopete already 
> has #ifdef code to support old/new version of some libaries. If libotr api 
> was not changed too much, I'd like to see support for 3.2.0...
>     
>     > Is that the only exception needed to be able to push this to 4.11 or is 
> there something else needed?
>     
>     Another problem is hard feature freeze and dependences freeze - it is not 
> possible to update dependences on external libraries (increasing minimal 
> libotr version).
>     But because this is review request for bug fixing and specialy allow to 
> compile kopete with new version of library there is chance (after fixing new 
> strings) to include it into 4.11. But libotr 3.x cannot be dropped in 4.11.
> 
> Michael Zanetti wrote:
>     Right... Ok... In that case I'd say the 4.11 train has gone. Lets target 
> for 4.12 then.
>     
>     Regarding the backwards compatibility to 3.2.0: this is the document 
> describing the API changes between 3.2.0 and 4.0.0: 
> http://www.cypherpunks.ca/otr/UPGRADING-libotr-4.0.0.txt
>     IMO too much to #ifdef...

Ok, then after KDE 4.11 branch will be created and kopete master branch will be 
ready for new features and changes, go ahead and commit this patch. I belive 
that in next 6 months distributions update libotr package to new version.


- Pali


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/111227/#review35015
-----------------------------------------------------------


On June 25, 2013, 12:49 a.m., Michael Zanetti wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/111227/
> -----------------------------------------------------------
> 
> (Updated June 25, 2013, 12:49 a.m.)
> 
> 
> Review request for Kopete.
> 
> 
> Description
> -------
> 
> This patch ports kopete-otr to libotr 4.0.0
> 
> Changes:
> -> support for translating libotr-generated messagess.
> -> incoming unencrypted messages during an ecrypted session can now be 
> threated as normal messages producing a notification and saved to history 
> (Bug 204502)
> -> Along with the port I worked around the crash with incoming messages 
> during key generation (Bugs 195328, 218570, 298681, 304105, 306276, 309987, 
> 318255)
> -> Multiple chat session to the same user with different resources can now be 
> distinguished
> 
> Because the original code was rather badly formatted the change also contains 
> some formatting changes which I'm not sure what exactly is the difference. I 
> tried to make as few coding style changes as possible.
> 
> 
> Oh, and it does add some new strings, but shouldn't change any existing ones. 
> The ones newly added are the ones that have been inside libotr before - 
> hardcoded to english. So even if we would ship this patch without updating 
> translations, translation wise it would look the same as before. So, do you 
> guys think we could squeeze it into 4.11 still?
> 
> 
> This addresses bugs 195328, 204502, 218570, 298681, 304105, 306276, 309987, 
> and 318255.
>     http://bugs.kde.org/show_bug.cgi?id=195328
>     http://bugs.kde.org/show_bug.cgi?id=204502
>     http://bugs.kde.org/show_bug.cgi?id=218570
>     http://bugs.kde.org/show_bug.cgi?id=298681
>     http://bugs.kde.org/show_bug.cgi?id=304105
>     http://bugs.kde.org/show_bug.cgi?id=306276
>     http://bugs.kde.org/show_bug.cgi?id=309987
>     http://bugs.kde.org/show_bug.cgi?id=318255
> 
> 
> Diffs
> -----
> 
>   CMakeLists.txt 18896be 
>   cmake/modules/FindLibOTR.cmake e9dbf31 
>   plugins/otr/authenticationwizard.h a4dea40 
>   plugins/otr/authenticationwizard.cpp e4d5d13 
>   plugins/otr/kopete_otr.desktop 2429868 
>   plugins/otr/otrguiclient.h 0347075 
>   plugins/otr/otrguiclient.cpp 11329db 
>   plugins/otr/otrlchatinterface.h 5bb6e1d 
>   plugins/otr/otrlchatinterface.cpp 56ffd0c 
>   plugins/otr/otrlconfinterface.h 0cfdf29 
>   plugins/otr/otrlconfinterface.cpp 7d58462 
>   plugins/otr/otrplugin.h 35ecb7b 
>   plugins/otr/otrplugin.cpp ef973ed 
>   plugins/otr/otrpreferences.h 7aa742c 
>   plugins/otr/otrpreferences.cpp ece403a 
>   plugins/otr/privkeypopup.h 4b271e9 
>   plugins/otr/privkeypopup.cpp 3fea9e2 
> 
> Diff: http://git.reviewboard.kde.org/r/111227/diff/
> 
> 
> Testing
> -------
> 
> Tested with XMPP against Pidgin (OTRv3) and Adium (OTRv2). The overall 
> handling of messages (escaping, parsing etc) did not change with this patch, 
> so it shouldn't break any other protocol. Still, some more testing wouldn't 
> hurt if you have some other protocols in use you can test with.
> 
> 
> Thanks,
> 
> Michael Zanetti
> 
>

_______________________________________________
kopete-devel mailing list
kopete-devel@kde.org
https://mail.kde.org/mailman/listinfo/kopete-devel

Reply via email to