--===============8124471604401006182== MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit > On Jan. 6, 2017, 11:29 a.m., Kevin Funk wrote: > > src/syntax/data/cpp.xml, line 1566 > > > > > > Looking at that again. Do we really, really need to add all classes to the syntax highlighting file? > > > > Did anyone ever measure the performance impact of this? > > > > IMO that's the job of KDevelop to auto-complete... If you want to have proper auto-completion, use KDevelop. > > > > Highlighting Qt macros & Qt global functions (starting with a 'q') are still fine though. There are way less items for those. > > Friedrich W. H. Kossebau wrote: > If all Qt classes, also from all the non-core Qt modules, are added to this non-ISO-C++ syntax file, would it not be consequent to also add all KDE Frameworks classes & macros as well? That is something I would expect from a data file being part of KDE Frameworks itself. Not only for promotional reasons. > If the answer is no, the reasoning for that should perhaps be applied to deciding what Qt classes to include here as well. > > Dominik Haumann wrote: > KDE classes are only useful for KDE developers - and let's face it: that is us, and essentially pretty much no one else. > > Qt is by far more widely used, so we indeed can argue that is is useful. As mentioned already in another post, I still am against adding /both/ Qt4 and Qt5 classes, especially since Qt5 is mostly Qt4 compatible. > > With respect to performance: We could add a unit test in the syntax-highlighting framework, then you will know for sure. > > But besides that, given the keyword list is a binary search, I think this is fast, even when we have 200 or so more items in the list. My opinion would be: 1) Its a Hash for the keywords, I doubt it makes a measurable diff to have there some more or some less ones. 2) I would kill Qt4 and only keep Qt5, Qt4 is more or less dead for "new" things 3) Qt is one of the only toolkits around for C++, I think it makes sense to have per default highlighting for it. - Christoph ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/128631/#review101833 ----------------------------------------------------------- On Aug. 8, 2016, 7:26 p.m., Thomas Fischer wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/128631/ > ----------------------------------------------------------- > > (Updated Aug. 8, 2016, 7:26 p.m.) > > > Review request for Kate. > > > Repository: ktexteditor > > > Description > ------- > > Adding Qt5 classes and a non-class (`QStringLiteral`) to C++ syntax highlighting > > > Diffs > ----- > > src/syntax/data/cpp.xml c3a5cbe > > Diff: https://git.reviewboard.kde.org/r/128631/diff/ > > > Testing > ------- > > > Thanks, > > Thomas Fischer > > --===============8124471604401006182== MIME-Version: 1.0 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: 7bit
This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/128631/

On January 6th, 2017, 11:29 a.m. UTC, Kevin Funk wrote:

src/syntax/data/cpp.xml (Diff revision 1) 

    1566
    
  <list name="Qt5Classes">

  


 




  
Looking at that again. Do we really, really need to add all classes to the syntax highlighting file?


Did anyone ever measure the performance impact of this?


IMO that's the job of KDevelop to auto-complete... If you want to have proper auto-completion, use KDevelop.


Highlighting Qt macros & Qt global functions (starting with a 'q') are still fine though. There are way less items for those.

 




 
On January 6th, 2017, 1:31 p.m. UTC, Friedrich W. H. Kossebau wrote:

 

  
If all Qt classes, also from all the non-core Qt modules, are added to this non-ISO-C++ syntax file, would it not be consequent to also add all KDE Frameworks classes & macros as well? That is something I would expect from a data file being part of KDE Frameworks itself. Not only for promotional reasons.
If the answer is no, the reasoning for that should perhaps be applied to deciding what Qt classes to include here as well.

 






 
On January 6th, 2017, 7:08 p.m. UTC, Dominik Haumann wrote:

 

  
KDE classes are only useful for KDE developers - and let's face it: that is us, and essentially pretty much no one else.


Qt is by far more widely used, so we indeed can argue that is is useful. As mentioned already in another post, I still am against adding /both/ Qt4 and Qt5 classes, especially since Qt5 is mostly Qt4 compatible.


With respect to performance: We could add a unit test in the syntax-highlighting framework, then you will know for sure.


But besides that, given the keyword list is a binary search, I think this is fast, even when we have 200 or so more items in the list.

 











My opinion would be:


1) Its a Hash for the keywords, I doubt it makes a measurable diff to have there some more or some less ones.
2) I would kill Qt4 and only keep Qt5, Qt4 is more or less dead for "new" things
3) Qt is one of the only toolkits around for C++, I think it makes sense to have per default highlighting for it.








- Christoph






On August 8th, 2016, 7:26 p.m. UTC, Thomas Fischer wrote:











 

  


Review request for Kate.


By Thomas Fischer.




Updated Aug. 8, 2016, 7:26 p.m.












 Repository: 
ktexteditor





Description 

 

 

  
   
Adding Qt5 classes and a non-class (QStringLiteral) to C++ syntax highlighting

  		
 







Diffs 


    

 
  • src/syntax/data/cpp.xml (c3a5cbe)
    • 





View Diff







  		
 










  

 



--===============8124471604401006182==--