[Kexi-devel] RFC: plan for starting the Qt5/KF5 port

[Boudewijn Rempt]

On Wed, 25 Feb 2015, Dmitry Kazakov wrote:

> Hi!
> 
> My 2 cents :)
> 
> 1) astyle
> 
> Last time astyle was applied to Krita code (something around 2010-2011, it was applied partially?) I really didn't like the result. At least the thing it did with braces and indentation. I guess we just need
> to choose really carefully what to change and what not. E.g. one-line-return-if-error ifs might not have braces. That is I don't agree that the policy statement "Use curly braces even when the body of a
> conditional statement contains only one line" should be followed blindly.
> 
> For the rest, e.g. include style, tab vs spaces, lines with trailing whitespace I'm ok with fixing it automatically.
> 
> 2) If astyle applied, it must be applied to master-only. Under no circumstances to calligra/2.9. Without being able to use 'git blame' it'll be tough to fix many kinds of bugs and do bisecting.

Okay.

> 3) #pragma once
> 
> To tell you the truth I cannot remember a single case of facing a problem with include guards for last several years. Why should we bother about any non-standard compiler extensions then? My emacs scripts
> handle it quite perfectly, I am pretty sure that QtCreator can also crack it. I just don't understand the reasoning. Does MacOS compilers support it well enough?
>

I have fixed a lot of bugs caused by people copy-pasting header files
and not changing the header guards. Of course, everything is fine when 
creating new files, but people don't just create new files. As for 
portability, check once again:

http://en.wikipedia.org/wiki/Pragma_once#Portability

The only compiler that doesn't support it is Solaris Studio C/C++, which 
is not a target platform.

But as long as Qt Creator doesn't support it, I cannot use it.

Boudewijn