[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#950845: Flags cannot be created on scoped enums



Hi Lisandro, 

Many thanks for taking care of this!


Le ven. 7 févr. 2020 à 16:06, Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com> a écrit :
tag 950845 moreinfo
thanks

Hi Denis!

First of all, thanks for filing the bug. In this way is easier for us to track
issues.

Secondly: I'm keeping this bug open for now, even if it's already fixed in
testing/unstable, as I clearly see you need it solved in Debian stable.

On 20/02/07 12:21, Denis Rouzaud wrote:
> Package: qtbase-default
> Version: 5.11.5
>
> Flags cannot be created on scoped enums
>
> See issue: https://bugreports.qt.io/browse/QTBUG-47652

> This has been fixed in Qt 5.12
> https://codereview.qt-project.org/c/qt/qtbase/+/235284

In you [original mail] you say this bug needs to get solved in order to use it
in QGIS. The fix is already in the archive in testing/unstable, only stable is
still affected. So if you are requesting this change is (at least) because of:

- You are preparing binary uploads for Debian stable, possibly non-official
  (Debian stable has qgis compiled against Qt 4).

Yes this is it.
We are willing to increase the minimum Qt version to allow using scoped enums for flags (because it provides way better API documentation).
Increasing to 5.12 but 5.11+fix on debian would be enough for us.
It would be hard to justify not to support Buster (or to package Qt ourselves).
 
- You need to update qgis in stable (I doubt so, as per the above comment).
- You are preparing a backport.

>From a Qt point of view is indeed a rather important bug to be fixed, but from a
Debian stable point of view is not so:

- Code already in stable does not suffers from this: it's clearly working.
- The bug has been present since Qt 5.5.

I see this. The bug isn't that bad indeed, and explained above it's mainly for documentation purpose we're looking to have it fixed.

I think it will be difficult to convince the stable release team to accept the
patch if the use case is not directly involved in Debian itself.

However I'll try to add the patch and try it myself on a stable system to check
it does not breaks something obvious. For what I can tell from the patch it
shouldn't, but...

I understand that. In any case, thanks a lot for giving it some time. 

Whatever extra information you can provide us to justify the stable upload would
certainly help.


[original mail] <https://lists.debian.org/debian-qt-kde/2020/02/msg00055.html>


Regards, Lisandro.


Best wishes,

Denis 

Reply to: