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

Bug#803423: #803423 libqt5gui5: Ugly GUI (no icons) / Weird interpretation of XDG_CURRENT_DESKTOP and GNOME_DESKTOP_SESSION_ID



On Friday, 22 April 2016 12:35:45 AM AEST Lisandro Damián Nicanor Pérez Meyer 
wrote:
> I have seen this question asked before on Qt mailing lists' and the answer
> is: which would be a sane default on Linux? DEs need to set
> XDG_CURRENT_DESKTOP.

My use case does not involve DE. At least not directly. I run applications on 
virtual X screen and export using Xpra to another machine.


> So basically this is not a bug in Qt5 "per-se", but the lack of something
> that creates a platform theme to use "by default".

I think this is a bug and it should be forwarded. I insist that lack of 
sensible defaults is a bug if application starts crippled without certain 
environment variables set. I wonder how did they even arrive on such sick 
idea??

Another problem is such defaults selected silently without any warnings 
printed to console... QT5 should really say something when expected essential 
environment variables not set...


> Upstream said it will not provide such a thing because it's up to the DE to
> set the appropiate value. I have two choices here: close the bug or mark
> it as wontfix. As I consider upstream's choice a valid one (even if I
> might not fully agree) I'm opting to close it.

I think there are more than two options. :)
We could consider patching the problem and set up defaults according to 
relevant recommended packages (if any). I understand if you don't want to do 
that with contradiction to upstream decision in which case it should be OK to 
close this bug. I tend to treat bugs as TODO list entries so I believe it is 
OK to close bugs that require no action from maintainer.

However in this case I'd much appreciate if you could set "forwarded" header 
to relevant upstream bug.


> But thanks to Arch's [wiki] I've learnt about [qt5ct] "which provides a DE
> independent Qt5 QPA and a configuration utility". So basically we need
> someone to package it.

I'm already contemplating packaging "qt5ct" but I feel that would not be an 
ideal solution. According to principle of least surprise, application should 
behave predictably. Before this mis-innovation of QT5 most GUI applications 
behaved normally when started on plain X.
Now, much like myself, one can spend hours trying to find out what's wrong 
and even bug innocent projects that inherited this misfeature from QT5... :(

-- 
Regards,
 Dmitry Smirnov.

---

If any remedy is tested under controlled scientific conditions and
proved to be effective, it will cease to be alternative and will simply
become medicine. So-called alternative medicine either hasn't been
tested or it has failed its tests.
        -- Richard Dawkins, 2007

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: