When both Qt4 and Qt5 are available, let the user pick one at compile time

This issue was migrated from bug 10793 in our old bug tracker.

Original bug information:

Reporter: Jeroen Roovers
Status: RESOLVED FIXED
Product: Wireshark
Component: Qt UI
OS: Gentoo
Platform: x86
Version: 1.99.x (Experimental)

Attachments:

wireshark-1.99.0-qt5.patch: Add configure --with-qt{4,5}

Jeroen Roovers said:

Created attachment 13335

Add configure --with-qt{4,5}

Build Information:

Paste the COMPLETE build information from "Help->About Wireshark", "wireshark -v", or "tshark -v".

Currently configure chooses Qt5 over Qt4 with no mechanic to let the user pick the one they want when both are installed.

added oslinux uiqt versiondev labels

Anders Broman said:

There is issues with Qt4, Qt5 is recomended if available so I'm not sur we want to do this.

Please submitt patches troug git/gerrit https://www.wireshark.org/develop.html

Jeroen Roovers said:

(In reply to Anders Broman from comment #1)  
> There is issues with Qt4, Qt5 is recomended if available so I'm not sur we  
> want to do this.  

That's good to know I guess, but Gentoo is still working out the kinks in its Qt5 deployment and in the mean time we should be able to manage what to use instead of having configure pick something automatically.

> Please submitt patches troug git/gerrit  
> https://www.wireshark.org/develop.html  

I stared at it for a while but it wasn't obvious how I could simply drop a patch there, except by maybe painstakingly rewriting line after line.

Alexis La Goutte said:

(In reply to Jeroen Roovers from comment #2)  
> > Please submitt patches troug git/gerrit  
> > https://www.wireshark.org/develop.html  
>  
> I stared at it for a while but it wasn't obvious how I could simply drop a  
> patch there, except by maybe painstakingly rewriting line after line.  

It is more easy for us, to follow patch and review code
You can look also http://wiki.wireshark.org/Development/SubmittingPatches

Anish Bhatt said:

>  
> I stared at it for a while but it wasn't obvious how I could simply drop a  
> patch there, except by maybe painstakingly rewriting line after line.  

This is actualy super easy to do, but I submitted it to gerrit for you since there seemed to be no followup :
https://code.wireshark.org/review/#/c/6260/

Anish Bhatt said:

(In reply to Anish Bhatt from comment #4)  
> >  
> > I stared at it for a while but it wasn't obvious how I could simply drop a  
> > patch there, except by maybe painstakingly rewriting line after line.  
>  
> This is actualy super easy to do, but I submitted it to gerrit for you since  
> there seemed to be no followup :  
> https://code.wireshark.org/review/#/c/6260/  

I think I am going to abandon this changeset unless you're interested in a followup Jeroen, the patch you submitted does not work as-is, and I do not know this base enough to fix it. I can probably poke around enough to fix it, but based on comment 2, I am not sure if we want to.

Gerrit Code Review said:

Change 8052 had a related patch set uploaded by Guy Harris: Add a major version number argument to --with-qt.

https://code.wireshark.org/review/8052

Gerrit Code Review said:

Change 8052 merged by Guy Harris: Add a major version number argument to --with-qt.

https://code.wireshark.org/review/8052

Guy Harris said:

Fixed, instead, by having --with-qt take an optional version argument, so that:

    --with-qt attempts to find either Qt 5 or Qt 4, preferring Qt 5;

    --with-qt=4 attempts to find only Qt 4;

    --with-qt=5 attempts to find only Qt 5.

And, yes, this is different from how GTK+ is handled and, yes, GTK+ arguably should be handled the same way as Qt is now handled.

closed