Message: previous - next
Month: April 2012

Re: [trinity-devel] Re: Qt:: or TQt:: ?

From: "E. Liddell" <ejlddll@...>
Date: Mon, 16 Apr 2012 22:38:58 -0400
On Mon, 16 Apr 2012 13:23:05 -0700 (PDT)
Darrell Anderson <humanreadable@...> wrote:

> > > I just think it's a bit silly that there are multiple
> > > KDE3 forks. I don't see how our project has any incentive to keep our
> > > code the same. Since the changes are nominal, a quick search and
> > > replace will probably make it usable with other kde3 versions.
> > 
> > One of the reasons behind the multiple forks is the TQt VS
> > Qt issue. That's why it is important to have a good motivation and
> > explanation around this. Especially if you consider multiple forks silly
> > and you want to work towards more unification.
> > 
> > I understand that TQt inside the Trinity source was part of
> > an approach that is replaced partly now by another approach on the
> > library side. I hope Timothy can be a bit more clear about all the details.
> > 
> > At one point there was the question of compiling Trinity
> > code against either Qt3 or Qt4 with TQt. Now there seems to be the
> > approach to combine Qt3 and Qt4 usage in a single application (but I
> > also see TQt mentioned then). Right now it has become a bit unclear for
> > most people which things serve exactly which purpose. Maybe there is
> > also already some documentation on this that I'm overlooking?
> There are no technical papers at the wiki explaining the benefits, reasons, or how the TQt layer functions. 
>Without this information, TQt becomes a "black box" or more crudely, a pain in the butt because nobody 
>understands the vision or the technical functionality in order to embrace a positive attitude.

The intention as I understand it is to be able to use QT3 and QT4 in the same program.  The use for this 
that's most often mentioned is replacing Konqueror's aging KHTML rendering engine with Webkit.

*However*, a closer examination of Webkit itself suggests that, if this is the only real use case for QT3 and
QT4 being in the same program, we may be better off writing a QT3 port of Webkit and forgetting about QT4.
More than half of the Webkit <-> QT4 glue code seems to consist of trivial one-line functions, and there's no 
reason to believe a QT3 version would need to be more complex.

Can anyone suggest other uses for QT3 and QT4 together?  Does the new QT4 styling engine require this?
> A few people have left or reduced participation with Trinity because of the TQt debate. I am aware that some 
>people with OpenSuse are keeping KDE3 alive. Yet I am unaware of other KDE3 forks. Where are these projects?

Serghei said he was attempting a fork.  So that's two.