GTK3 is ready?

[process_1]

Is it GTK3 ready?

According to "Project Roadmap" it is still in Alpha stage, but it is not to trust these info, as it is seldom updated.

https://wiki.lazarus.freepascal.org/Roadmap

Zeljko?

[trev]

See GTK3 Interface last updated April 20, 2020.

[af0815]

Sorry for using this old post (more than 2 yrs old)

But the question ist still actual. Is there any step forward to gtk 3 ? The informations i found are not actual.

[JuhaManninen]

But the question ist still actual. Is there any step forward to gtk 3 ? The informations i found are not actual.

The best progress information is in the Git development branch "main".
Do for example:

Code: [Select]

$ git log lcl/interfaces/gtk3/or use one of the GUI tools for Git.
The progress is slow. Nobody actively maintains it. Contributions are welcome.

[dbannon]

af, Qt5 is ready now and qt6 only head back by an upstream problem. This is obviously an huge credit to Zeljko but might also be an indication that GTK3 is just that bit harder too.

Davo

[PierceNg]

https://blog.gtk.org/2020/12/ - GTK4 development started in Oct 2016 and was released in Dec 2020, two years ago.

I have no idea the effort involved in building an LCL widget set interface, but maybe consider going straight to GTK4 and dropping GTK3? Could be better than being always in catch-up mode.

[JuhaManninen]

https://blog.gtk.org/2020/12/ - GTK4 development started in Oct 2016 and was released in Dec 2020, two years ago.
I have no idea the effort involved in building an LCL widget set interface, but maybe consider going straight to GTK4 and dropping GTK3? Could be better than being always in catch-up mode.

Not really. GTK is notorious for changing their API also after a release. During the years their API stabilizes. It is actually a good strategy if developers know it and understand what it means.
For LCL bindings it means we should not jump to the latest GTK version right away.
GTK3 has been stable enough for few years. The slow development of LCL bindings is caused purely by lack of resources. Apparently nobody wants it badly enough.
A (big enough) bounty could help but it is not certain. It still requires somebody with the skills, time and motivation to do it.

Somebody could say that a stable API is the most important priority and GTK should change their API.
OK, a stable API has its merits but what if you continue it for decades? For example Xlib is very stable but how many people want to use it?

[Bogen85]

A (big enough) bounty could help but it is not certain. It still requires somebody with the skills, time and motivation to do it.

Time I think is the biggest factor for many of us...

[Tony Stone]

A (big enough) bounty could help but it is not certain. It still requires somebody with the skills, time and motivation to do it.

Time I think is the biggest factor for many of us...

I lack the skill.  The motivation for me is there some what.  But time is some what limited also...  I guess I would be willing to put $1000.00 on it if enough other people wanted to chip in and create a large bounty for the right developer.  Maybe every forum member can throw some money into a pot and motivate the properly skilled developer?   

[Leledumbo]

Looks like Zejlko is only interested in Qt these days, which I'm a user of. I only fallback to GTK when there's no other way, as I find Qt aesthetically more pleasing. Sometimes I got held back by libQt<n>Pas, but it's not for long usually.

[af0815]

BTW it is also a stucked since 2020 Freepascal and Lazarus Foundation Projekt (Source: https://foundation.freepascal.org/projects/complete-gtk3-widgetset )

[dbannon]

I think ".... but no actual development has been done" would not be a fair statement, if you watch the commits ( https://gitlab.com/freepascal.org/lazarus/lazarus/-/commits/main/lcl/interfaces/gtk3 ) they do come in but there is an awful lot to be done.

Parts of it are working, just lots and lots of debugging.

Davo

[3oheicrw]

There is still occasionally updates you could check via github.

[af0815]

But it will a good idea to keep this https://wiki.lazarus.freepascal.org/Roadmap more up to date. For QT5 and partly QT6 it is done.

Only looking in the git logs is a good point for developer, but for user of the lcl is is not read/understandable, except the information, here are guys on work.

[theo]

There is still occasionally updates you could check via github.

There is no real progress. just a few details here and then.
That's not the kind of progress it really needs.
Even the existing code should be revised from scratch imho.
But that's not easy. Understanding GTK3 / Pango / Cairo is not easy and a deep understanding of the LCL is also necessary.
I think only a talent like Zeljko can do that.