The new browser / QT5

[Dave Yeo]

There's not many options besides the QT based browser. FF52 could still be ported and then the various forks such as PaleMoon could be ported. Really, now there is the Webkit browsers, Firefox and its forks.

[Tuure Linden]

And portin RUST to OS/2 is absolutely impossible?

Personally I think it has been a bad idea to come up with such a programming language. Many legacy systems won't get Firefox releases anymore because of that... For example TenFourFox for PowerPC Macs is now doomed.

[Jochen Schäfer]

Nothing is impossible, but even porting QT5 in C++ is a major undertaking.
In theory, one could build a Rust cross compiler from the C/C++ Rust compiler and then compile Rust with the cross compiler. That's what I read about how most ports are done.
But it's time consuming and it's also a 4 GB dead end.

[Martin Iturbide]

Hi.

And portin RUST to OS/2 is absolutely impossible?

I don't think it is impossible, but not being a programmer I can not estimate the effort. Maybe the logical path was that it was less effort (I'm not saying that is easy or effortless) to port Qt 5 and a Qt browser than posting Rust and compiling Firefox. So that makes more sense for a community like us, which is struggling for resources. (money, developers, etc)

Personally I think it has been a bad idea to come up with such a programming language....

I can not say that Rust is technically good or bad (haven't used it), but I what is interesting is that is a new language that already has a "killer/flagship application" created under his technology, which is Firefox. But, how relevant is Firefox today? or, how strong is Mozilla's funding to push Rust harder on the market/community?. As my personal opinion, at the moment I'm not convinced of the Rust adoption in software, maybe the future will slap me in the face, but I would prefer Qt5, Node.js, OpenJDK to be ported to OS/2 before Rust.

Regards

[Olafur Gunnlaugsson]

And portin RUST to OS/2 is absolutely impossible?

Personally I think it has been a bad idea to come up with such a programming language. Many legacy systems won't get Firefox releases anymore because of that... For example TenFourFox for PowerPC Macs is now doomed.

It is not impossible to port the rust compiler, but it is quite a big project and the only use for it really is for the Mozilla ecosystem, we have had problems maintaining much more popular programming tools in the past due to a lack of available programming talent and/or programmers willing/able to give up their free time to do maintenance  them, so focusing what little resources the community has on it will bring only limited amount of software to the platform.

The language features are also strangely old fashioned with lack of memory management (optional or otherwise) and modularity, especially given that the main problems associated with Mozilla browsers in the past have been related to memory leaks, the lack of a modularity is a pain since the rust toolchain is terribly slow making GGC seem fast by comparison. It would in its current state also need redesigning for use on OS/2, as it stands now it is too memory hungry, in fact has not possible to build the toolchain on 32 bit systems since 2014 (modularity would help a lot with this), although this is a language implementation issue and not a technical issue, i.e. it is perfectly possible to create or modify a rust compiler that works on 32 bit systems, but it will add considerably to the effort needed to port it.

The odd thing about rust while it is indeed a safer language than C++ and this is touted by Mozilla as the main reason for its existence, it is still not a safe language, making it yet another safeish functional language variant of which there are already a few thousands. It is a bit perplexing why mozilla choose to invent a new language that sports no new features rather than using an existing language and toolchain ....

[Dave Yeo]

Porting Rust would be a huge effort. The whole toolchain would need reporting including GCC to use ELF instead of AOUT, then LLVM ported and then could start on Rust. Rust itself was quite hard to even port to the BSD's. Possibly years for a compiler guru to do. Then it sounds like a memory hog which would need more then 4GBs to build Firefox. OS/2 was actually one of the last 32bit platforms that could build Firefox.
A cross compiler might be easier, but still a big job really requiring a developer who really groks compilers and OS/2.

[Tuure Linden]

And there's nothing wrong with the QT besides the short lifecyckle of major versions. It makes writing cross platform software very easy if I have understood right. I've never used the environment myself. What might be the reason why the make new versions so frequently, I have no idea... Back in the day I liked software made with QT3 and KDE 3 series Desktop was great in Linux. I absolutelu hated KDE4 and I don't like KDE5 so much neither. There is Trinity Desktop that is based on the latest KDE 3 release but it has compatibility issues with newer Linux distros and it seems that it's quite difficult to maintain QT3 software in the world of QT4 and 5.

[Paul Smedley]

FYI... from the Qt mailing list.....

It is time to start freezing Qt 6.0.0 initial schedule. I have had some discussion and based on those here is the initial schedule for Qt 6.0.0:

- Qt 6.0 Feature freeze 31.8.2020
- Qt 6.0 Alpha 14.9.2020
- Qt 6.0 Beta 1 1.10..2020
- Qt 6.0.0 RC  17.11.2020
- Qt 6.0.0 Final 1.12.2020

[Pete]

Hi All

Looking at Pauls post I guess we should be thinking about porting qt6 rather than finishing qt5...

Any volunteers to tell Silvan?  :-)


Regards

Pete

[Andi B.]

Hi All

Looking at Pauls post I guess we should be thinking about porting qt6 rather than finishing qt5...

Any volunteers to tell Silvan?  :-)


Regards

Pete

Not a good idea. Stopping running projects in favor of starting new ones usually ends up in finishing nothing. But I'm pretty sure Silvan knows the bigger picture and knows what to do. I for myself see no need to discuss this with Silvan. Think it's better when he does his usual work than discussing such things.

[Martin Iturbide]

Hi

I agree with Andi B. It is better to complete the milestone of the new browser with the Qt5 port and later check the possibility and what is needed to port Qt6. Otherwise it will cause to delay too long the deliverable and support may fade.

But it is always good to keep an eye on what is going on with the Qt community.

Regards

[Dave Yeo]

I think Pete was joking, he even added the

[Pete]

Hi All

Indeed Dave - I suspect Silvan would give about the same answer as Andi, possibly not as politely  ;-)

However, we need to be aware that there will probably be a need to fund a qt6 port in the near future - not sure how near - which starts with people organising funding a little in advance of need. Not my area of expertise so possibly a Roderick/OS2VOICE "project" in the near future...


Regards

Pete

[Martin Iturbide]

Hi

And we also need to remember that developers takes some time to adopt newer versions of the framework before updating their software. Maybe in a year there will be nice open source things that runs on Qt6.

If possible, after having a solid Qt5 and a modern browser on OS/2, I think it will be interesting to fund Bitwise to update OpenJDK. But it would be good to discus the relevance of Java today and if there are interesting Java applications that would be great to run on OS/2 today.

The other thing that catch my eye are the javascript frameworks that young people are using today to produce multiplatform applications like node.js and others. I would like to have more time to evaluate and understand that frameworks.

Regards

[Joop]

If possible, after having a solid Qt5 and a modern browser on OS/2, I think it will be interesting to fund Bitwise to update OpenJDK. But it would be good to discus the relevance of Java today and if there are interesting Java applications that would be great to run on OS/2 today.

In my search for Java programs, still counting and up to 150 today!, there are lots of programs which have an update or which need a higher level of Java. New features but also databases, new apps which can deal with the new rules from Google or Youtube or what about full featured daw's, the only solution for OS/2 to run serious sound applications which is bugged or even absent in present jdk6. New games, new whatever you can think off. If we could run them all the count will go up to may even 250 programs. So, yes, Java is still important special for OS/2 because it does have applications which are not build native on OS/2. Also lots of applications which uses Java with QT, so the need for update is also in this corner. And to be honest, I'm a little disappointed about the QT spread of applications, had expected more. But that's my opinion. OpenJDK version 11 or the next major version is needed for OS/2 in order to have an interesting spread of useful applications. Even if they start today with porting and so on, it will take also not only money, but also a year or so to do the job.