Notable News Items for Rebol
In a project I would not have the time to undertake, Barry (@draegtun) has made a listing of some changes:
https://github.com/metaeducation/ren-c/blob/master/CHANGES.md
For those who are not aware, the engine we codename 'Ren-C' is designed to replace the core of Rebol 3. And as far as the stakeholders involved think, it *is* Rebol 3. We just have a lot of work to do to tidy it up and make sure it's ([better] [along] [every] [metric]). That means if something like cryptography and HTTPS has added weight to the executable over R3-Alpha, we need a way to disable it so you get an executable without it...so we can make it lighter/thinner
So don't mind the executable sizes yet, we're trying to modularize things like the FFI and such so they're optional. But @GrahamChiu has been setting up a build farm for some of the current state:
http://metaeducation.s3.amazonaws.com/index.html
For anyone who thinks it's not happening, it is. Atronix is integrating and adopting the code, and has a business stake in making sure it works. I'm not personally 'ready' with what I see to be the research aspects of the project, but what's there is quite well documented.
I'll also mention that there is a Disqus based forum now. I'm not particularly involved with that, but I will say I've been impressed by the progress that forum software has made since I first looked into it:
https://forum.rebol.info/
So while I was skeptical of Disqus at first, I'd suggest everyone (Nick too) give it a shot and try to have a few conversations there. I'm feeling warmer to it than I did when the idea was first being pitched.
posted by: Fork 12-Jun-2017/15:35:40-7:00
> but what's there is quite well documented.
If it's not obvious, what I meant to say was "well commented":
https://github.com/metaeducation/ren-c/blob/master/src/core/c-eval.c
"Documentation" is different. But I will say, and I think @ShixinZeng can back me up on this, making changes to Rebol's C code and feeling confident about those changes is light years ahead of how it was when we first got the code.
posted by: Fork 12-Jun-2017/15:38:09-7:00
I'm grateful to see such serious work, and excited to hear Atronix has a stake in it. Has Ren-C replaced their existing core code, and been integrated with /View in a production environment? I don't think any other comparable improvements are being made to R3 anywhere else, and if a usable /view version comes out of this, there's clearly no other fork which should be considered the mainstream leader. I'll put some permanent links here to hopefully direct some more traffic that way. Also, it's great to see Android releases in the build farm. Are there any other platforms planned?
posted by: Nick 13-Jun-2017/3:07:35-7:00
> Has Ren-C replaced their existing core code, and
> been integrated with /View in a production environment?
Not yet; last status report was June 1: "On the news about View on Ren-C, I got a few simple demos working."
https://chat.stackoverflow.com/transcript/message/37421190#37421190
Since much of their UI is a custom control, I don't know exactly how much emphasis there's going to be on "traditional" /View.
Personally, I don't want to see a situation where there's more than one GUI dialect with various incompatibilities. It seems to me that since Red is focusing nearly exclusively on the GUI, then that's where people should be putting their effort. If they spec it and document it, and it excites people on HackerNews or wherever, I'm hoping that by the time that comes around it would be relatively little work to simulate that using Qt or in the browser or wherever.
There's actually now efforts on "Qt Lite", so it might not even be all that huge:
http://blog.qt.io/blog/2016/08/18/introducing-the-qt-lite-project-qt-for-any-platform-any-thing-any-size/
> it's great to see Android releases in the build farm.
> Are there any other platforms planned?
@giuliolunati has begun some work for emscripten on the build farm. But emscripten is a particular area where the build size matters a lot. Converting the C to JavaScript bloats it up from ordinary native code, and we want to get as many things that are redundant with the browser to be turned off.
As for other builds...I made a HaikuOS virtual machine, and was able to still build for it. So although we are pushing some more boundaries, the basic premises of portability are still being honored.
posted by: Fork 13-Jun-2017/10:52:55-7:00
Since making Linux executables has come up...I'll also mention that there is Rebol3 progress on things like encapping, which we do in usermode as a Rebol script, not with C:
https://github.com/metaeducation/ren-c/blob/master/src/os/encap.reb
It's just starting out. But the way we're doing it is fancier than what Rebol2 does (for ELF and Windows PE formats). Rebol2 seems to just append the data, this means it won't survive STRIP or other modifications, such as maybe changing icon resources on Windows which would shuffle and add segments to the file. For instance: try doing `strip` on Red's encapped executable, and it stops working.
posted by: Fork 13-Jun-2017/19:17:39-7:00
Thanks Fork.
posted by: SamTheTruck 9-Aug-2017/0:12:53-7:00
|