JogAmp iOS Arm64 Port: First Visuals …

After the GlueGen iOS arm64 bring-up
<https://jausoft.com/blog/2019/06/17/jogamp-ios-arm64-bring-up/>

I can finally show some visuals, having a working 'core' Jogl
port to iOS arm64 using our own OpenJDK build
<https://jausoft.com/blog/2019/06/23/jogamp-ios-arm64-port-first-visuals/>

More work to be done here of course, the 2.4.0 release is also overdue.

All the best,

~Sven

Hi Sven,

Most impressive, but what happens if Apple drops support for OpenGL? Would you end up paying for something like
MoltenGL?

Tom

On 6/24/19 4:46 PM, ThomasR [via jogamp] wrote:
> Hi Sven,
>
> Most impressive,

Thank you.

> but what happens if Apple drops support for OpenGL? Would you
> end up paying for something like
> MoltenGL?

Seemingly current OpenGL driver is based on Metal partially,
so the RENDERER string.

I doubt that they can drop OpenGL, that would destroy quite a market
for Apple considering all professional applications.

Well, next API goal is Vulkan - but we surely will not touch
the walled gardened Metal API.

~Sven

>
> Tom

Hello

Will we keep both .jnilib and .dylib files?

On 6/25/19 10:21 AM, gouessej [via jogamp] wrote:
> Hello
>
> Will we keep both .jnilib and .dylib files?

Will remove jnilib for macos as well ..,
I guess that was the consensus right?

Then I merge the path in NativeLib.. or where I duplicated it
.. source of your good question IMHO.

~Sven

Yes the consensus consists in keeping only .dylib files as this is what Java >= 8 expects.

Good job :) I hope that you'll find a sponsor to work on Vulkan too as it will probably require a greater effort than supporting iOS.

On 6/25/19 12:49 PM, gouessej [via jogamp] wrote:
> Yes the consensus consists in keeping only .dylib files as this is what Java
>>= 8 expects.

Yes. I already started work on Java9+ issues, some in ios branch and
some in java11 branch of GlueGen.
Looks like it is easy to maintain both.
So when we make the Java11 switch, my proposal is to:
- Use OpenJDK >= 11 for building, otherwise this would explode
  the ant build file complexity.

- Build will produce a Java8 baseline compatible bytecode,
  since that seems to be the current Android 'floor' as well.

So, I guess a good compromise.

>
> Good job :) I hope that you'll find a sponsor to work on Vulkan too as it will
> probably require a greater effort than supporting iOS.

MacOS and iOS were and are probably the most time consuming ports we have.

And yes, we certainly need more companies (using JogAmp) paying for it,
sharing & balancing the costs between them. Also allowing 1-2 more
contributors working on it, being compensated.

Well, maybe some companies are following this 'drama' and do the right thing ;-)

Laters & Cheers,

~Sven

Yeah, the -release flag in the jdk 11 no longer requiring a bootstrap classpath seems way easier on the build system side.

Consider this a big +1 to your plan (Build on jdk 11+ with java 8 bytecode )from me.

Harvey

JogAmp iOS Arm64 Port: NEWT

<https://jausoft.com/blog/2019/07/08/jogamp-ios-arm64-port-newt/>

Links & video inside blog, text copy below:

+++

This video demonstrates a working NEWT iOS port:

    Mitigates FBO Depth Buffer iOS Quirk
    Using GearsES2 demo
    PixelScale 1f, 2f and 0f – last two using max pixel scale
    Touch w/ GearsES2 works:
        1 finger rotate
        2 finger drag
        2 finger pinch-zoom
        NEWT gesture detection (last two above)
    Transparency (JogAmp logo in UIWindow behind NEWT window)

Git ios branches will be merged soon to master branch and promoted to all our
git repos

    GlueGen https://jausoft.com/cgit/gluegen.git/log/?h=ios
    Jogl https://jausoft.com/cgit/jogl.git/log/?h=ios

JogAmp https://jogamp.org/
Forum http://forum.jogamp.org/
My related blogs https://jausoft.com/blog/category/computer-stuff/3d-opengl/

We like to thank CCT International for having ordered the JogAmp port running
on Apple’s iOS devices.

+++

On 6/23/19 8:44 AM, Sven Gothel [via jogamp] wrote: