Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-04-24 Thread Harish Raja Selvan via Tiff
Hi Roger,
I’ve created a merge request to add Windows on ARM cross-compilation support to 
the CI:
https://gitlab.com/libtiff/libtiff/-/merge_requests/867
Could you please take a look and share your thoughts when convenient?
Thanks for your time.
Best regards,
Harish
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-03-19 Thread Harish Raja Selvan via Tiff
Hi Roger,
Just following up on my previous message regarding Windows on ARM 
cross-compilation support in CI.
Please let me know your thoughts on this when convenient.
Thanks for your time.
Best regards,
Harish
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-03-09 Thread Harish Raja Selvan via Tiff
Hi Roger,
I just wanted to follow up on my previous message regarding the proposed 
Windows on ARM cross-compilation support in CI.
Please let me know your thoughts on this.
Thanks for your time.
Best regards,
Harish
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-03-02 Thread Harish Raja Selvan via Tiff
Hi Roger,
Yes, the Windows ARM64 container images do exist, but Docker Desktop on Windows 
on ARM currently supports Linux containers only and does not support running 
Windows containers on WoA at the moment. This is noted in the Docker Desktop 
documentation: https://docs.docker.com/desktop/setup/install/windows-install/
Given this, I’ve worked on enabling Windows on ARM cross-compilation using the 
existing Windows x64 Docker-based setup, with a few adjustments in the 
Dockerfile and the GitLab PowerShell build script. This approach builds 
successfully for WoA. The test suite relies on ctest, so the Invoke-Ctest step 
is skipped for the ARM64-related jobs.
If this approach seems reasonable to you, I’d like to propose adding WoA 
cross-compilation coverage to CI in this form. In that case, would VS2026-based 
cross-compilation be sufficient, or would you prefer VS2022 coverage to be 
included as well?
Kind regards,
Harish
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-02-23 Thread Roger Leigh via Tiff
On 27 Jan 2026, at 07:29, Harish Raja Selvan 
 wrote:
> 
> Thanks for the explanation.
> I wanted to share that I’ve successfully built libtiff natively on Windows on 
> ARM64 (X Elite laptop) using the MSVC compiler, without requiring any code 
> changes. 
> Since Windows on ARM Docker images are not currently supported, I wanted to 
> ask whether it would be possible to set up a self-hosted ARM64 runner, if we 
> are able to provide suitable WoA hardware.
> Please let me know your thoughts.

Note that Windows on ARM docker images *are* supported and exist.

See https://mcr.microsoft.com/v2/windows/servercore/tags/list and note at the 
bottom:

"ltsc2025-amd64",
"ltsc2025-arm64”,

Our CI is now using the ltsc2025-amd64 image.  But there is an ARM64 version 
too.

The provision of suitable hardware is a kind offer, but the main problem is the 
ongoing support and administration of the systems to keep them running and 
up-to-date.  That’s the primary reason I switched the Windows AMD64 builds over 
to use containers ( https://gitlab.com/codelibre/containers?filter=libtiff ), 
since it’s then possible to maintain all of the build environment definitions 
and update them under our control, as well as having a clean and unique 
environment for each build.  If an equivalent set of containers can be built 
for ARM64 then it becomes a bit more viable, but it’s still a big time 
commitment to set all that up.  Troubleshooting and support of the old Windows 
build systems and the bring-up and testing and container building of the new 
Windows build systems has already taken several working days out of this month 
for me, and I don’t want to increase this further, especially when it’s not 
really increasing the effective test coverage that much.  I’d like to see how 
the existing setup copes for a few months before adding more systems.  It may 
be viable later on, but for now the existing ones are more than enough of a 
task to manage.

Kind regards,
Roger

___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-02-05 Thread Harish Raja Selvan via Tiff
Hello,
I just wanted to follow up on my earlier message regarding possible Windows on 
ARM64 CI support and the option of providing WoA hardware for a self-hosted 
runner.
Please let me know your thoughts on this.
Thank you for your time.
Best regards,
Harish
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-01-26 Thread Harish Raja Selvan via Tiff
Thanks for the explanation.
I wanted to share that I’ve successfully built libtiff natively on Windows on 
ARM64 (X Elite laptop) using the MSVC compiler, without requiring any code 
changes.
Since Windows on ARM Docker images are not currently supported, I wanted to ask 
whether it would be possible to set up a self-hosted ARM64 runner, if we are 
able to provide suitable WoA hardware.
Please let me know your thoughts.

Best regards,
Harish
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff


Re: [Tiff] Request to add Windows on ARM CI support for LibTIFF

2026-01-07 Thread Roger Leigh via Tiff
On 5 Jan 2026, at 06:54, Harish Raja Selvan via Tiff  
wrote:
> 
> Hello Everyone,
> I am Harish Raja Selvan from MulticoreWare, India. We are currently working 
> on enabling native builds of various open-source libraries for the Windows on 
> ARM (WoA) platform.
> I’d like to start a discussion about extending the LibTIFF CI pipeline to 
> include Windows on ARM (ARM64) support. I wanted to ask whether setting up a 
> self-hosted Windows on ARM runner would make sense for more comprehensive 
> testing and long-term stability.
> 
> Adding Windows on ARM coverage in CI would help ensure consistent build and 
> test validation on this platform. I’d be happy to help contribute 
> configuration changes if this direction seems reasonable.

The current CI setup is designed to cover a range of combinations of hardware 
platforms, operating systems and compilers, but can’t be an exhaustive set of 
every combination.  We don’t currently have any coverage for ARM64; I 
previously had FreeBSD/clang on ARM64 on a Raspberry Pi 4 in the test matrix, 
but had to take it out due to reliability problems with the storage.  I can 
look at bringing it back.

The main limitations for adding more combinations are logistical.  Having and 
maintaining the correct hardware, software, and the ongoing maintenance burden 
of keeping all of the systems up to date.

For our current Windows build, I’m currently looking at migrating from a 
self-hosted runner using the shell executor to a Windows Server system running 
docker, so we can create and use Windows docker images.  This makes keeping the 
build image up to date much simpler, and also has improved security and the 
possibility for using multiple images on the same system such as we have for 
Linux.  We would probably want to do the same for ARM64 Windows to keep it 
manageable.

At present, I am hosting the hardware, but I don’t currently have any hardware 
suitable for running Windows on ARM64.

That said, libtiff should build and run on Windows ARM64 without any trouble at 
all as it stands today, either with MSVC or GCC or Clang.  Have you tried 
building it directly or with vcpkg?


Kind regards,
Roger
___
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff