On 10/31/24 02:24, Daniel P. Berrangé wrote:
On Wed, Oct 30, 2024 at 09:04:26PM -0700, Pierrick Bouvier wrote:
MacOS and Linux are straightforward, but Windows needs a bit more
details.
Signed-off-by: Pierrick Bouvier <pierrick.bouv...@linaro.org>
---
docs/about/build-platforms.rst | 4 +-
docs/devel/build-system.rst | 100 +++++++++++++++++++++++++++++++++
2 files changed, 103 insertions(+), 1 deletion(-)
+Windows
+-------
+
+You first need to install `MSYS2 <https://www.msys2.org/>`_.
+MSYS2 offers `different environments
<https://www.msys2.org/docs/environments/>`_.
+x86_64 environments are based on GCC, while aarch64 is based on Clang.
+
+We recommend to use UCRT64 for windows-x86_64 and CLANGARM64 for
windows-aarch64
+(only available on windows-aarch64 hosts).
I don't believe we should be recommending ucrt64 for windows. We have two
testing jobs, one cross compiled from Mingw Fedora, and one native under
Msys2, both of which are validating mingw64. We have no ucrt64 testing.
IIUC, Stefan's QEMU installer binaries are also mingw32/mingw64 based,
not ucrt64.
I have no strong opinion, and was just following default recommendation
of MSYS2 project to switch to UCRT64.
All msys2 environment use mingw64, but mingw64 can target different
windows runtime. From what I understand, any difference or bug would be
a bug for mingw64 project, not for us.
UCRT is now the default target used by mingw64 [1].
More information on this difference here [2].
Migrating to UCRT64 seems to be the best way in the future.
All that said, I'll update this recommendation to MINGW64 for V2 of this
series,
[1] https://www.mingw-w64.org/changelog/#v1200-2024-05-29
[2]
https://github.com/mingw-w64/mingw-w64/blob/master/mingw-w64-doc/howto-build/ucrt-vs-msvcrt.txt
With regards,
Daniel
