Re: MinGW build on master fails with Error 127

["Óscar Fuentes" <ofv@wanadoo.es>]

[Resending because messages posted through gmane take several hours to
appear on the mailing list. Sorry for the duplicate.]

Eli Zaretskii <eliz@gnu.org> writes:

> And as long as I have your attention: what about the switch to UCRT64?
> Does Emacs need something in its Makefile's to specify linking against
> MSVCRT, even though UCRT64 is the default?  The information at
>
>   https://www.msys2.org/news/#2022-10-29-changing-the-default-environment-from-mingw64-to-ucrt64
>
> doesn't provide any details about that.

See

https://github.com/msys2/MINGW-packages/tree/master/mingw-w64-emacs

There is a patch specific for UCRT support. It's quite simple although
the workaround for close_stream doesn't inspire confidence, hints that
there might be subtle differences on the behavior of C runtime
functions.

This is the relevant commit message:

commit 01be90ce74e3b69e7b03ab886d5f4ac59da0f646
Author: Christoph Reiter
Date:   Thu Jun 23 21:34:12 2022 +0200

    emacs: some hacky patches to make it work on ucrt64
    
    for both stdout and stderr fclose fail during atexit as if they
    were already closed. the function in theory has a workaround for that
    in that it will continue if errno is EBADF, but fclose doesn't set that
    on Windows.
    
    This shows at least what is missing/broken if upstream wants to have a look.


Of course the Emacs UCRT build lacks the real-world testing that the
Mingw-w64 one has, so we need to wait for bug reports.

BTW, there is a patch for building with Clang on Windows.