Re: What's cooking in git.git (Jul 2025, #02; Mon, 7)

From: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Date: Jul 10, 2025, 04:25 UTC
Message-ID: <tuglsm3r64mkgzdh2zdrewpzydz7xmyatqxwbbzpqmltwseefc@65cbcs3pq2qg>
In-Reply-To: <668ce4d4-3b60-47e0-a4a5-dbf73efd1e75@kdbg.org>

On Wed, Jul 09, 2025 at 05:04:51PM -0800, Johannes Sixt wrote:

Show 8 quoted lines

> Am 09.07.25 um 12:41 schrieb Carlo Marcelo Arenas Belón:
> > CC: Johannes Sixt on feedback for SIGCHLD in mingw which uses 17 instead
> >     of 20 (cygwin and others) and therefore seem to cause signal to err
> >     without setting errno as it should.
> 
> I don't know what the relevance of the actual number of a signal is.
> Nobody is supposed to use the number, only the macro. But I may be
> misjudging the situation as a bystander of this discussion.

In d28250654f (Windows: Fix signal numbers, 2009-01-22), the value for SIGCHLD was changed from 0 to 17 mentioning some sort of POSIX system.

the relevance comes that if I build a program in Windows that calls `signal(SIGCHLD)` using that number I get an error back, but nothing on errno, but if the number would be instead 20, like in Cygwin/macOS suddenly the behaviour changes and we get the expected EINVAL(20) in errno.

> Furthermore, there is no such thing as "a signal" on Windows (outside of
> Cygwin), let alone a SIGCHLD. Therefore, this topic about SA_RESTART is
> irrelevant for the MinGW port.

So do you have any objection on changing the value (which should be otherwise unused, and irrelevanted as you pointed out) to 20 then?

Carlo