[PATCH v2] Cygwin: signal: Do not handle signal when __SIGFLUSHFAST is sent

[Jeremy Drake]

On Thu, 9 Jan 2025, Corinna Vinschen wrote:

> Hi Jeremy,
>
> Can you perhaps bisect 3.5.5 to tell us which signal-related patch
> exactly has introduced this specific hang?  It may give us a clue.

I believe Johannes was planning to do this, after holidays/vacation.  It
seems that Git for Windows' test suite is the most reliable way to
reproduce this.  There are occasional hangs in MSYS2 builds, but they
go away on a re-run of the build.  FWIW, I don't think we've even proven
that this is signal-related.  The latest backtraces seem to suggest pipes,
but I wouldn't bet on anything at this point.  The only reason signals
came into it is that the last (more common) hang was signal-related.