On 6/6/2024 4:09 PM, Dan Shelton via Cygwin wrote:
On Wed, 6 Mar 2024 at 01:08, Mark Geisert via Cygwin <cygwin@cygwin.com> wrote:
On 3/5/2024 2:42 PM, Dan Shelton via Cygwin wrote:
[...]
posix_spawn() was added to POSIX so a Win32 implementation can use Win32 spawn()
...now I see what you're getting at:
If posix_spawn() is intended to launch truly unrelated processes, with
minimal or no coordination with the launching process, why can't it just
use Windows' CreateProcess? I assume here that's what Win32 spawn() does.
Kindly let me interject this first.
Could the implementor(s) of Cygwin's posix_spawn() possibly comment on
why it ends up doing a fork(), if I understand correctly, rather than
doing what Win32 (or even Cygwin's) spawn() does?
I'm not sure, but I think this would benefit Cygwin greatly to
implement a native posix_spawn(). It should be faster and scale
better. Also the only stuff which do really need fork() are
bash/dash/ksh, and maybe daemon stuff like sshd.
I think I remember back in the last century Cygwin's 'make' got some
attention because fork()s were judged to be slow during Cygwin's early
years. I don't know if the shells got the same attention then or since.
Somebody would have to look.
Re CreateProcess: Maybe CreateProcess() breaks access token or newgrp support?
We prefer dealing with detailed bug reports as opposed to speculation :-)
That said, Cygwin does implement system(), popen(), and the spawn*()
family of functions that all launch subprocesses with CreateProcess()
rather than using fork(). Perhaps one of them would work for you?
Does Cygwin do bounties (paid bug fixes), or Google Summer Of Code?
Not that I know of.
..mark
--
Problem reports: https://cygwin.com/problems.html
FAQ: https://cygwin.com/faq/
Documentation: https://cygwin.com/docs.html
Unsubscribe info: https://cygwin.com/ml/#unsubscribe-simple
