bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#9010: sigprocmask -> pthread_sigmask for Emacs


From: Eli Zaretskii
Subject: bug#9010: sigprocmask -> pthread_sigmask for Emacs
Date: Fri, 08 Jul 2011 10:17:48 +0300

> Date: Wed, 06 Jul 2011 11:11:08 -0700
> From: Paul Eggert <eggert@cs.ucla.edu>
> CC: Eli Zaretskii <eliz@gnu.org>
> 
> To address this issue I plan to install the following patches, after
> some more testing.  The first patch is the gnulib part (almost all
> automatically-generated); the second patch is the patch to Emacs
> proper.  The gnulib part now generates a substitute file lib/signal.h
> for older hosts that don't have pthread_sigprocmask.
> 
> For the Windows port, I expect the simplest fix would be to add a
> pthread_sigmask function that turns around and calls sigprocmask, but
> maybe Windows has pthread_sigmask already?  And I don't expect the
> Windows port will need to worry about the signal.h wrapper.

Windows doesn't have pthread_sigmask, no.  The whole signal mask
feature doesn't make sense on Windows, because there are no Posix
signals per se (SIGINT is emulated, but only in console programs; the
Windows build of Emacs fakes SIGINT in a few more cases).

After your patch, will there be any code in Emacs that still uses
sigprocmask?  If not, we could just rename the w32.c sigprocmask stub
to pthread_sigmask.  If sigprocmask will still be used, then your
suggestion for pthread_sigmask to call sigprocmask is TRT.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]