Message-Id:	<3.0.6.32.19990212201020.00818c40@pop.netaddress.com>
X-Sender:	pderbysh AT pop DOT netaddress DOT com
X-Mailer:	QUALCOMM Windows Eudora Light Version 3.0.6 (32)
Date:	Fri, 12 Feb 1999 20:10:20 -0500
To:	djgpp AT delorie DOT com
From:	Paul Derbyshire <pderbysh AT usa DOT net>
Subject:	Re: Re-entrancy questions.
In-Reply-To:	<199902122156.QAA21804@envy.delorie.com>
References:	<3 DOT 0 DOT 6 DOT 32 DOT 19990212164718 DOT 0081b960 AT pop DOT netaddress DOT com>
	<3 DOT 0 DOT 6 DOT 32 DOT 19990212164718 DOT 0081b960 AT pop DOT netaddress DOT com>
Mime-Version:	1.0
Reply-To:	djgpp AT delorie DOT com

At 04:56 PM 2/12/99 -0500, you wrote:
>
>> Is the DJGPP 'new' thread-safe already, or should someone developing
>> a threading library for DJGPP wrap 'new' to be thread-safe?
>
>I don't think anyone has put any efforts into making djgpp's libraries
>re-entrant or thread safe, so I suppose the answer is no unless
>someone does the research to prove otherwise or make it so.

What about on other systems? E.g. anything other than DOS. Unix,
principally. Is malloc, new etc. always thread-safe there? Not just
interprocess safe but thread safe too, so if you use pthreads or LWPs
they'll work without needing to be redefined (in pthread.h or lwp.h or user
code).

-- 
   .*.  "Clouds are not spheres, mountains are not cones, coastlines are not
-()  <  circles, and bark is not smooth, nor does lightning travel in a
   `*'  straight line."    -------------------------------------------------
        -- B. Mandelbrot  |http://surf.to/pgd.net
_____________________ ____|________     Paul Derbyshire     pderbysh AT usa DOT net
Programmer & Humanist|ICQ: 10423848|

- Raw text -