X-Spam-Check-By: sourceware.org
Date: Thu, 16 Nov 2006 07:42:11 -0500
From: Christopher Faylor <cgf-no-personal-reply-please AT cygwin DOT com>
To: cygwin AT cygwin DOT com
Subject: Re: Excessive thrashing when popen()ing after a large malloc()
Message-ID: <20061116124211.GA27414@trixie.casa.cgf.cx>
Reply-To: cygwin AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
References: <Pine DOT LNX DOT 4 DOT 58 DOT 0611141146430 DOT 9574 AT u53026a DOT matapp DOT unimib DOT it> <4559D376 DOT 9030707 AT byu DOT net> <loom DOT 20061115T171109-921 AT post DOT gmane DOT org> <20061115165133 DOT GA21115 AT trixie DOT casa DOT cgf DOT cx> <455BD867 DOT 8090702 AT cwilson DOT fastmail DOT fm>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <455BD867.8090702@cwilson.fastmail.fm>
User-Agent: Mutt/1.5.11
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Precedence: bulk
List-Unsubscribe: <mailto:cygwin-unsubscribe-archive-cygwin=delorie DOT com AT cygwin DOT com>
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com

On Wed, Nov 15, 2006 at 10:17:59PM -0500, Charles Wilson wrote:
>Christopher Faylor wrote:
>
>>Fully implementing posix_spawn is quite a bit of work although I did
>>tell Linus Torvalds that I'd look into implementing it in Cygwin
>>eventually (it would make git-on-cygwin faster).
>>
>>I have a patch in my sandbox which implements a more lightweight
>>approach for popen using cygwin internals.
>
>Don't you go thru this exercise about every two years?
>
>1) fork/exec implementation in cygwin sux
>2) attempt to rewrite cygwin internals to use symantics of 
>spawn|vfork|some-other-clever-idea
>3) it breaks a bunch of stuff
>4) try to patch around the problems, repeat #3
>4) revert back to original fork/exec implementation
>
>What's different this time?  (Really, I'm not being snarky, and maybe 
>I'm misremembering, but I'd hate to see precious developer time spent in 
>#2 thru #4 if #5 will happen anyway -- if the past is any guide.)

I did try one time previously to implement vfork/exec transparently as
just a call to spawn* but my implementation caused more problems than it
solved.  That is not the same thing as implementing popen via spawn which
is a more tractable issue.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/