Message-ID: <B0000054717@stargate.astr.lu.lv>
From: "Andris Pavenis" <pavenis AT lanet DOT lv>
To: DJ Delorie <dj AT delorie DOT com>, djgpp-workers AT delorie DOT com
Date: Thu, 10 Dec 1998 16:06:11 +0200
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: egcs-1.1.1
In-reply-to: <199812101321.IAA08001@envy.delorie.com>
References: <Pine DOT A41 DOT 4 DOT 05 DOT 9812101154060 DOT 41782-100000 AT ieva01 DOT lanet DOT lv>	(message from Andris Pavenis on Thu, 10 Dec 1998 12:07:45 +0200 (WET))
X-mailer: Pegasus Mail for Win32 (v3.01d)
Reply-To: djgpp-workers AT delorie DOT com

On 10 Dec 98, at 8:21, DJ Delorie wrote:

> 
> > I wrote earlier I had some problems with experimental feature of
> > egcs-1.1.1 that allows to exclude need of separate cpp.exe for compiling
> > C, C++ (and perhaps also Objective C) sources. Of course cpp.exe is
> > still used if somebody wants to get preprocessed sources or assembler
> > source should be preprocessed.
> 
> I hear that the built-in cpplib doesn't parse source the same way as
> the external one, and that the performance isn't as good as the
> external one (spawns included).  They're still working on it.
> 

Yes there are things to test. There can also be differences in output of 
'gcc -E' as 'gcc -E ...' uses cpp then can make diagnostic of possible 
problems more difficult.

About performance:
   I perhaps should check both variants when building some large project
   in some variants (with RAM disk and without it)

One more thing. As I understand M$ DPMI servers leaking is proportional
number of processes started (maybe I'm wrong). If so reducing number
of processes can slightly help.

Andris