Date: Thu, 10 Dec 1998 13:26:53 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Andris Pavenis <pavenis AT lanet DOT lv>
cc: djgpp-workers AT delorie DOT com
Subject: Re: egcs-1.1.1
In-Reply-To: <Pine.A41.4.05.9812101154060.41782-100000@ieva01.lanet.lv>
Message-ID: <Pine.SUN.3.91.981210132356.2237H-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com


On Thu, 10 Dec 1998, Andris Pavenis wrote:

>    Could it be acceptable to use this feature for release (or one more
>    test release)?

Why not?  Since cpp is supplied, falling back in case of trouble should 
be no problem, right?

> I think it would give us faster compilation as starting separate
> cpp.exe and writting preprocessed source to temporary file and reading
> it again could be avoided. It would give much for compiling large
> source files but more for many small ones.

In my experience, the overhead of a separate cpp pass is barely noticed 
(no matter how large is the source), unless the disk cache is badly 
misconfigured.

But that is not a reason not to release the package with this feature, if 
that's what you think is best.

> Problem: Perhaps such version will not able to call gcc-2.8.1
> or other earlier version of compiler using -V command line option of gcc
> (also oposite will not work)   

Why would somebody want to mix two versions of gcc?  I thought every 
compiler comes with its own gcc.exe, no?