From:	pavenis AT lanet DOT lv
Message-ID:	<B0000090432@stargate.astr.lu.lv>
To:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>, djgpp-workers AT delorie DOT com
Date:	Mon, 14 Jun 1999 14:27:33 +0300
MIME-Version:	1.0
Subject:	Re: Small changes to DJGPP.ENV
References:	<B0000090429 AT stargate DOT astr DOT lu DOT lv>
In-reply-to:	<Pine.SUN.3.91.990614141154.8278B-100000@is>
X-mailer:	Pegasus Mail for Win32 (v3.11)
Reply-To:	djgpp-workers AT delorie DOT com

On 14 Jun 99, at 14:15, Eli Zaretskii wrote:
> 
> On Mon, 14 Jun 1999 pavenis AT lanet DOT lv wrote:
> 
> > > Why does this break things?  This should only be a problem if the two 
> > > versions of cpp.exe are different.  Or did I miss something?
> > 
> > They ARE different. gcc-2.95 installation installs cpp.exe it in $prefix/bin.
> 
> Sorry, I'm confused.  It seems that gcc-2.95 installs two versions of 
> cpp.exe: one is cpp itself, the other is actually a variant of gcc.exe, 
> or its symlink (the latter goes to $DJDIR/bin).  This much I understand.

It's not simply symlink. It's a different executable. It appears there 
beginning from gcc-2.95 prereleases.

> But how all this is relevant to GCC 2.8.1?  I don't see two copies of 
> cpp.exe there, and I have gcc281b.zip from October 1998, which is the 
> latest, I think.  Yet you still say that this problem exists in 2.8.1 as 
> well?

I mentioned earlier versions (gcc-2.8.1 and egcs-1.1.2) to say that 
proposed removal of definition of COMPILER_PATH does not break 
them. Even more: if one reads what is written in gnu/gcc-
2.81/readme.DJGPP than he/she removes definition of 
COMPILER_PATH anyway.

Presence of COMPILER_PATH does not break anything if earlier 
versions is used but will break things if we decide to include 
bin/cpp.exe in gcc-2.95 binary distribution. 

Andris

- Raw text -