Mailing-List: contact new-httpd-help@apache.org; run by ezmlm
From: TOKILEY@aol.com
Message-ID: <c7.24a7430.25e1a21a@aol.com>
Date: Sun, 20 Feb 2000 15:01:30 EST
Subject: Re: "free" borland C++ compiler
To: new-httpd@apache.org
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit


In a message dated 00-02-20 12:48:38 EST, Marc Slemko writes...

> The summary: Borland C++ 5.5 is available as a free download.  It isn't
> clear if this is just the compiler or the IDE as well (command line
> support would seem to be the most important anyway), but this makes me
> wonder if supporting compiling Apache on win32 with this. 

I have been compiling the Win32 version for Apache using Borland
for almost 2 years now ( See Apache message archives.. search
for 'Borland' ).

There is always a noticeable speed improvement and the resulting
code is always smaller, tighter, and faster. ( Look at .ASM output ).
The Borland Compiler for Win32 has always been a 'better' 'C' 
compiler than MSVC. It even catches MANY coding mistakes that
MSIE simply lets ride on by.

> I don't know about the technical issues involved

There are some. Most of them stem from the fact that there are some
bugs in the base Apache code that MSVC doesn't even 'see'.
Borland catches them and you have to fix them.
The others are mostly the whole 'stdlib' game where Microsoft
has added underscores ( God knows why ) to some stdlib calls.
Borland provides their own 'MSVC stubs' that take care of all this
but you have to include a special header.
Again... see Apache message archive. I believe I left a document
in there that details most of the issues when Brian Behlendorf
first asked me about all this almost 2 years ago.

I believe the Borland makefiles were uploaded as well.

> It still seems that this could be a win in terms of getting 
> people to contribute and helping them in being
> able to do so.

Probably anyone who is even remotely capable of understanding
the Apache code already has MSVC of some flavor but with
Borland's superior ability to catch programming errors it would
be a good idea to use it.

Sometimes... even if I compile a new release with MSVC... I give
it all 1 pass through Borland just to see if it catches any new
'hidden' mistakes in the code that MSVC is going to miss.
Works like a champ every time.

Yours
Kevin Kiley
CTO, Remote Communications, Inc.
http://www.RemoteCommunications.com