What is up with gcc on MacOS ?

[Paul Johnson]

Greetings, Bill.

I don't use MacOS, but I am curious what you mean by this. I've been 
digging around in the gcc archives and Mac web pages.

I think here you refer to the version of gcc-3.1 that is distributed 
with the MacOS X operating system v.10.2, right? One would not want to 
use the gcc that was packaged with v10.1, since that was gcc2.

I read in the MacOSX 10.2 release info that one can rebuild their gcc.

If you go here:
http://www.opensource.apple.com/projects/darwin/1.4/projects.html
you see "Darwin" related open source projects, and there you find a link 
to this source code:
http://www.opensource.apple.com/projects/darwin/1.4/source/other/gcc-932.1-6.tar.gz

It is not clear to me what the version number means, and I wonder if it 
is the same as 3.1 that they distribute with v10.2 or if it is an 
updated snapshot.

IT is true, from what I read, that the apple-extended version of gcc 
will never be exactly the same as the original gcc.  But much of the 
Objective-C work did get patched into the FSF version of gcc 3.2. The 
Apple folk are the ones doing much  of the Obj-C work on the current gcc 
snapshots.  I believe the snaps of gcc reflect the Apple magic. The 
release notes for these gcc-3.2 snaps indicate that messaging in Obj-C 
is accelerated by 2 or 3 times.  Stan Shebs <shebs at apple dot com> 
posted a note indicating that gcc-3.2 includes most general compiler 
optimizations from Apple.

So, are you really certain that the newest gcc-3.2 is an "unsupported" 
compiler from the Mac point of view?


address@hidden wrote:
> Hi Alex
>
> I think you are using the FSF build of the compiler rather than the Apple 
> one.
>
> AFAIK this is a very unsupported combination.   It is also sub-optimal 
> because it omits a substantial number of important optimisations which 
> have been added by Apple but not yet accepted by the FSF people.  The 
> speed difference has been reported as around 40%!
>
> Do you have some compelling reason not to use Apple's standard compiler 
> (accept for the one function call patch)?  This version is well tested. 
> Indeed it almost certainly much the best tested of any of the gcc 
> objective-C compiler versions, because it is actually used to build the 
> operating system.
>
> Bill Northcott
>
> address@hidden wrote on 12/09/2002 08:15:22 AM:


--
Paul E. Johnson                       email: address@hidden
Dept. of Political Science            http://lark.cc.ku.edu/~pauljohn
University of Kansas                  Office: (785) 864-9086
Lawrence, Kansas 66045                FAX: (785) 864-5700


                 ==================================
  Swarm-Support is for discussion of the technical details of the day
  to day usage of Swarm.  For list administration needs (esp.
  [un]subscribing), please send a message to <address@hidden>
  with "help" in the body of the message.