Re: Mark Mitchell as C++ co-maintainer

[Doug Semler <doug at seaspace dot com>]

> 
> 
>   In message <Pine.LNX.4.10.9904152103150.1698-100000@tahallah.demon.co.uk>yo
> u 
> write:
>   > On Thu, 15 Apr 1999, Jeffrey A Law wrote:
>   > 
>   > > In hindsight, the improvements to the C++ compiler have been the
>   > > single most important reason people have converted from gcc2 to egcs.  
>   > > Three cheers for the g++ team! :-)
>   > 
>   > My employers standardises on gcc 2.8.1. I've been trying to persuade them
>   > to move to egcs-1.1.2 for a while because 2.8.1 doesn't do namespaces! =)
>   
>   > The sheer irony is that our NT Microsoft Visual DevStudio 5.0 development
>   > environment supports namespaces. How embarrassing.
> http://www.cygnus.com/egcs/egcs-1.0/c++features.html
> http://www.cygnus.com/egcs/egcs-1.1/c++features.html
> http://www.cygnus.com/egcs/egcs-1.0/features.html
> http://www.cygnus.com/egcs/egcs-1.1/features.html
> 

A couple things that might want to be mentioned...
   gcc 2.8.1 as released is now a year old - egcs follows the C++ standard
     much more closely - plus egcs incorporates the changes made to gcc 2.8.1.
   egcs has a much faster turnaround time on bug fixes and enhancements.
   I can get the source tree for the compiler and build all the drivers
     and support libraries in one pass.  If I was using gcc, I used to have
     to go out and get the drivers, libraries, and g77 separately, patch
     gcc so g77 would work, and build everything separately.  Major pain in
     the arse.
   
Actually, the only drawbacks to using egcs here is that the bug fixes come
out almost *too* quickly, and technically speaking, egcs is "experimental."

But I think the experimental tag is really a misnomer...the releases are
of very high quality and very stable.  In fact, it sometimes compiles
our code too correctly...eg it exposes bugs in our code which are not
found by gcc 2.8.1 or DevStudio...

When egcs first came out, I had absolutely no problem convincing management
to use it...for the very reasons mentioned in the web pages above :)
Actually the big reasons are architectural, with the support for IRIX 6,
and Cygwin, we have a common development environment across all our 
platforms, and are able to port over to a new one with ease....we just
added a new port to Linux, and all we had to do was build the compiler
and build the software.  That's it, no muss no fuss.

Now if all the U*ix's would agree on header file/library locations and
contents.....:)

---
Doug Semler                       | doug@seaspace.com
SeaSpace Corporation              | Garbage In -- Gospel Out
Least Senior Software Developer;  | Minister of things to do Next Quarter
Low Man on the Totem Pole         | (but will Never Be Done) DNRC  O-
A closed mind is a terrible thing | Bus Error (passengers dumped)
  
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS/M d---(pu) s++:- a-- C++ UILSH+++$ P--- L++ E--- W+
N++ o-- K? w--(++$) O- M-- V- PS+ !PE Y PGP t(+) 5+++ X+
R- tv+(-) b+(++) DI++++ D G e++>++++ h!>--- r% y+>+++++**
------END GEEK CODE BLOCK------