FSF offical gcc and egcs gcc: Remarks by RMS (fwd)

[atai at ece dot ucsd dot edu (Andy Tai)]

This mail by RMS is directly relevant to egcs and the future of gcc.
Since no one has mentioned this mail, I post it to this mailing list.

I am sure the FSF leadership and the egcs team can resolve
this potential confusion in the best possible way.     

Of course, we shall thank the egcs team and the FSF people for their work in
gcc which benefits millions of people.  There is nothing that we (the users) can
ask of them, we are given gcc for free.  Because of the importance of gcc, I am
sure  most users do not hope to see gcc becomes like Emacs/XEmacs.   Splits
like that creates confusion that GNU/free systems like Linux can be hurt. 
  



Forwarded message:
> From rms@santafe.edu Wed Dec 17 01:42:57 1997
> From: Richard Stallman <rms@santafe.edu>
> In-Reply-To: <199712162213.OAA12009@vision.ucsd.edu> (atai@ece.ucsd.edu)
> Subject: Re: We still need those old frame functions in libgcc.a
> 
>     Mr. Stallman, as purely a gcc user, I hope you can reconsider your decision
>     regarding compatibility with egcs.  Egcs (gcc egcs-2.9.x) release 1.0 has been
>     warmly welcomed and adapted by many users.
> 
> This was not supposed to happen.  Gumby (David Wallace) of Cygnus
> assured me last August that EGCS was meant to be a way to enable
> people to work on changes for installing them in GCC later on--and
> definitely not intended to make a fork in GCC development.
> 
> The fact that you (and probably other users) think of it as an "actual
> user release" is just what he said would be avoided.  It is disturbing
> because if this continues, it could lead to a real fork.  I think it
> was a mistake to announce a "release" of EGCS, because that invited
> users to think as you do.
> 
> It is normal when trying out new features that sometimes a feature
> gets tried but not permanently adopted.  That is not a problem.  But
> leading people to use EGCS for other than testing causes them to get
> problems from it.
> 
> So I would like to ask you and other users not to think of EGCS as an
> actual user release.  Please use it for testing so that the new
> features can be tried out and developed, but don't think of it as a
> release for ordinary use.
> 
>     If due to incompatibility
>     gcc users have to switch between two versions of compilers and even libraries
>     the trust users put on free software may be harmed.  
> 
> I agree, and therefore the people working on EGCS ought to avoid
> presenting it as a "release" which could become, in effect, a forked
> version of GCC.
> 


-- 
Li-Cheng Tai (Andy Tai)                       e-mail: atai@ece.ucsd.edu