Re: FreeBSD 4.0

[Marc Espie <espie@quatramaran.ens.fr>]

In article < 16930.937487529@upchuck.cygnus.com > you write:
>  In message < 199909160846.KAA29697@quatramaran.ens.fr >you write:
>  > As far as OpenBSD goes, we are willing to fix our header files over time.
>  > The NULL issue is now solved, and everything will be alright in that
>  > regards for 2.6.
>That doesn't matter.  GCC still needs to work on those systems where you have
>not updated your header files.  Or if we find another need to update a header
>file it is not acceptable to have to wait for another OpenBSD release to fix
>the header files.

>Quite simply, the USER_H overrides are wrong and must go.  Period.

No, there are two issues at work.

I concur that, if the end user want to fiddle with a newer gcc snapshot,
USER_H is in the way.

HOWEVER, from the OpenBSD release point of view, we want to go on fixing
include files, and provide our own.

I believe that both goals are not contradictory. Specifically, when 
installed as the system compiler, gcc WILL use the system includes, and
it should work with them.   If fixincludes has to tweak those headers,
OR if gcc should head its own headers, well... then it's a bug in our
system headers, and we want to fix that...  worse than that: it's a deadly
bug; there are parts of the system, such as the kernel, which are built
-nostdinc, and which desperately need the system headers to be correct.

If a user wants to install a more recent gcc to fiddle with it (or a gcc on an
ancient system that, for some strange reason, they don't want to upgrade), 
no problem with having it install its more modern version of include files 
and running fixincludes and so on...

If you prefer, there are two software packages to maintain here: gcc and
OpenBSD. 

Your goal is to ensure gcc's headers and fixincludes work.
My goal is to ensure OpenBSD system headers are fine.

Allow me to defend my own turf...