This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Time to drop fixinc.svr4?


Mark Mitchell wrote:
> >>>>> "Bruce" == Bruce Korb <bkorb@cruzio.com> writes:
> 
>     Bruce> Hi Robert,
> 
>     Bruce> "udk" is.  What is "udk"?
> 
> I think that's the "Unix Development Kit".  I bet Robert (hi, Robert!)
> will find that he can't bootstrap due to the same wchar_t problem
> everyone else is having...  

There are a couple of versions of the UnixWare DK.  I have only the
latest here and it's actually not ever bootstrapped GCC. It is C99-aware
and gets all fouled up by the signatures in the new math libraries and
in the includes.  I haven't had time to teach gcc enough about this
target to make them play nice with each other.  I've waffled between
making the system headers dumber (stripping out the C99-isms) via
fixincludes or trying to make the libstdc++ stuff smarter by teaching
them about C99.

Your fix to wchar_t fixed OpenServer well enough that it did
successfully bootstrap this morning, so I think the correct fix for this
problem has been characterizied.

> The same goes for *-*-sysv4 (I think that covers things like NCR
> MP-RAS).

On IA32, this means MP-RAS and the ports from Microport, Consensys,
UHC, Dell, and a few others that are probably of waning concern.  It
probably includes several non-IA targets.  At one point, there were over
a hundred licensees of SVR4.

RJL

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]