This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Intent to remove old VMS wart
- To: Clive Nicolson <srwmcln at baby dot bedroom dot gen dot nz>
- Subject: Re: Intent to remove old VMS wart
- From: Zack Weinberg <zack at wolery dot cumb dot org>
- Date: Thu, 1 Jun 2000 12:55:07 -0700
- Cc: gcc at gcc dot gnu dot org
- References: <009EAFA0.EEECAD00.1@baby.bedroom.gen.nz>
On Thu, Jun 01, 2000 at 10:26:45PM +1200, Clive Nicolson wrote:
> >I'm doing this for three reasons: (1) to the best of my knowledge,
> >DEC's current VMS C compilers do not include this misfeature; (2) it
> >clashes, badly, with the new tokenizing lexer; (3) VMS support is
> >nonfunctional anyway.
>
> The VMS functionallity of gcc was not particularly broken, but I've not built
> from cvs for a while. I see no problem with the removal of that
> oddball VAXC-ism
Do I understand correctly that you have access to a VMS machine?
Could you please try a snapshot after, um, 18 April 2000? The new
preprocessor doesn't know how to convert Unix pathnames to VMS ones.
It might not be necessary anymore - I've heard that the newer C RTLs
handle this for you - but I'd like to know, either way.
(If it doesn't work, please look at hack_vms_include_specification()
in cppfiles.c and figure out when it needs to get called.)
zw