I think Jarkko has mentioned that the perly.[ych] that came with 8200
were sub optimal and that one ought to revert the the ones that were
distributed with 8150. I am not sure if that kit's perly files needed
to have vms_yfix.pl run on them or not.
Peter Prymmer
On Tue, 19 Dec 2000, Craig A. Berry wrote:
> When building 8200 with DEC C V5.2-003 on OpenVMS Alpha V7.1, I get the
> following error as it tries to insert the perly object code into the object
> library:
>
> %LIBRAR-E-DUPGLOBAL, global symbol PL_YYDEBUG from file
>DISK8:[BERRYC.PERL]PERLY.OBJ;1 already in library DISK8:[BERRYC.PERL]LIBPERL.OLB;1
>
>
> perly.c now contains the line:
>
> int yydebug; /* nonzero means print parse trace */
>
> which appears to be new, and vms/vms_yfix.pl turns the int into a dEXT int,
> which on VMS defines to be the VMS-specific globaldef int (globaldef allows
> us to shove it into a named program section whose attributes can be
> specified).
>
> The PL_yydebug symbol already appears in mg and globals, and I believe it is
> the define of PL_yydebug in perlapi.h that puts it there and makes it an
> external symbol of some kind:
>
> #define PL_yydebug (*Perl_Iyydebug_ptr(aTHXo))
>
> I'm not sure I've got this analysis 100% right, and I don't know what to do
> about it other than hack vms/vms_yfix.pl to treat yydebug differently, but
> that seems kludgy at best. Advice welcome.
>
>
