[EMAIL PROTECTED]:
# At 21:51 on 06/07/2002 PDT, "Brent Dax" <[EMAIL PROTECTED]> wrote:
#
# > # Of these, 2 are not defined by ANSI C89:
# > # read (in core_ops.o,core_ops_prederef.o)
# > # write (in core_ops.o,core_ops_prederef.o)
# > #=20
# > # Pretty good!
# >
# > Indeed. Those should probably be surrounded with ifdefs--does
# > Configure generate a HAS_HEADER for these?.
#
# Not sure- haven't looked at them yet.
#
# > The only thing that jumped out at me were the direct calls
# to fprintf
# > with warnings--perhaps you should use the standard warnings
# mechanism
# > instead.
#
# You're right- I meant to go back and fix those and totally forgot.
#
# > I'd suggest that your next steps include modifying
# > config/gen/config_h.pl to output a has_header.h with only
# ANSI headers
# > enabled. (Make sure that Configure tests for them in the
# first place,
# > though!)
#
# It's actually doing that already (it's in config/auto/headers.pl)
So you did. I kinda skipped past that. *smacks his forehead*
# > The other thing I'd suggest is that we wrap mmap and munmap in
# > something. I think Windows has similar functionality through a
# > different API, so Parrot_map and Parrot_unmap might be in order.
#
# Agreed. I'll take a look at that as well.
If you do that, I'd further suggest that on ANSI's Parrot_map you just
malloc() the memory and slurp the file in, then free() it in
Parrot_unmap. That way we never have to check to see if mmap or an
equivalent is actually available.
--Brent Dax <[EMAIL PROTECTED]>
@roles=map {"Parrot $_"} qw(embedding regexen Configure)
Early in the series, Patrick Stewart came up to us and asked how warp
drive worked. We explained some of the hypothetical principles . . .
"Nonsense," Patrick declared. "All you have to do is say, 'Engage.'"
--Star Trek: The Next Generation Technical Manual
