On Wed, 8 Nov 2000 [EMAIL PROTECTED] wrote:
> "Black, Andrew" <[EMAIL PROTECTED]> writes:
> >> Slightly O/T - just to say that CHECKSUM does not always produce different
> >> results, even on very different files.]
> >
> > Also be warned that same file (when backed up and restored) might have a
> > DIFFERENT checksum. This can be caused by data "hidden" beyond the end of
> > file byte for the file.
>
> Well, damn.
>
> Okay, well maybe DIFF is the right way to go...I just wish that the $STATUS
> values for differ/same were something other than "random strings of hex
> digits". If they had values that we could use F$message to translate, for
> example.
>
> Hmm...looks like the message is in SYS$MESSAGE:PRGDEVMSG (on my system)
> %DIF-S-SAMEFILE, files are identical
> %DIF-I-FILAREDIF, files are different
> And the two status code correspond to "ID#s" 1 and 2 (with different
> severity, of course).
>
> I'll see about cooking up a patch based on DIFF and see how it goes.
I suppose that DIFFERENCES would be a good tool to use. But am I
understanding the proposed architecture correctly?
Goal: determine from configure.com (hence limited to DCL) that the
perl_src:[vms]perly_c.vms file is nearly the same as
perl_src:[000000]perly.c except for the bits that the perly
fixer script adds to declare globals etc.
Algorithm: carry around another file perl_src:[vms]perly_old.vms
that corresponds to the last version of [.vms]perly_c.vms
that was known to work and make sure that [.vms]perly_c.vms
either differs or has a different checksum from either
perl_src:[000000]perly.c or perl_src:[vms]perly_old.vms
is that a correct synopsis? Does having the extra file seem like a high
price to pay to ensure that the unix perl pumpking has run a perl script
to update things?
Apologies in advance if I have mischaracterized the proposal. Please
understand that I have not had very much time to read it in detail yet.
Peter Prymmer
