develooper Front page | | Postings from November 2003

Re: New situation ....

Thread Previous
H.Merijn Brand
November 12, 2003 12:51
Re: New situation ....
Message ID:
On Wed 12 Nov 2003 18:43, wrote:
> <> wrote on 11/12/2003 11:47:13 AM:
> > > I hope that helps.
> >
> > Yes and no.
> >
> > Yes, since 0) is what I did, I now know I did the right thing
> Jarkko's reply would indicate that it is OK for you to proceed
> with a p4 edit and a p4 submit.  For his part Jarkko might try to
> issue p4 revert sooner rather than later (if necessary preceeded
> by an appropriate setting of $P4CLIENT for the environment).

I will prceed :)

> > No, I did get no respons to the patch announcement, from not even one
> > person on the Cc list, so either
> > - nobody cares
> > - all agree
> > - nobody understands
> >
> > Now I'd just like to know what Jarkko is changing, so I could maybe
> > anticipate on that. Let's just hope he did not accidently leave a
> > dangling open p4 edit
> Apparently he did.  Since your comment solicitation was so nicely worded
> allow me to proceed with my own remarks for your consideration.  I presume
> you wanted commentary on the new *.cbu file patch (please correct if I am
> wrong about that).

Actually on all three points, but I'm happy with every feedback.

> Among other things I noted that it seemed to introduce
> new filenames such as for example:

No, it just *moves* some code. No new code is involved, no new filenames are
involved. What this patch changes (it has already been changed in the
metaunits, but that should promote to Configure), is that the already
supported call-back units will now be called whatever the variable they
control is set to. Before the patch, the call-back was only called if the
variable was set.

> > if $test -f uselargefiles.cbu; then
> >       echo "Your platform has some specific hints regarding large file builds, using them..."
> >       . ./uselargefiles.cbu
> > fi
> If I recall correctly Configure still needs to be able to run
> under the port of bash to djgpp on MS-DOS and/or PC-DOS which
> still have the 8.3 file naming limitation.  I guess
> that I would recommend shortening long CBU file names such as
> uselargefiles.cbu, uselongdouble.cbu, etc. so that such call
> back units might be used on platforms with odd filename limitations.
> Admittedly it is unlikely that large files or long doubles will
> be in the C implementations on djgpp, hence the test -f can
> safely bypass those two examples on djgpp.  But I think you
> are trying to set a precedent with the *.cbu files and you
> might want to be mindful of the limitations imposed on odd
> platforms such as djgpp on DOS.

Though your argument is perfect, and you are right, it is probably a lot more
work to change this back. I will put this on my todo list to investigate. A
new problem might arise in backporting this to 5.6.2 and 5.8.0, whereas the
current approach will work as expected. Looking at the state of the hint files
that we have (and IIRC Raphael just blindly copied them back to 5.6.2) this
will be completely safe. Even the patch that will be made to the newly
generated Configure might be easy to apply to 5.6.2 and 5.8.3

> Peter Prymmer

Thanks for making me feel not alone :)

To summarize, if I've done things *right*, nobody will ever notice the changes.
I've only opened doors to new hint possibilities in the future :)

H.Merijn Brand        Amsterdam Perl Mongers (
using perl-5.6.1, 5.8.0 & 633 on HP-UX 10.20 & 11.00, AIX 4.2, AIX 4.3,
  WinNT 4, Win2K pro & WinCE 2.11.  Smoking perl CORE:
send smoke reports to:, QA:

Thread Previous Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About