develooper Front page | perl.perl5.porters | Postings from April 2007

Re: VMS and Win32 Makefiles in change 31059

Thread Previous | Thread Next
From:
demerphq
Date:
April 25, 2007 03:26
Subject:
Re: VMS and Win32 Makefiles in change 31059
Message ID:
9b18b3110704250325n361d5a90l7e80fc1a76b83a4@mail.gmail.com
On 4/25/07, Nicholas Clark <nick@ccl4.org> wrote:
> On Wed, Apr 25, 2007 at 10:06:04AM +0100, Steve Hay wrote:
> > Nicholas Clark wrote:
> > >However, I can't test Win32 or VMS. I've tried to write the makefile
> > >sections
> > >based on cut&paste from the rules for miniperl and x2p, but I suspect I
> > >have
> > >bugs:
> >
> > There were a few slips in the Win32 changes, mostly relating to paths
> > (the Win32 build is done from win32/, so the generate_uudmap.* paths are
> > all ..\ from there, plus the globals.c file first gets compiled into the
> > mini\ sub-directory), but also a missing "<<" to end a temp file and the
> > EMBED_EXE_MANI needed for VC8 was missed.
> >
> > I'm going to test things with the attached changes, and will apply it
> > (or similar) later.
>
> Thanks.
>
> I forgot to add last night that I had no idea where to start on makefile.ce,
> so didn't touch it. Therefore I infer I broke the WinCE build. Sorry.
>
> I couldn't see a good way to ship a default uudmap.h and have it correctly
> rebuilt locally. (Without all sorts of Makefile mess). Maybe we should put
> a copy of the ASCII uudmap.h into Porting/, to keep the sample config.h
> company.
>
> I hope it's fairly easy to recover, as generate_uudmap only needs to be
> built and run on the host machine. It's just that with a cross compile,
> one has to start paying attention to what is the host compiler and what is
> the cross compiler, and with no way to test any of this, I suspected that
> any guess on my part would have been sufficiently far from working code that
> a clean slate would be easier.

Im just curious, why is it necessary to go to this extent? Surely the
only place where the table would differ would be on an EBCDIC machine?
In which case a script in Porting and a reminder for EBCDIC buillders
to regenerate as needed would suffice.

I ask mostly because what i describe is precisely how regcharclass.h
is generated (and it too will have problems on EBCDIC, or rather, it
doesnt know about EBCDIC, so its likely)

Cheers,
yves

-- 
perl -Mre=debug -e "/just|another|perl|hacker/"

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About