On 4/18/05, Michael G Schwern <[EMAIL PROTECTED]> wrote:
> On Sun, Apr 17, 2005 at 09:37:10PM +0200, demerphq wrote:
> > > Which assumes that things are being run from exactly two levels below the
> > > perl source directory such as ext/Foo/. This hard coded assumption had
> > > not shown up before because prior to 6.25_07 there were not tests which
> > > involved executables. Now there is.
> >
> > I looked into this and the line in question doesnt seem to be directly
> > at fault, as the times where the error occurs $self->{PERL_CORE} is
> > set to 0 despite $ENV{PERL_CORE} being set to 1. I figured out how to
> > fix that, and then the error you meant popped up.
>
> Thanks.
I hope that was the right place, I never did quite work out where
ENV{PERL_CORE} gets stuck into $self->{PERL_CORE} in the first place.
> > The attached patch resolves the issue with a hack that works fine for
> > win32 but might cause problems with winCE.
>
> I see what you're doing. Ah, it can use $self->{PERL_SRC} to find the
> source directory!
>
> Good, now I don't have to tear pl2bat apart.
Heh, my thoughts exactly. If $self->{PERL_SRC} is set correctly then
getting rid of the yukky s/// on $^X would be much nicer.
>
> > Theres a minor issue that the /wince branch has its own copy.
>
> The difference between the two is so trivial. They can't be unified?
Im sure they can, its just I was feeling kinda muddleheaded yesterday
and wasnt up to the work/research required. I just wanted to have
EU::MM pass test so I could look into updating the regex patches.
Ill give it a go if nobody else does, but I dont know how to tell if
one is on WinCE and not Win32 from inside perl. I guess ill have to
poke around in the winCE directory to see how different it is from the
win32 directory and see how the make process works out.
Cheers,
Yves
--
perl -Mre=debug -e "/just|another|perl|hacker/"
