Hi,
Just pushed Inline-0.49_01 to CPAN.
Most significant change is the attached alteration to Inline.pm's sub
create_config_file().
The renaming of 'my @INC' to 'my @_inc' is just to remove any thoughts that
this might be altering perl's @INC. If someone sees a problem with that,
please let me know - it's *not* part of the fix (that's mentioned next), so
can be reverted if there's a need.
The rest of this patch should render Inline::CPP once more buildable by the
smokers that the cpan-testers use - and should see the end of the FAIL
reports such as:
http://www.cpantesters.org/cpan/report/3b90f976-4d0c-11e1-9d6f-f6dbfa7543f5
It should also mean that anyone else who happens to have Inline and
Parse::RecDescent in different PERL5LIB locations will also be able to build
Inline::CPP - which is something they currently can't do, afaict. (At least,
I got the same errors as the smokers when I created such a set up.)
Should this fix be implemented differently ?
Another solution is to remove the code (in the same sub) that local()ises
$ENV{PERL5LIB} - but I didn't want to take that path for fear of breaking
something. I'm assuming that local()isation does something crucial ... does
anyone know ?
If 0.49_01 tests ok, and no-one has any problems with it, it will shortly
become 0.50 (stable) - so that the Inline::CPP can (hopefully) get back to
getting useful results from the cpan-testers.
For the other changes that have been made since 0.49, see the top-level
Changes file, and the C/Changes file.
Cheers,
Rob