Front page | perl.perl5.porters |
Postings from July 2010
On Sun, Jul 04, 2010 at 12:12:31PM -0400, George Greer wrote:
> On Sun, 4 Jul 2010, Nicholas Clark wrote:
> >Odd. I thought that $ENV{PERL_CORE} was set to true in the environment of
> >the
> >test. (There's an exception list in t/TEST, and ../cpan/Test-Harness is not
> >in it)
>
> Ah, I must've been running the test directly. Under harness there's:
>
> PERL_CORE=d:\tmp\source\t\perl.exe
>
> which makes for a much better patch.
>
> >>patched the batch file to locate perl.exe itself. There is a small danger
> >>that it could locate an errant perl.exe for a CPAN install (it just checks
> >>"../../perl.exe").
> >
> >I'd feel more comfortable if we could work out how to get PERL_CORE into
> >its
> >environment, and use that as a test, if practical.
>
> http://m-l.org/~perl/misc/0001-Make-Test-Harness-source_handler.t-use-the-build-per.patch
>
> http://github.com/greerga/perl/commit/8ed8cbb86a646614a2f538937d1e34784e6509f9
>
> or attached.
Thanks, applied as 08017dae61874ea838912f67e02ec9a1577b9894
I'd forgotten that $ENV{PERL_CORE} is set to $^X
That's a rather useful trick. It seems that it wasn't even the most important
reason to set it to $^X - the main reason was to ensure always set, and always
consistent with tainting:
http://perl5.git.perl.org/perl.git/commitdiff/e447daf9e3940e57
Nicholas Clark
Thread Previous
|
Thread Next