On Sun, Apr 14, 2002 at 06:56:02PM -0500, Craig A. Berry wrote:
> There's more to this than I thought because it's not just that the
> new PREFIX functionality doesn't work.  All the INSTALL* and *VENDOR*
> macros are broken.  There will be no installing of extensions on VMS
> using MM 5.90 until this is resolved.

Ooops.  I didn't mean to leave it in nearly that bad a shape.


> So far I've added a boatload of new %Config variables and have
> managed to silence some, but not all, of the uninitialized value
> warnings in the installation tests.  I've added usevendorprefix and
> set its value to C<''> as Michael suggested, but AFAICT nothing
> checks this and all the vendor-related stuff is expected to be there
> even if I've requested that it not be used.

Damn, this is going to involve more special casing than I'd like.
I'll figure it out.


> There are a couple of oddities in addition to the test failures.
> Running a Makefile.PL seems to do everything twice:
> 
> $ perl Makefile.PL
> Checking if your kit is complete...
> Looks good
> Checking if your kit is complete...
> Looks good
> <snip>
> 
> In this generated descrip.mms (Makefile) excerpt, something is
> inserting a space before the colon that indicates the path is
> absolute:
> 
> PREFIX = perl_root :
> SITEPREFIX = perl_root :
> VENDORPREFIX = perl_root :

Odd.

> And the test failures I'm still seeing:
> 
> t/inst................
> #     Failed test (t/inst_prefix.t at line 88)
> #                   '[.foo.bar.lib.perl5.5_7_3.VMS_AXP]'
> #     doesn't match '(?-xism:^\$\(PREFIX\))'

*snip*

These results are basically correct, just not the same as Unix.
Normally the result would just be:

    $(PREFIX)/lib/perl5/5.7.3/

or something, but of course you can't do that so I assume you
pre-expanded the prefix.  I'll adjust the tests.



-- 

Michael G. Schwern   <[EMAIL PROTECTED]>    http://www.pobox.com/~schwern/
Perl Quality Assurance      <[EMAIL PROTECTED]>         Kwalitee Is Job One
GOD made us funky!

Reply via email to