New M::B beta on CPAN

The uploaded file

     Module-Build-0.2808_01.tar.gz

has entered CPAN as

   file:
$CPAN/authors/id/K/KW/KWILLIAMS/Module-Build-0.2808_01.tar.g
z
   size: 203445 bytes
    md5: d272b213e748a9c70990df37b7e8cee4

I posted this for some formal beta testing, with an eye
toward  
synchronizing with the code freeze process for perl 5.10. 
As of  
right now, M::B in blead is a mishmash of 0.2808 and some
patches  
from various people.  I would like 0.2809 to be what ships
with 5.10  
if possible so that it coordinates well with the CPAN
release.

If there are any important bugs we need to fix before
calling this  
0.2809, please inform.

As luck would have it, I just had my appendix removed, so
I've got a  
couple recovery days where I can do patches etc.

There are plenty of changes since version 0.2808, see
below.

  -Ken


0.2808_01 - Wed Oct 24 11:51:25 2007

  - All .pm files in the Module-Build distribution (except
for
    M::B::Version.pm, which is kind of tied to version.pm)
now have the
    same $VERSION number explicitly specified.

  - When checking prerequisites, the required version of
perl is now
    checked before we start finding the $VERSION declaration
of the
    distribution, which results in much more intuitive error
messages
    e.g. if the author is using 5.6-isms in their
declaration but the
    user doesn't have 5.6. [Slaven Rezic]

  - Added 'artistic_2' license, corrected 'lgpl' license url
(bug  
#29783)
    [David Thomas]

  - VMS find_perl_interpreter() is just $^X (bug #29810)
[Craig A.  
Berry]

  - Some large VMS fixes, mostly having to do with the
non-case-
    preserving nature of most VMS filesystems, but also
correcting for
    illegal characters in VMS file specs. [John E. Malmberg
and Craig
    A. Berry]

  - Fixed the _detildefy() method on VMS. [John E. Malmberg
and Craig
    A. Berry]

  - We now use a much more reliable method when Data:umper-in
g saved
    state data. [Yves]

  - When a module had 0.000 as its version, a few places in
the code
    thought the module had no version at all.  This is now
    fixed. [Andrew "Zefram" Main]

  - When finding the default set of script_files, we now
compute them
    as relative paths, not absolute. [Spotted by Curtis
"Ovid" Poe]

  - Got rid of a call to eliminate_macros, which isn't
needed in
    Module::Build since there is no external make utility
involved.
    Override expand_test_dir to make up for the fact that
the
    home-grown glob() returns absolute, not relative, paths.
[Craig
    A. Berry]

  - Fixed a catdir() that needed to be catfile() in the
.packlist
    creation code. [John E. Malmberg]

  - If a *.PL file ended abnormally during the build,
processing should
    have stopped, but it didn't.  Fixed. [Matthew Cast and
David
    Golden]

  - Module::Build::Compat adds "require 5.XXXXX"
to Makefile.PL when
    'perl' is specified as a 'requires' prerequisite [David
Golden]

  - Refactored t/compat.t for modularity and transparency;
added
    labels for all tests; supressed subprocess output to
    STDOUT and STDERR [David Golden]

  - Fixed bug in perl_version_to_float when version is
already a float
    [David Golden]

  - Removed a mention of $build-> from the
documentation, the
    official interface to Config.pm settings is now via the
    $build->config() and has been for some time.
[Suggested by Michael
    Schwern]

  - Tweaked some text in the Cookbook to bring it into the
modern age,
    and added a recipe for accessing Config.pm settings.
[Ibid]

  - Lots of POD link/readability improvements to the
Module::Build::API
    documentation [Salve J. Nilsen]

  - Added configure_requires as a new type of prereq. 
[Suggested by  
Adam
    Kennedy]

  - Patch 31156 from bleadperl: some filename dot and
extension help
    for Module::Build on VMS. [Craig Berry]

  - Reworked the _detildefy() method so it doesn't depend on
glob()
    anymore.  This gets rid of a problem with spaces or
other special
    shell characters in things like 'prefix' or
'install_path'
    entries. [Prodding by Eric Wilhelm]

  - Added midnightbsd to the list of Unix-like OSes we know
about
    [Rafael Garcia-Suarez]

Thomas Klausner wrote:
> I spent some time on this issue this morning, see my
latest reply/patch 
> in RT. Maybe the problem actually lies in version.pm
(but I didn't look 
> into this yet)

Yeah, that does seem to be my problem (version.pm) rather
than Ken's. 
I'll take a look tonight...

John

p.s. I'm also getting ready to bump version to 1.0 when Perl
v5.10.0 
ships, so I'll be producing a patch to M::B to update the
embedded 
version.pm code as well...