[DBIx-Class-Devel] Busted Releasing on Windows

[Peter Rabbitson]

On Thu, Dec 13, 2012 at 11:57:58AM -0500, Graham Knop wrote:
> My suspicion about the M::I brokenness being due to Windows isn't
> based on much other than "ugh windows".

It depends which brokenness you are referring to. The "dist with
partial inc/" thing has happened on linux as well[1]. I am still
not sure *how* they did it, but maybe my last overhaul of handling
of the MANIFEST in 374cd092 will make it go away.

> The exec call to actually do the POD injection could be fixed to work
> on Windows.  I'm also not sure how important the permissions issues
> are given how MakeMaker behaves.

The "in Make or die" is extremely important due to __FILE__ being
involved in the mix. I am contemplating replacing the entire assembly
with something else over the weekend (given that my distdir assembly
is now much more straightforward).

> But it still feels like it's not worth spending the effort to fix
> release time only issues for Windows.

SineSwiper seems bent on continuing to run in hell, and I like
the challenge, so I guess overall it's worth it. It also makes the
distbuilder more robust in the long run.

Cheers