Test-Deep as_test_deep_cmp branch

--LQksG6bCIzRHxTLp
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

I've filed a PR against Test-Deep:

  https://github.com/rjbs/Test-Deep/pull/52

  addresses

  https://github.com/rjbs/Test-Deep/issues/50

The idea is:

When you call cmp_deeply($have, $want), objects found in $want are either
Test::Deep::Cmp objects or not.  If they are, they do their special test ma=
gic.
If they are not, they are more or less compared to the data in $have without
respect to blessing.  For example, a blessed hashref in $want is compared to
the matching data in $have for keys and values.

With my changes, objects in $want are checked for a as_test_deep_cmp method,
which can be called to return a Test::Deep::Cmp to be used instead of the
object.  I expect this to be rarely, but sometimes, useful.  Issue #50 goes
into the details of why it came up.  In that, I mention using this with
JSON::Typist/Test::Deep::JType.  I have implemented the described feature at
https://github.com/rjbs/JSON-Typist/pull/2

I'm interested in any thoughts on this before I make a trial release.

--=20
rjbs

--LQksG6bCIzRHxTLp
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJYVphXAAoJEOYby6cMccU5W/wH/3eYWgmPZ0l9rX8Wo/gLpwQz
X+fCEOZ8NSxEtSHJq6ws9pAsqn2wZ+U8xPVy8u/U4KYnjOthHHwCAxTBER3JFQJp
gyMwVCNkPyAhE6Kec3Ebn/o+l1KC1LJeA9xPAL8cb+DKEQdL7+i3sLPh2nXD9J5U
KQoQBXEyWks2HvUcarNRdjIZRYltIVz0I5Id8JfwNFQcb4HNm4L8UKE5i52Q+N5f
15p6oNh5yhOPoDxeUYUNDeVNpfghZbeXRuIsYik54W2sMFbZSiyIQSdxVim74iUq
peD9VtLerO2B7dRlRpyYxq0mRgJhuVCCKhul2KIh0x/Hc9hufnTJdb9i4Y0DTSI=
=Kadw
-----END PGP SIGNATURE-----

--LQksG6bCIzRHxTLp--