Tests: SBCL passes test-lisp and test-upgrade
Also:
* big refactoring of test infrastructure
* contrib/debug.lisp and (asdf/driver:asdf-debug) for debugging.
* remove aif and it, use the cleaner if-bind.
* backtrace support, with improvements from trivial-backtrace.
* integration of asdf-condition-control (originally from xcvb-driver)
* simplify the upgrade heuristic: always tries to upgrade once.
  The previous heuristic was too complex with too many failure cases.
  You are thus assumed to want a given version asdf
  if and only if it is in your source-registry.

* Get rid of *stamp*, use lexical scoping everywhere
* Better handle non-filesystem actions in compute-action-stamp
* Have a simpler mechanism for backward compatibility with operation-done-p
* Have a more robust make-sub-operation, and use it where appropriate
* Test twiddling.

* Refactor load-system to use *load-system-operation*,
and change require-system and module-provide-asdf to use that.
* Remove broken-fasl-support: first, cormanlisp is supposed to have fixed it;
 second, cormanlisp is dead;
 third, one can now (setf *load-system-operation* 'load-source-op) instead.
* Reindent default-source-registry
* Fix a test in light of some previous package frobbing.

The issue fixed is that most Lisps don't like
*default-pathname-defaults* being a logical-pathname.
Added a test for logical-pathname functionality;
it all works on CCL, and works better on other Lisps after the fix,
but there are still logical-pathname issues on SBCL, CLISP and more.
Also, a package cleanup of our test suite.

Use probe-file for probe-file* in LispWorks.

Make ABCL's jar translation happier with
the recent compile-file-pathname* change for Stelian by redoing cfp myself.

Make truenamize work when the merged pathname is relative.
Makes it happier when e.g. the *default-pathname-defaults* is #p"",
which is the case notably on CCL.

Try and fail to support GCL 2.6 and have 2.7 fail less;
too many bugs, in CONDITIONs, PATHNAMEs, and more,
plus compiler bug that breaks coerce-pathname.

Also, use coerce-pathname in asdf-ecl and deprecate merge-component-name-type.

Started as an attempt to "enhance" it,
until thinking about why I was breaking tests told me I couldn't.

Simplify test scripts.

There were bugs in the tests having to do with repeatedly calling
quit-on-error in a single test script.  Only one of those calls
will be evaluated, and then the script will exit.  That means that
multiple calls to quit-on-error masked failures in the tests.

Also fixed a package-related error in test1.script that I conjecture
led to the mistaken use of multiple calls to quit-on-error.

perl -pi.bak -e 's|"../asdf"|\(compile-file-pathname "../asdf"\)|;' *.script

Found that the current load recipe (load "../asdf") in the test
scripts was brittle, because the presence of asdf.fasl was obscured
if there was an "asdf/" directory in ..  Unfortunately, the result
of makeinfo --html asdf.texinfo created just such a directory!
fe[nl]ix kindly provided a pointer to compile-file-pathname that
solved the problem.

Added &allow-other-keys to operate. Added test/test-force to test that forcing a system works (at least in a simple case).