* Make each test file runnable using `module+'.
* Make the space tests not print anything (unless they fail, of course).
* Make the `first-class?' case even "more first class" by defining their
and/or as functions. (Doesn't make a real difference for this test,
but nice to test more of the language.)
The `or' and `and' functions already worked correctly; this
fix is for `or' and `and' as applied directly to arguments, which
expands to the `racket/base' `or' and `and' forms.
* A long-standing bug, which happened with
(let ([x (lazy (delay 1))]) (force x) (force x))
not being properly handled.
* Added `delay/strict', mostly for the below.
* Made srfi/45 reprovide it as `eager'.
* Also restricted the exports from srfi/45 to its interface.
All of these issues were reported by Andreas Rottmann.
- `!!' now scans the same kind of data that `make-reader-graph'
handles (except that hash-tables are not implemented)
- this means no structs, no mpairs, and a bunch of other stuff
- `!!!' is gone (lazy procedures are not wrapped)
- dealing with multiple values moved into lazy/lazy.ss itself (and
in the future everything will move in there)
* Removed lazy/promise, and use scheme/promise instead. Also remove
the docs for lazy/promise that were bogus (since scheme/promise *is*
doing the same thing now).
* Other adjustments to the docs. They should be considered incomplete
now, and will need a major rewrite when the whole thing works again
(multiple values things are just commented out for now).
* Added a test macro and a quick test suite for lazy/promise.
* The lazy tests are added to the nightly build tests
svn: r11042
