Don't assume that non-simple strings always have fill-pointers when  serializing. Thanks to pixel@kepibu.org.

Fix a bug in the creation of transaction-ids (bug reported by Klaus Harbo).

Don't use wildcards but delete only the four rucksack files when
opening an existing Rucksack in :SUPERSEDE mode (bug reported
by Volkan YAZICI).

EQUAL argument wasn't supplied.

by persistent btrees for non-unique slot indexes.

and using a lazy-cache.  Fixed some small bugs.  Added a few handy functions and
macros.

In detail:

Added P-PUSH and P-POP.

Improved btree efficiency by switching to a different data structure
for the bindings.  Instead of using a persistent cons for each key/
value pair, we now put the keys and values directly into the bnode
vector.  This speeds up most btree operations because it reduces
persistent consing when adding new values and it reduces indirections
when searching for keys.

Renamed BTREE-NODE to BNODE, BTREE-NODE-INDEX to BNODE-BINDINGS,
BTREE-NODE-INDEX-COUNT to BNODE-NR-BINDINGS, FIND-BINDING-IN-NODE to
FIND-KEY-IN-NODE.

Fix a missing argument bug in REMOVE-CLASS-INDEX.

Added a LAZY-CACHE which just clears the entire hash table whenever
the cache gets full.  This improves memory usage, because the normal
cache queue kept track of a lot of objects that for some reason
couldn't be cleaned up by the implementation's garbage collector.

Added the convenience macros RUCKSACK-DO-CLASS and RUCKSACK-DO-SLOT.

Made RUCKSACK-DELETE-OBJECT an exported symbol of the RUCKSACK
package.

Fix a bug in TEST-NON-UNIQUE-BTREE: it should call
CHECK-NON-UNIQUE-CONTENTS instead of CHECK-CONTENTS.

Fixed a garbage collector bug reported by Sean Ross. When the garbage
collector deletes object ids from the object table (because the
objects are dead and we may want to reuse their ids later for other
objects), it should also remove that object from the cache.  If it
doesn't, there's a possibility that the object id will be reused later
for a new object and the cache wil still refer to the old in-memory
object.

Class and slot indexes now map directly to objects instead of
object-ids.  This fixes a bug where the garbage collector forgot
to add all indexed objects to the root set. (Suggested by Sean
Ross.)

Increased default cache size to 100,000 objects.

to fix a compiler warning reported by Brad Beveridge.

causes a transaction rollback.

Add Brad Beveridge's basic unit test suite (modified to work
with lisp-unit instead of 5am).

Add Chris Riesbeck's lisp-unit library to help with creating unit test suites.

Use (ARRAY-DIMENSION buffer 0) instead of LENGTH in
LOAD-BUFFER, because we want to ignore the fill pointer
here.  Thanks to Sean Ross.

- Fix some typos (:VALUE should be :VALUE=) in index.lisp.
- Version 0.1.11.

the position of the key.  Reported and fixed by Leonid Novikov.

Abe, not by Henrik Hjelte.

from a leaf, we should update the parents so they'll use the key that
has now become the biggest.  (Henrik Hjelte.)

Try to signal an error when an incompatible value is given to indexed
slots, e.g. trying to put a string into a slot with a :symbol-index.
(Henrik Hjelte)

Signal an error during when putting duplicate values into a slot for
which duplicate values are not allowed.  (Henrik Hjelte)

Use BTREE-VALUE-TYPE, not BTREE-KEY-TYPE, when type checking a value
during BTREE-INSERT.  (Henrik Hjelte)

Wrap COMPILE-FILE calls in a WITH-COMPILATION-UNIT to prevent
superfluous warnings about undefined functions.

Add RUCKSACK-DELETE-OBJECT, RUCKSACK-DELETE-OBJECTS and
RUCKSACK-ROOT-P (suggested by Henrik Hjelte).  I haven't
tested these functions yet.

works on SBCL.  (Thanks to Levente Mészáros.)

(thanks to Attila Lendvai).

 * add new parameter *collect-garbage-on-commit*
 * add (:inhibit-gc nil) keyword arg to with-transaction
 * add without-rucksack-gcing macro
 * only collect garbage on transaction-commit when
   *collect-garbage-on-commit* is not nil

 * use binary search instead of linear search in find-subnode and
   find-binding-in-node

 * btree-max-node-size now defaults to 32 instead of 100

 * propogate unique fropm the direct slot-definition(s) to the
   effective slot definition

The :EQUAL parameter for MAP-INDEX-DATA wasn't handled correctly
for indexes with non-unique keys (reported by Cyrus Harmon).

Take care of some differences between the MOP implementations of Lispworks
and SBCL.  Lispworks doesn call (SETF SLOT-VALUE-USING-CLASS) in
SHARED-INITIALIZE, but SBCL does.  Lispworks calls FINALIZE-INHERITANCE
after a class is redefined and a new instance is created, but SBCL doesn't.
All tests now work for Lispworks (5.0) and SBCL (0.9.16).

Some work on a copying GC.

Handle updates of in-memory persistent objects by writing a method for
Lispś UPDATE-INSTANCE-FOR-REDEFINED-CLASS that marks the object as dirty
and calls Rucksack's UPDATE-PERSISTENT-INSTANCE-FOR-REDEFINED-CLASS.

Get rid of the Lispworks specific PROCESS-A-SLOT-OPTION stuff and handle
the slot options in a way that's compatible with AMOP.

Remove INITARGS argument for UPDATE-PERSISTENT-INSTANCE-FOR-REDEFINED-CLASS,
because it turns out to be unnecessary (see details in notes.txt)

Add explanation to test-index-1a.lisp about the use of
  (eval-when (:compile-toplevel :load-toplevel :execute) ...)

Replace *RUCKSACK* by RS in the test-*.lisp files.

the :INDEX class option in a way that's compatible with the AMOP.

Add list of things to do, list of things done, and some random notes.

Write test case for slots with redefined indexes.  This also tests
the default method for UPDATE-PERSISTENT-INSTANCE-FOR-REDEFINED-CLASS.