Wed Mar 19 16:03:51 1997  Stephen Peters  <speters@cygnus.com>

	* kdb_oci.c: Fixes to allow renaming from backup databases.

	* mk_kerbnet.sql: New file for creating Oracle database.

Wed Mar  5 15:51:57 1997  Stephen Peters  <speters@cygnus.com>

	* kdb_oci.c (krb5_oci_db_set_name): Fixed to allow different
	  locations for the database.

Sat Mar  1 00:19:51 1997  Marc Horowitz  <marc@cygnus.com>

	* configure.in: removed USE_KDB5_LIBRARY, because it caused a
 	circular dependency building the library. Added USE_DB_LIBRARY and
 	USE_OCI_LIBRARY instead.

Tue Feb 25 11:45:01 1997  Stephen Peters  <speters@cygnus.com>

	* kdb_oci.c: Move Oracle-specific info from static areas to
		  per-context buffers.

	* t_kdb.c (do_testing): Ignore a lack of error from Oracle DB
		  set_name function.

	* configure.in: Created KDB_FILE substitution.  Added
		  USE_KDB5_LIBRARY for t_kdb test program.

	* Makefile.in: Use @KDB_FILE@ instead of GNU make's ifneq().
		  Removed -ldb for t_kdb test program.
		  
Fri Feb 14 10:17:58 1997  Stephen Peters  <speters@cygnus.com>

	* t_kdb.c: Added entries to end of both kdb5_dispatch_tables.

	* kdb_dbm.c (krb5_dbm_db_bounded_iter): No longer a stub function,
		  it now actually works.

Wed Feb 12 18:55:36 1997  Stephen Peters  <speters@cygnus.com>

	* kdb_oci.c (krb5_oci_err_report): Use generic logging interface.
		  Also added #include for oracle.h.

Tue Feb 11 10:10:39 1997  Stephen Peters  <speters@cygnus.com>

	* keytab.c: Changed to use krb5_db_... instead of krb5_db_dbm_...

	* kdb_dbm.c (krb5_dbm_db_bounded_iter): Stub function for bounded
		  iterators. 
	
	* kdb_oci.c: New file for handling Oracle database functions.

Fri Jan 31 11:58:49 1997  Jeremy Allison <jra@cygnus.com>

	* kdb_dbm.c: Added O_BINARY to lock file opens. Added
	__CYGWIN32__ specific code to close the main database
	before doing a rename and re-open it afterwards. Needed
	as NT cannot rename an open file.

Tue Dec  3 21:15:41 1996  Mark Eichin  <eichin@cygnus.com>

	* t_kdb.c (swatch_on, swatch_eltime): use krb5_gettimeofday.

Thu Oct 17 20:15:10 1996  Mark Eichin  <eichin@cygnus.com>

	* kdb_dbm.c: ditch openclose conditionals, just fix locking.
	Revert krb5_dbm_db_init_real and krb5_dbm_db_init_fast, just let
	the callers add extra locking themselves.
 	(krb5_dbm_db_unlock): only close on zero refcount.

Tue Oct  8 23:16:49 1996  Mark W. Eichin  <eichin@cygnus.com>

	* kdb_dbm.c (krb5_dbm_db_init, krb5_dbm_db_init_fast): new
 	interfaces that select between open-close-mode with per-operation
 	locking and single-open-mode with overall locking.
	(krb5_dbm_db_init_real): store argument in db_openclose member of
 	context; change DB_OPENCLOSE code to runtime conditional.
	(krb5_dbm_db_fini): in fast mode, unlock database on close.
	(krb5_dbm_db_lock, krb5_dbm_db_unlock): change DB_OPENCLOSE code.
	(krb5_dbm_db_put_principal, krb5_dbm_db_delete_principal): change
 	DB_OPENCLOSE code. Note that previously it was backwards! With
 	DB_OPENCLOSE *not* defined, it would always close and reopen the
 	database.

Tue Oct  8 23:16:49 1996  Mark W. Eichin  <eichin@cygnus.com>

	(krb5_dbm_db_lock, krb5_dbm_db_unlock, krb5_dbm_db_put_principal,
 	krb5_dbm_db_delete_principal): clear db_dbm_ctx properly on
 	KDBM_CLOSE error.

Tue Oct  8 23:16:49 1996  Mark W. Eichin  <eichin@cygnus.com>

	* kdb_dbm.c (krb5_dbm_db_rename): grab errno from rename *before*
 	calling krb5_dbm_db_end_update, to avoid "not a typewriter"
 	syndrome.
	(krb5_dbm_db_put_principal, krb5_dbm_db_delete_principal):
 	KDBM_STORE can fail (in case of database corruption, for example)
 	*without* causing errno to be set.  If errno is zero, use
 	KRB5_KDB_DB_CORRUPT instead. (If it is non-zero, it may still be
 	wrong, but at least something gets reported.)

Wed Sep  4 17:11:05 1996  Ken Raeburn  <raeburn@cygnus.com>

	* Makefile.in (DEPLIBS): Don't define at all.

Sat Aug 31 18:00:31 1996  Ken Raeburn  <raeburn@cygnus.com>

	* Makefile.in (DEPLIBS): Use static suffixes, in case shared
	versions aren't being built.

Sat Aug 17 22:57:45 1996  Ken Raeburn  <raeburn@cygnus.com>

	* keytab.c (krb5_ktkdb_get_entry): Pass 0 as second argument to
	krb5_dbm_db_get_mkey; the previous value was the wrong type, and
	it doesn't get used by krb5_dbm_db_get_mkey anyways.

Tue Apr 30 18:00:08 1996  Chris Provenzano  <proven@cygnus.com>

	* kdb_cpw.c(add_key_pwd()): If salt type = KRB5_KDB_SALTTYPE_AFS3
	then pass salt length as -1 to krb5_string_to_key().
	Then reset the salt length to its proper value before
	calling krb5_dbekd_encrypt_key_data().

