Thu Feb  6 02:55:19 1997  Erez "HWank1" Zadok  <ezk@shekel.mcl.cs.columbia.edu>

	* amd/info_ndbm.c (ndbm_search,ndbm_init): fixed the code so that
	it works on freebsd, where they map dbm_pagfno a non-existing
	macro on purpose, to force you to port the code to Berkeley DB.

Fri Jan 31 01:35:09 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/ops_afs.c (mount_toplvl): added a sleep(1) right before the
	actual mount_toplvl, to avoid some possible race conditions.

Thu Jan 30 02:31:49 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/rpc_fwd.c: removed all TLI (HAVE_T_OPEN) code.

Wed Jan 29 18:25:07 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* lib/wire.c (getwire): removed HAVE_T_OPEN (TLI code)

	* amd/amd.c (main): fix local address using htonl(). Free up
	buffers allocated by netdir_*() functions.

Wed Jan 22 23:53:24 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/fix-amd-map.in: new script that will convert old-style amd
	maps to new one.

Tue Jan 21 01:18:16 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* FIRST AMD IN AM-UTILS FINALLY WORKS (mounts itself as toplvl,
	and replies back to simple nfs/link type mount queries).

	* conf/mtab/mtab_svr4.c: reorganized the code in this file to be
	easier to read, and fixed a few bugs with un/setting the lock file
	descriptor.

Wed Jan  8 15:27:49 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/sched.c: removed HAS_SVR3_SIGNALS code to handle older
	signals in Svr3 systems.  Only stellix ever used this.

	* amd/rpc_fwd.c (fwd_reply): removed all code that was triggered
	by DYNAMIC_BUFFERS, since it was not in use.

Tue Jan  7 00:45:59 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/opts.c (eval_opts): removed AMD_COMPAT code.  Amd no longer
	takes '=' sign as old selector.  Fix your maps.
	(deslashify): removed APOLLO code that treated two '//' as one.

Sun Jan  5 01:25:09 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/ops_nfsx.c (nfsx_init): removed HARD_NFSX_ERRORS, code which
	was not used.  It made any submount error fail the whole group.

	* amd/ops_nfs.c (make_nfs_auth): removed code dependent on
	HAS_NFS_QUALIFIED_NAMES.
	(mount_nfs_fh): removed macro for NFS_ARGS_NEEDS_PATH, assuming it
	is never on.  So code that was triggered by #ifndef
	NFS_ARGS_NEEDS_PATH (osf1, aix3, and bsd44 variants) may have to
	be fixed later.
	Removed all ULTRIX_HACK code.  Ultrix is DEAD!
	removed all PRESET_AC code, which was only on for Linux to preset
	the attribute cache values (otherwise they default to zero).
	(nfs_umounted): revobed KICK_KERNEL code (IRIX systems).  Will
	have to fixed up later.
	(mount_nfs_fh): removed code that was never included, which
	attempted to set the port on which amd's nfs mounts on.  This
	isn't supported by the ping algorithm yet.  In any case, it is all
	done in nfs_init().

	* amd/srvr_nfs.c (find_nfs_srvr), amd/ops_nfs.c (mount_nfs_fh):
	removed HAS_TCP_NFS code.  The whole use of NFS V.3 with TCP or
	specially hacked NFS V.2 that uses TCP (BSD 4.4 variants) will
	have to be addressed later.

	* amd/nfs_subr.c (nfsproc_getattr_2): removed all code dependent
	on PRECISE_SYMLINKS.

	* amd/mapc.c (mapc_meta_search): removed code that depended on
	HAS_DEPOT.  It is known buggy code.  The whole DEPOT support needs
	to be rethought.
	(mapc_add_kv): ported code to use standard regular expression
	(regexp) library.

	* amd/{map.c,ops_sfs.c,ops_ufs.c}: removed all code dependent on
	FLUSH_KERNEL_NAME_CACHE and the macro itself.

Sat Jan  4 15:46:58 1997  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/map.c (exported_ap_alloc): removed all code triggered by
	SHRINK_EXPORTED_AP and the macro itself.

	* amd/info_nis.c: removed HAS_NIS_RELOAD macro, and left the code
	in all the time.  Assuming that there is always NIS realod
	available.
	(yp_all_fixed): removed code for BROKEN_YP_ALL.  If your yp_all()
	function is broken such that it does not release filedescriptors
	it created, then you are hosed.  Get a fix for yp_all from your
	vendor or switch operating systems.

Tue Dec 31 15:34:36 1996  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/info_hesiod.c: renamed macro HAS_HESIOD_RELOAD to
	HAVE_HS_ZONE_TRANSFER.

Mon Dec 30 17:33:23 1996  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/{nfs,host}_ops.cremoved INFORM_MOUNTD macro, so its code is
	always included.  We should always try to inform the remove mount
	daemon of mounts we performed.  If we want an option for this, it
	should be a run-time option.

	* amd/host_ops.c: removed all that optionally did not make the
	directory mount points for amd.

	* HOST_EXEC code removed from everywhere.  There is no indication
	that it was used anywhere.  If it needs to be used, then a
	run-time switch should control it.

Thu Dec 26 00:06:52 1996  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amd/*_ops.c: removed all references to mntent_t's fields
	mnt_freq and mnt_passno, since they are no longe rused (and never
	have been).

	* amd/am_ops.c: removed function sunos4_match, which was off
	anyway.  This was one of the places where an '=' sign was allowed
	as an assignment operator (which is now ':=').

	* amd/afs_ops.c (afs_bgmount): removed SUNOS4_COMPAT code which
	was off anyway.  This was one of the places where an '=' sign was
	allowed as an assignment operator (which is now ':=').

Tue Dec 24 01:53:52 1996  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* amq/amq.c: ported to am-utils.  Working.

	* conf/mount/mount_svr4.c (mount_svr4): changed mount type on svr4
	to be string, as it should be (rather than a string mapped to int
	and back mapped to string).

Sun Dec 22 13:26:05 1996  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* lib/mount_fs.c (mount_fs): removed MNTINFO_PREF code.

	* lib/umount_fs.c: split this into three files: default, osf, and
	bsd44, that are in conf/umount/umount_*.c.

Thu Dec 19 17:33:46 1996  Erez "HWank1" Zadok  <ezk@lorien.cs.columbia.edu>

	* aux/configure.in, aux/aclocal.m4, aux/acconfig.h: added new
	macro to automatically figure out if an external definition for
	sys_errlist[] exists.

	* lib/xutil.c (xfree): removed "#undef free" because it's not
	needed there.  Any system that redefines free() needs help.

	* conf/mtab/mtab_file.c (REWRITE_MTAB_IN_PLACE): removed code that
	was dependent on REWRITE_MTAB_IN_PLACE.  It does not appear to be
	in active use anywhere.  If added later, it should be a run-time
	configuration option.

	* conf/mtab/mtab_mach3.c: removed code that was #ifdef'ed not to
	be MOUNT_TABLE_ON_FILE, since mach2 always stores its mount tables
	in a file in /etc.

	* lib/util.c (str3cat): removed _AIX code.  AIX should have a
	working realloc, or else find a cleaner solution to this.
	(mkdirs): removed SUNOS4_WORKAROUND code.  This was to do a sync()
	after immediately making directories so that fsck will be able to
	fix the filesystem in the event of an immediate crash.  This was
	because a bug in UFS which is fixed by now.

	* lib/hutil.c (domain_strip): removed PARTIAL_DOMAINS triggered
	code.  Does not appear to be on and used anywhere.
	(dofork): removed MACH3 code, b/c mach3 NFS bugs are fixed by now.

	* lib/resvport.c: split the binding of reserved ports code to a
	separate file.  A #define (HAVE_T_OPEN) separates the TLI (SVR4)
	from the BSD code.

Thu Dec 19 02:08:35 1996  Erez "HWank1" Zadok  <ezk@shekel.mcl.cs.columbia.edu>

	* lib/Makefile.am (amu_LIBADD): includs @LIBOBJS@ now, auto-filled
	in by automake and configure, for optional sources that need to be
	built as part of libamu.a.

	* aux/aclocal.m4: Make sure AC_CHECK_MOUNT_STYLE adds mountutil.o
	to LIBOBJS.

	* lib/memcmp.c (memcmp): new file added to replace a possible bad
	implementation of memcmp.

Wed Dec 18 22:20:23 1996  Erez "HWank1" Zadok  <ezk@shekel.mcl.cs.columbia.edu>

	* converted the sources to use memset/memmove/memcmp instead of
	bzero/becopy/bcmp.

	* ran all sources through gnu Indent w/ options specified in
	aux/amindent.

Wed Dec 11 22:19:29 1996  Erez "HWank1" Zadok  <ezk@shekel.mcl.cs.columbia.edu>

	* At this point the initial genration of config.h via autoconf and
	some preliminary makefiles via automake is done.  I will not begin
	modifying sources, so it's a good time to document from here.

