1999-10-11  Ales Hvezda  <ahvezda@geda.seul.org>

	* scheme/*.scm: Removed all gnetlist:set-netlist-mode

	* src/g_register.c: Removed register call to gnetlist:set-netlist-mode

	* src/g_netlist.c: Removed gnetlist:set-netlist-mode

	* scheme/gnet-spice.scm: Integrated latest spice backend by Bas 
          Gieltjes

	* tests/: Created directory to hold gnetlist tests
	
	* tests/*: Populated tests directory with output from correct netlist
	  runs

1999-10-10  Ales Hvezda  <ahvezda@geda.seul.org>

	* gnetlist/gnetlist.c: Fixed up s_log_init to use the new prototype

1999-10-09  Ales Hvezda  <ahvezda@geda.seul.org>

	* examples/test.sch: Renamed (replaced) test_verilog.sch

	* examples/test_verilog.sch: Translated it away from borders

	* examples/7447.sch: Translated it away from borders

1999-10-06  Ales Hvezda  <ahvezda@geda.seul.org>

	* scheme/gnetlist.scm: Fixed a dangling reference to bom:strip1 
	  (changed to strip1)

1999-10-06  Matthew Ettus   <matt@ettus.com>

	* scheme/gnet-vipec.scm:  New file, not functional yet.

	* scheme/gnet-PCB.scm:  Cleaned up functions, moved common ones to
	main gnetlist.scm file. 

	* scheme/gnet-allegro.scm:  Clean up, remove dead code

	* scheme/gnet-bom.scm:  Clean up, remove dead code

1999-10-04  Magnus Danielson  <cfmd@swipnet.se>

	* scheme/gnet-vhdl.scm: Finally fixed the write-port-clause stuff.
	This is now splitted into three routines which forms a much better
	basis for further development and also fixes the bugs which occurs if
	any or more of the in, out and inout sublists happends to be empty.
	Stuck in support for signal type.

1999-10-03  Ales Hvezda  <ahvezda@geda.seul.org>

        * In all *.c included define.h before struct.h

1999-09-28  Magnus Danielson  <cfmd@swipnet.se>

	* scheme/gnet-vhdl.scm: Fixed bug with no ports to entity.

9/25/99	Integrated Magnus's latest VHDL backend .scm code

9/23/99	Split up gnetlist.scm into seperate files.  The basic functions will
	live gnetlist.scm while the individual backends will be in gnet-*.scm
	
	Made some changes to the code so that the right backend *.scm code
	is loaded when appropriate

	Integrated Magnus Danielson's VHDL backend

9/19/99	Make sure gnetlist runs with guile-1.3.2a
	
	Fixed all warnings

9/17/99	Fixed that really nasty bug/problem that dealt with loading up 
	multiple sheets (see below).  The postprocess was happening too
	early and was for some odd reason nuking previously post processed
	nets.

9/16/99	Finally integrated Bas Gieltjes's latest spice netlist (sorry about
	the delay)  (this basically consisted of adding the code to 
	gnetlist.scm)  I'm not 100% this still works with the current 
	gnetlist, especially the new net= attribute

	Added code in gnetlist.c to only load up schematics which haven't
	been loaded already

	Found a NASTY bug/problem when you try to load up more than one
	schematic sheet with gnetlist.  Added a line in s_traverse_start
	which should have helped, but problem/bug still exists

9/14/99 Integrated Matt Ettus's Allegro and PCB gnetlist backends.  Thanks!

	Updated man page to reflect the backends

9/12/99	Fixed a netlist print function to that it doesn't try to print
	valid null strings

	Fixed up s_net_name_search so that it tells the user if a multiple 
	named net is found

	Implemented the concept of net name priority (net_name_has_priority)
	which controls which name has the highest priority.

	Added net-naming-priority keyword which controls which attribute
	(either net= or label= ) has priority when both are attached to a net

	Added i_vars.c which containes the rc mode init funcs which are 
	used by Kazu's rc mode functions

	More work on getting net= attribute to override previously defined
	pins (all controlled through net-naming-priority)

	Removed some obsolete #if 0'd out code from s_net.c

	Added s_rename.c which contains all the code to handle the renaming
	(aliasing) of net names.

	Integrated s_rename functions into the code

	Fixed a bug in s_netattrib_net_search where the outside of the 
	component wasn't being search for overriding net= attributes

	Fixed s_netlist_post_process, (the net naming), to only name
	nets of components which have a uref.  Components which don't
	have a uref are power/gnd/special components

	Updated verbose legend to include the R (rename of nets)

	Added gnetlist:get-renamed-nets which takes a dummy parameter and
	returns a list of lists (which contain src -> destination net 
	renaming pairs)

	Updated the geda netlist backend to output the above renaming
	list pairs into the netlist file as a new section

9/11/99	Created s_netattrib.c which contains all the code to handle the net=
	attribute

9/7/99	Start work on getting the net= attribute to work 

9/5/99	Moved test_verilog.sch into the examples directory (from 
	gschem/examples)

9/3/99	Fixed a nasty stupid bug in the nethash datastructure creation,
	used a constant instead of the pound define

8/31/99	Renamed ntext to text (finally)
	
	Fixed g_rc.c to work on WinNT

8/30/99	Renamed all ales to conn and ALES to CONN

8/27/99	Set the program version to 19990829 to force Ales to release
	on the 29th :)
	
	Updated all *.c and *.h files to have the right address for the
	FSF in the copyright/licence header

8/19/99 Updated Matt Ettus's BOM backend with thi's formatting changes

8/17/99	Added some bus related globals.

8/11/99	Added the real code to implement gnetlist:get-toplevel-attribute

	Integrated Matt Ettus's bill of material code (cool stuff!)

8/8/99 	Moved mk_verilog_syms to utils (created utils and Makefile.am)

	Minor updates to example schematics since I renamed some of the
	verilog symbols to have the -1 prefix.

8/5/99	Added some stuff to configure.in for win32 builds

8/3/99	Updated libtool to 1.3.3

	Integrated Mike Jarabek latest verilog backend
		- Patched g_netlist.c
		- Patched g_register.c
		- Patched gnetlist.scm
		- Added examples/sch/*.sch to examples
		- Added mk_verilog_syms to src
		- Added symbols to ../symbols

8/2/99	Added g_netlist_get_attribute, which allows you go find any toplevel
	attributes (not yet functional)

7/21/99 Removed all trace of s_passing.h

7/15/99	Added support for libgdgeda 1.6 (mainly means searching for libpng)

7/6/99  Changed configure and Makefiles to cut down on the number of 
	libraries which are linked

	Didn't really disable libgdgeda when using --disable-gdgeda
        Now fixed.

7/5/99 	Verified that gnetlist works with alphanumeric slotting

	Fixed all warnings 

	Renamed s_netlist_post_resolve to s_netlist_post_process

	Put Manu's configure.in changes into gnetlist as well
	
	Got configure.in to be almost identical to gnetlist/configure.in

	Updated man page slightly

7/4/99 	Integrated latest verilog backend from Mike 

	Changed Mike's all-nets function to use get-all-unique-nets 
	(equate same named nets together)

	Added some improved feedback when doing the net traversal with
	verbose mode on

7/3/99	Removed almost all executing traces of connected_to_[1|2] (these
	was replaced by connected_to)

	Added gnetlist:get-all-unique-nets which returns all the unique
	nets in the design

	Added gnetlist:get-all-connections which returns all connections
	which are associated with the argument netname 

	Worked on gnetlist.scm to use above functions.

	Got geda netlist format to use above functions correctly.

	Got tango netlist format to use above functions correctly.
	(all others are okay)

	Reformated the tango backend scheme code to be more scheme like
	Reformated the gEDA backend scheme code to be more scheme like

	Went through all the code removing old obsolete variables from
	the internal representation.

	As of the today gnetlisting works again :-)

7/2/99  Fixed the ./configure stuff dealing with libgdgeda so that it works
        when libgdgeda isn't installed in a system directory

7/1/99 	Got support for gdgeda in configure.in (required because you
	are linking against libgeda)

6/29/99 Got the core of nethash working with the rest of gnetlist

	Things seem to be working better now.  Core traversal of nets
	works now as well as the naming of nets.

6/25/99 Added support to configure to look for optarg and friends otherwise
        include the right externs...

6/17/99 Finally put all the changes in to make gnetlist work with the 
	new connection system 

6/8/99	Integrated old gnetlist into the current development dist scheme


--------------- new devel stuff starts here ----------------------------


5/14/99 Updated manual page a bit 

5/13/99 Integrated Mike Jarabek's verilog netlister 

4/11/99	Added some more debugging information when reading in the scm file

4/9/99	Fixed Makefile.am so that gnetlist.1 is installed

3/26/99 Added some newlines in a few places in gnetlist.scm

3/20/99 Got spice backend guile functions to follow the new 
	backend:function_name naming convention

3/19/99	Updated g_rc to print more diag message to the log file when
	it can't find an rc file (like gschem)

3/18/99 Updated g_rc to use new macros for finding system-*rc files

3/11/99 Added back the original implementation of the write-net-name-... 
	guile code for the spice netlist.  It works right for transistors.
	Mine did not.  Bummer.  Spice netlist doesn't support slotting for
	now.

3/10/99	Fixed a coredump when you pass in an "unknown" for the pin number
	in get-nets.  Return a list which is ("ERROR"). (pointed out by
	Bas Gieltjes)

	Changed ales-netlist format to geda-netlist format.  Changed
	gnetlist.scm appropriately	

	Renamed all C guile functions to be called gnetlist:<whatever>
	This makes it clear where functions come from 

	Added code to set netlist_mode which enables/disables certain
	netlist features.  Each backend should call this with something
	even if it's the default mode: gEDA

	Added gnetlist:set-netlist-mode to g_register and g_netlist which 
	sets the above mode

	Integrated tango netlist backend by Nuno Sucena (thanks)

	Changed the highest level function in all backends to take a 
	parameter which is the netlist output filename.  This is a required
	parameter

	Added -o flag which specifies output filename (for above), if you
	don't specify an output filename the default is output.net

	Integrated spice backend by Bas Gieltjes

	Had to add a kludge to gnetlist.c which sets netlist_mode based
	on the guile_proc string (for spice mode which needs to be set
	before the .sch files are read in)

	Changed the default unnamed_counter to be one instead of zero

	Fixed the incorrect behavior when you have slotted components and
	the spice backend.  

	Played a bit with the way power connections show up in the netlists
	(s_net.c and the POWER string code)  Not completely happy with 
	any of this.

3/9/99	Fixed a few variables which were uninitialized (width, height,
	snap_size)

1/24/99 Removed some // comments (shouldn't use those in C code)

12/3/98 Added thi's small gnetlist patch

11/8/98	Changed the ales-netlist slightly (less newlines)

	Started adding code to resolve same net names

	Found a really nasty bug, wasn't initalizing net_name in cpinlist!

	Found several really nasty bugs, where something was initalized
	incorrectly (strings = 0 ?!?) (s_net.c in add function)

	Found another missing init (object_ptr) in s_netlist.c (add function)

	Found another missing init (original) in s_cpinlist.c (add function)

	All of the above bugs were apparently the cause of the mysterious
	guile 1.3 problem.  Further testing is needed

10/25/98 Cleaned up all warnings 

10/24/98 Discovered that gnetlist doesn't like guile 1.3

	 Fixed a memory overrite in s_net_name, also added a return(NULL)

	 Fixed a place where I was freeing a null pointer in g_get_packages

	 Found/fixed another "trying to free null data"

 	 Updated gnetlist.scm with display-connections (output.net looks 
	 better)

10/20/98 Fixed a bug which prevented gnetlist to work with guile 1.3

	 Change rc file guile primative version to gnetlist-version

10/18/98 Changed o_attrib_search_name_single to pass a null in for the last
	 argument

10/12/98 Added guile code to handle unconnected pins (gnetlist.scm)

	 Found a place where U? wasn't being set.

10/11/98 Added parsecmd.c (commandline argument parsing)

         Simplified the loading of schematic sheets and intergrated above

	 Added the actual command line options (-v for verbose mode) and
	 (-i for guile interactive mode)  Default for both of these is
	 off

	 Added usage() function (which exits when called)

	 Added quiet_mode to the command line arguments, to shut off all
	 messages

	 Removed all warnings

	 Worked on gnetlist.scm to get some guile code to drive some of 
	 the new netlist code
	
	 Added and registered g_get_package_attribute guile function which
	 returns the wanted attribute from the specified package uref

	 Added command line option -g guile_proc so you can execute guile
	 scripts from the command line 

	 Added code to support above

	 More work into gnetlist.scm to get some sort of netlist writer 
	 going
		

10/10/98 Added file s_traverse.c which creates flat netlist datastructure

	 Serious work on getting netlist code to work!

	 Add s_net.c (functions that deal with the net structure) 

	 Add s_netlist.c (functions that deal with the netlist struct)

	 Add s_cpinlist.c (functions that deal with the cpinlist struct)

	 Completely preliminary code for netlist generation (first cut,
	 highly inefficient, but works! :)

	 Updated g_get_packages to use the new available structures (much
	 faster)

	 Updated g_get_pins to use the new available structures (much
	 faster)

	 Added (exit) and (quit) guile functions

	 Found a condition where my attempt to speed up the netlist post
	 processing, but it found a nasty bug, so now we have a 3 pass
	 post process. 
	

0.0.5
-------

9/19/98 Improved the input gathering code for the guile shell a bit.

9/13/98 Changed netlist.scm to gnetlist.scm

	Fixed g_rc_scheme_directory to store the string in the structure 

9/4/98	Made then necessary changes to have gnetlist use libgeda

	Fixed a few minor tidbits related to above
	
8/30/98 Added embedded component support from gschem to gnetlist
 
8/29/98 Added complex_parent pointer to object structure

	Added code to fill above pointer 

	Added s_netlist.c and added it to Makefile.am

	Added first cut of GUILE netlist routines	
		g_get_packages

	Added s_scratch.c (misc scratch buffer routines)

	Made some changes to the attribute handling, head attrib node 
	points to the object where the attributes belong to.  Be sure 	
	to propegate this change back to gschem (done)

	Made sure that text_string is only set when it should be by 
	setting p_text_string[0] = '\0'.  Be sure to propegate this change
	back to gschem (done)

	Added a bunch of attribute search routines to support all this
	Be sure to propagate these attribute routines back go gschem

	Wrote some test scheme code which is in ../bin/netlist.scm

8/25/98	Added code into the repository and devel tree

	Fixed logging mechanism

	Added appropriate lines in all the upper makefiles to build this code

	Code doesn't do anything other then read in the schematics