make lrecord UID's have a separate UID space for each object, resurrect debug SOE code in extents.c -------------------- ChangeLog entries follow: -------------------- src/ChangeLog addition: 2010-03-15 Ben Wing <ben@xemacs.org> * alloc.c: * alloc.c (c_readonly): * alloc.c (deadbeef_memory): * alloc.c (make_compiled_function): * alloc.c (make_button_data): * alloc.c (make_motion_data): * alloc.c (make_process_data): * alloc.c (make_timeout_data): * alloc.c (make_magic_data): * alloc.c (make_magic_eval_data): * alloc.c (make_eval_data): * alloc.c (make_misc_user_data): * alloc.c (noseeum_make_marker): * alloc.c (ADDITIONAL_FREE_string): * alloc.c (common_init_alloc_early): * alloc.c (init_alloc_once_early): * bytecode.c (print_compiled_function): * bytecode.c (mark_compiled_function): * casetab.c: * casetab.c (print_case_table): * console.c: * console.c (print_console): * database.c (print_database): * database.c (finalize_database): * device-msw.c (sync_printer_with_devmode): * device-msw.c (print_devmode): * device-msw.c (finalize_devmode): * device.c: * device.c (print_device): * elhash.c: * elhash.c (print_hash_table): * eval.c (print_multiple_value): * eval.c (mark_multiple_value): * events.c (deinitialize_event): * events.c (print_event): * events.c (event_equal): * extents.c: * extents.c (soe_dump): * extents.c (soe_insert): * extents.c (soe_delete): * extents.c (soe_move): * extents.c (extent_fragment_update): * extents.c (print_extent_1): * extents.c (print_extent): * extents.c (vars_of_extents): * frame.c: * frame.c (print_frame): * free-hook.c: * free-hook.c (check_free): * glyphs.c: * glyphs.c (print_image_instance): * glyphs.c (print_glyph): * gui.c: * gui.c (copy_gui_item): * hash.c: * hash.c (NULL_ENTRY): * hash.c (KEYS_DIFFER_P): * keymap.c (print_keymap): * keymap.c (MARKED_SLOT): * lisp.h: * lrecord.h: * lrecord.h (LISP_OBJECT_UID): * lrecord.h (set_lheader_implementation): * lrecord.h (struct old_lcrecord_header): * lstream.c (print_lstream): * lstream.c (finalize_lstream): * marker.c (print_marker): * marker.c (marker_equal): * mc-alloc.c (visit_all_used_page_headers): * mule-charset.c: * mule-charset.c (print_charset): * objects.c (print_color_instance): * objects.c (print_font_instance): * objects.c (finalize_font_instance): * opaque.c (print_opaque): * opaque.c (print_opaque_ptr): * opaque.c (equal_opaque_ptr): * print.c (internal_object_printer): * print.c (enum printing_badness): * rangetab.c (print_range_table): * rangetab.c (range_table_equal): * specifier.c (print_specifier): * specifier.c (finalize_specifier): * symbols.c: * symbols.c (print_symbol_value_magic): * tooltalk.c: * tooltalk.c (print_tooltalk_message): * tooltalk.c (print_tooltalk_pattern): * window.c (print_window): * window.c (debug_print_window): (1) Make lrecord UID's have a separate UID space for each object. Otherwise, with 20-bit UID's, we rapidly wrap around, especially when common objects like conses and strings increment the UID value for every object created. (Originally I tried making two UID spaces, one for objects that always print readably and hence don't display the UID, and one for other objects. But certain objects like markers for which a UID is displayed are still generated rapidly enough that UID overflow is a serious issue.) This also has the advantage of making UID values smaller, hence easier to remember -- their main purpose is to make it easier to keep track of different objects of the same type when debugging code. Make sure we dump lrecord UID's so that we don't have problems with pdumped and non-dumped objects having the same UID. (2) Display UID's consistently whenever an object (a) doesn't consistently print readably (objects like cons and string, which always print readably, can't display a UID), and (b) doesn't otherwise have a unique property that makes objects of a particular type distinguishable. (E.g. buffers didn't and still don't print an ID, but the buffer name uniquely identifies the buffer.) Some types, such as event, extent, compiled-function, didn't always (or didn't ever) display an ID; others (such as marker, extent, lstream, opaque, opaque-ptr, any object using internal_object_printer()) used to display the actual machine pointer instead. (3) Rename NORMAL_LISP_OBJECT_UID to LISP_OBJECT_UID; make it work over all Lisp objects and take a Lisp object, not a struct pointer. (4) Some misc cleanups in alloc.c, elhash.c. (5) Change code in events.c that "deinitializes" an event so that it doesn't increment the event UID counter in the process. Also use deadbeef_memory() to overwrite memory instead of doing the same with custom code. In the process, make deadbeef_memory() in alloc.c always available, and delete extraneous copy in mc-alloc.c. Also capitalize all uses of 0xDEADBEEF. Similarly in elhash.c call deadbeef_memory(). (6) Resurrect "debug SOE" code in extents.c. Make it conditional on DEBUG_XEMACS and on a `debug-soe' variable, rather than on SOE_DEBUG. Make it output to stderr, not stdout. (7) Delete some custom print methods that were identical to external_object_printer().

2002-07-17  Marcus Crestani  <crestani@informatik.uni-tuebingen.de>
	    Markus Kaltenbach  <makalten@informatik.uni-tuebingen.de>
	    Mike Sperber <mike@xemacs.org>

	updated 2003-07-29

	New KKCC-GC mark algorithm:
	configure flag : --use-kkcc

	For better understanding, first a few words about the mark algorithm 
	up to now:
	Every Lisp_Object has its own mark method, which calls mark_object
	with the stuff to be marked.
	Also, many Lisp_Objects have pdump descriptions memory_descriptions, 
	which are used by the portable dumper. The dumper gets all the 
	information it needs about the Lisp_Object from the descriptions.

	Also the garbage collector can use the information in the pdump
	descriptions, so we can get rid of the mark methods.
	That is what we have been doing.

	
	DUMPABLE FLAG
	-------------
	First we added a dumpable flag to lrecord_implementation. It shows,
	if the object is dumpable and should be processed by the dumper.
	The dumpable flag is the third argument of a lrecord_implementation
	definition (DEFINE_LRECORD_IMPLEMENTATION).
	If it is set to 1, the dumper processes the descriptions and dumps
	the Object, if it is set to 0, the dumper does not care about it.
		

	KKCC MARKING
	------------
	All Lisp_Objects have memory_descriptions now, so we could get
	rid of the mark_object calls.
	The KKCC algorithm manages its own stack. Instead of calling 
	mark_object, all the alive Lisp_Objects are pushed on the 
	kkcc_gc_stack. Then these elements on the stack  are processed 
	according to their descriptions.


	TODO
	----
	- For weakness use weak datatypes instead of XD_FLAG_NO_KKCC.
	  XD_FLAG_NO_KKCC occurs in:
		* elhash.c: htentry
		* extents.c: lispobject_gap_array, extent_list, extent_info
		* marker.c: marker     
	  Not everything has to be rewritten. See Ben's comment in lrecord.h.
	- Clean up special case marking (weak_hash_tables, weak_lists,
	  ephemerons).
	- Stack optimization (have one stack during runtime instead of 
	  malloc/free it for every garbage collect)

	There are a few Lisp_Objects, where there occured differences and
	inexactness between the mark-method and the pdump description.  All
	these Lisp_Objects get dumped (except image instances), so their
	descriptions have been written, before we started our work:
	* alloc.c: string
	description: size_, data_, and plist is described
	mark: only plist is marked, but flush_cached_extent_info is called.
	      flush_cached_extent_info ->
		free_soe ->
		  free_extent_list ->
		    free_gap_array ->
		      gap_array_delete_all_markers ->
			Add gap_array to the gap_array_marker_freelist

	* glyphs.c: image_instance
	description: device is not set to nil
	mark: mark method sets device to nil if dead
	See comment above the description.