Mercurial > hg > xemacs-beta
view lisp/derived.el @ 2720:6fa9919a9a0b
[xemacs-hg @ 2005-04-08 23:10:01 by crestani]
ChangeLog addition:
2005-04-01 Marcus Crestani <crestani@xemacs.org>
The new allocator.
New configure flag: `MC_ALLOC':
* configure.ac (XE_COMPLEX_ARG_ENABLE): Add `--enable-mc-alloc' as
a new configure flag.
* configure.in (AC_INIT_PARSE_ARGS): Add `--mc-alloc' as a new
configure flag.
* configure.usage: Add description for `mc-alloc'.
DUMP_IN_EXEC:
* Makefile.in.in: Condition the installation of a separate dump
file on !DUMP_ON_EXEC.
* configure.ac (XE_COMPLEX_ARG_ENABLE): Add
`--enable-dump-in-exec' as a new configure flag.
* configure.ac: DUMP_IN_EXEC is define as default for PDUMP but
not default for MC_ALLOC.
* configure.in (AC_INIT_PARSE_ARGS): Add `--dump-in-exec' as a
new configure flag.
* configure.in: DUMP_IN_EXEC is define as default for PDUMP but
not default for MC_ALLOC.
* configure.usage: Add description for `dump-in-exec'.
lib-src/ChangeLog addition:
2005-04-01 Marcus Crestani <crestani@xemacs.org>
The new allocator.
DUMP_IN_EXEC:
* Makefile.in.in: Only compile insert-data-in-exec if
DUMP_IN_EXEC is defined.
lisp/ChangeLog addition:
2005-04-01 Marcus Crestani <crestani@xemacs.org>
The new allocator.
MEMORY_USAGE_STATS
* diagnose.el: Add new lisp function to pretty print statistics
about the new allocator.
* diagnose.el (show-mc-alloc-memory-usage): New.
modules/ChangeLog addition:
2005-04-01 Marcus Crestani <crestani@xemacs.org>
The new allocator.
Remove Lcrecords:
* postgresql/postgresql.c (allocate_pgconn): Allocate with new
allocator.
* postgresql/postgresql.c (allocate_pgresult): Allocate PGresult
with new allocator.
* postgresql/postgresql.h (struct Lisp_PGconn): Add
lrecord_header.
* postgresql/postgresql.h (struct Lisp_PGresult): Add
lrecord_header.
* ldap/eldap.c (allocate_ldap): Allocate with new allocator.
* ldap/eldap.h (struct Lisp_LDAP): Add lrecord_header.
nt/ChangeLog addition:
2005-04-01 Marcus Crestani <crestani@xemacs.org>
The new allocator.
New configure flag: `MC_ALLOC':
* config.inc.samp: Add new flag `MC_ALLOC'.
* xemacs.mak: Add flag and configuration output for `MC_ALLOC'.
New files:
* xemacs.dsp: Add source files mc-alloc.c and mc-alloc.h.
* xemacs.mak: Add new object file mc-alloc.obj to dependencies.
src/ChangeLog addition:
2005-04-01 Marcus Crestani <crestani@xemacs.org>
The new allocator.
New configure flag: `MC_ALLOC':
* config.h.in: Add new flag `MC_ALLOC'.
New files:
* Makefile.in.in: Add new object file mc-alloc.o.
* depend: Add new files to dependencies.
* mc-alloc.c: New.
* mc-alloc.h: New.
Running the new allocator from XEmacs:
* alloc.c (deadbeef_memory): Moved to mc-alloc.c.
* emacs.c (main_1): Initialize the new allocator and add
syms_of_mc_alloc.
* symsinit.h: Add syms_of_mc_alloc.
New lrecord allocation and free functions:
* alloc.c (alloc_lrecord): New. Allocates an lrecord, includes
type checking and initializing of the lrecord_header.
* alloc.c (noseeum_alloc_lrecord): Same as above, but increments
the NOSEEUM cons counter.
* alloc.c (free_lrecord): New. Calls the finalizer and frees the
lrecord.
* lrecord.h: Add lrecord allocation prototypes and comments.
Remove old lrecord FROB block allocation:
* alloc.c (allocate_lisp_storage): Former function to expand
heap. Not needed anymore, remove.
* alloc.c: Completely remove `Fixed-size type macros'
* alloc.c (release_breathing_space): Remove.
* alloc.c (memory_full): Remove release_breathing_space.
* alloc.c (refill_memory_reserve): Remove.
* alloc.c (TYPE_ALLOC_SIZE): Remove.
* alloc.c (DECLARE_FIXED_TYPE_ALLOC): Remove.
* alloc.c (ALLOCATE_FIXED_TYPE_FROM_BLOCK): Remove.
* alloc.c (ALLOCATE_FIXED_TYPE_1): Remove.
* alloc.c (ALLOCATE_FIXED_TYPE): Remove.
* alloc.c (NOSEEUM_ALLOCATE_FIXED_TYPE): Remove.
* alloc.c (struct Lisp_Free): Remove.
* alloc.c (LRECORD_FREE_P): Remove.
* alloc.c (MARK_LRECORD_AS_FREE): Remove.
* alloc.c (MARK_LRECORD_AS_NOT_FREE): Remove.
* alloc.c (PUT_FIXED_TYPE_ON_FREE_LIST): Remove.
* alloc.c (FREE_FIXED_TYPE): Remove.
* alloc.c (FREE_FIXED_TYPE_WHEN_NOT_IN_GC): Remove.
Allocate old lrecords with new allocator:
* alloc.c: DECLARE_FIXED_TYPE_ALLOC removed for all lrecords
defined in alloc.c.
* alloc.c (Fcons): Allocate with new allocator.
* alloc.c (noseeum_cons): Allocate with new allocator.
* alloc.c (make_float): Allocate with new allocator.
* alloc.c (make_bignum): Allocate with new allocator.
* alloc.c (make_bignum_bg): Allocate with new allocator.
* alloc.c (make_ratio): Allocate with new allocator.
* alloc.c (make_ratio_bg): Allocate with new allocator.
* alloc.c (make_ratio_rt): Allocate with new allocator.
* alloc.c (make_bigfloat): Allocate with new allocator.
* alloc.c (make_bigfloat_bf): Allocate with new allocator.
* alloc.c (make_compiled_function): Allocate with new allocator.
* alloc.c (Fmake_symbol): Allocate with new allocator.
* alloc.c (allocate_extent): Allocate with new allocator.
* alloc.c (allocate_event): Allocate with new allocator.
* alloc.c (make_key_data): Allocate with new allocator.
* alloc.c (make_button_data): Allocate with new allocator.
* alloc.c (make_motion_data): Allocate with new allocator.
* alloc.c (make_process_data): Allocate with new allocator.
* alloc.c (make_timeout_data): Allocate with new allocator.
* alloc.c (make_magic_data): Allocate with new allocator.
* alloc.c (make_magic_eval_data): Allocate with new allocator.
* alloc.c (make_eval_data): Allocate with new allocator.
* alloc.c (make_misc_user_data): Allocate with new allocator.
* alloc.c (Fmake_marker): Allocate with new allocator.
* alloc.c (noseeum_make_marker): Allocate with new allocator.
* alloc.c (make_uninit_string): Allocate with new allocator.
* alloc.c (resize_string): Allocate with new allocator.
* alloc.c (make_string_nocopy): Allocate with new allocator.
Garbage Collection:
* alloc.c (GC_CHECK_NOT_FREE): Remove obsolete assertions.
* alloc.c (SWEEP_FIXED_TYPE_BLOCK): Remove.
* alloc.c (SWEEP_FIXED_TYPE_BLOCK_1): Remove.
* alloc.c (sweep_conses): Remove.
* alloc.c (free_cons): Use new allocator to free.
* alloc.c (sweep_compiled_functions): Remove.
* alloc.c (sweep_floats): Remove.
* alloc.c (sweep_bignums): Remove.
* alloc.c (sweep_ratios): Remove.
* alloc.c (sweep_bigfloats): Remove.
* alloc.c (sweep_symbols): Remove.
* alloc.c (sweep_extents): Remove.
* alloc.c (sweep_events): Remove.
* alloc.c (sweep_key_data): Remove.
* alloc.c (free_key_data): Use new allocator to free.
* alloc.c (sweep_button_data): Remove.
* alloc.c (free_button_data): Use new allocator to free.
* alloc.c (sweep_motion_data): Remove.
* alloc.c (free_motion_data): Use new allocator to free.
* alloc.c (sweep_process_data): Remove.
* alloc.c (free_process_data): Use new allocator to free.
* alloc.c (sweep_timeout_data): Remove.
* alloc.c (free_timeout_data): Use new allocator to free.
* alloc.c (sweep_magic_data): Remove.
* alloc.c (free_magic_data): Use new allocator to free.
* alloc.c (sweep_magic_eval_data): Remove.
* alloc.c (free_magic_eval_data): Use new allocator to free.
* alloc.c (sweep_eval_data): Remove.
* alloc.c (free_eval_data): Use new allocator to free.
* alloc.c (sweep_misc_user_data): Remove.
* alloc.c (free_misc_user_data): Use new allocator to free.
* alloc.c (sweep_markers): Remove.
* alloc.c (free_marker): Use new allocator to free.
* alloc.c (garbage_collect_1): Remove release_breathing_space.
* alloc.c (gc_sweep): Remove all the old lcrecord and lrecord
related stuff. Sweeping now works like this: compact string
chars, finalize, sweep.
* alloc.c (common_init_alloc_early): Remove old lrecord
initializations, remove breathing_space.
* emacs.c (Fdump_emacs): Remove release_breathing_space.
* lisp.h: Remove prototype for release_breathing_space.
* lisp.h: Adjust the special cons mark makros.
Lrecord Finalizer:
* alloc.c: Add finalizer to lrecord definition.
* alloc.c (finalize_string): Add finalizer for string.
* bytecode.c: Add finalizer to lrecord definition.
* bytecode.c (finalize_compiled_function): Add finalizer for
compiled function.
* marker.c: Add finalizer to lrecord definition.
* marker.c (finalize_marker): Add finalizer for marker.
These changes build the interface to mc-alloc:
* lrecord.h (MC_ALLOC_CALL_FINALIZER): Tell mc-alloc how to
finalize lrecords.
* lrecord.h (MC_ALLOC_CALL_FINALIZER_FOR_DISKSAVE): Tell
mc-alloc how to finalize for disksave.
Unify lrecords and lcrecords:
* lisp.h (struct Lisp_String): Adjust string union hack to
new lrecord header.
* lrecord.h: Adjust comments.
* lrecord.h (struct lrecord_header): The new lrecord header
includes type, lisp-readonly, free, and uid.
* lrecord.h (set_lheader_implementation): Adjust to new
lrecord_header.
* lrecord.h (struct lrecord_implementation): The field basic_p
for indication of an old lrecord is not needed anymore, remove.
* lrecord.h (MAKE_LRECORD_IMPLEMENTATION): Remove basic_p.
* lrecord.h (MAKE_EXTERNAL_LRECORD_IMPLEMENTATION): Remove
basic_p.
* lrecord.h (copy_sized_lrecord): Remove distinction between
old lrecords and lcrecords.
* lrecord.h (copy_lrecord): Remove distinction between old
lrecords and lcrecords.
* lrecord.h (zero_sized_lrecord): Remove distinction between
old lrecords and lcrecords.
* lrecord.h (zero_lrecord): Remove distinction between old
lrecords and lcrecords.
Remove lcrecords and lcrecord lists:
* alloc.c (basic_alloc_lcrecord): Not needed anymore, remove.
* alloc.c (very_old_free_lcrecord): Not needed anymore, remove.
* alloc.c (copy_lisp_object): No more distinction between
lrecords and lcrecords.
* alloc.c (all_lcrecords): Not needed anymore, remove.
* alloc.c (make_vector_internal): Allocate as lrecord.
* alloc.c (make_bit_vector_internal): Allocate as lrecord.
* alloc.c: Completely remove `lcrecord lists'.
* alloc.c (free_description): Remove.
* alloc.c (lcrecord_list_description): Remove.
* alloc.c (mark_lcrecord_list): Remove.
* alloc.c (make_lcrecord_list): Remove.
* alloc.c (alloc_managed_lcrecord): Remove.
* alloc.c (free_managed_lcrecord): Remove.
* alloc.c (alloc_automanaged_lcrecord): Remove.
* alloc.c (free_lcrecord): Remove.
* alloc.c (lcrecord_stats): Remove.
* alloc.c (tick_lcrecord_stats): Remove.
* alloc.c (disksave_object_finalization_1): Add call to
mc_finalize_for_disksave. Remove the lcrecord way to visit all
objects.
* alloc.c (kkcc_marking): Remove XD_FLAG_FREE_LISP_OBJECT
* alloc.c (sweep_lcrecords_1): Remove.
* alloc.c (common_init_alloc_early): Remove everything related
to lcrecords, remove old lrecord initializations,
* alloc.c (init_lcrecord_lists): Not needed anymore, remove.
* alloc.c (reinit_alloc_early): Remove everything related to
lcrecords.
* alloc.c (init_alloc_once_early): Remove everything related to
lcrecords.
* buffer.c (allocate_buffer): Allocate as lrecord.
* buffer.c (nuke_all_buffer_slots): Use lrecord functions.
* buffer.c (common_init_complex_vars_of_buffer): Allocate as
lrecord.
* buffer.h (struct buffer): Add lrecord_header.
* casetab.c (allocate_case_table): Allocate as lrecord.
* casetab.h (struct Lisp_Case_Table): Add lrecord_header.
* charset.h (struct Lisp_Charset): Add lrecord_header.
* chartab.c (fill_char_table): Use lrecord functions.
* chartab.c (Fmake_char_table): Allocate as lrecord.
* chartab.c (make_char_table_entry): Allocate as lrecord.
* chartab.c (copy_char_table_entry): Allocate as lrecord.
* chartab.c (Fcopy_char_table): Allocate as lrecord.
* chartab.c (put_char_table): Use lrecord functions.
* chartab.h (struct Lisp_Char_Table_Entry): Add lrecord_header.
* chartab.h (struct Lisp_Char_Table): Add lrecord_header.
* console-impl.h (struct console): Add lrecord_header.
* console-msw-impl.h (struct Lisp_Devmode): Add lrecord_header.
* console-msw-impl.h (struct mswindows_dialog_id): Add
lrecord_header.
* console.c (allocate_console): Allocate as lrecord.
* console.c (nuke_all_console_slots): Use lrecord functions.
* console.c (common_init_complex_vars_of_console): Allocate as
lrecord.
* data.c (make_weak_list): Allocate as lrecord.
* data.c (make_weak_box): Allocate as lrecord.
* data.c (make_ephemeron): Allocate as lrecord.
* database.c (struct Lisp_Database): Add lrecord_header.
* database.c (allocate_database): Allocate as lrecord.
* device-impl.h (struct device): Add lrecord_header.
* device-msw.c (allocate_devmode): Allocate as lrecord.
* device.c (nuke_all_device_slots): Use lrecord functions.
* device.c (allocate_device): Allocate as lrecord.
* dialog-msw.c (handle_question_dialog_box): Allocate as lrecord.
* elhash.c (struct Lisp_Hash_Table): Add lrecord_header.
* elhash.c (make_general_lisp_hash_table): Allocate as lrecord.
* elhash.c (Fcopy_hash_table): Allocate as lrecord.
* event-stream.c: Lcrecord lists Vcommand_builder_free_list and
Vtimeout_free_list are no longer needed. Remove.
* event-stream.c (allocate_command_builder): Allocate as lrecord.
* event-stream.c (free_command_builder): Use lrecord functions.
* event-stream.c (event_stream_generate_wakeup): Allocate as
lrecord.
* event-stream.c (event_stream_resignal_wakeup): Use lrecord
functions.
* event-stream.c (event_stream_disable_wakeup): Use lrecord
functions.
* event-stream.c (reinit_vars_of_event_stream): Lcrecord lists
remove.
* events.h (struct Lisp_Timeout): Add lrecord_header.
* events.h (struct command_builder): Add lrecord_header.
* extents-impl.h (struct extent_auxiliary): Add lrecord_header.
* extents-impl.h (struct extent_info): Add lrecord_header.
* extents.c (allocate_extent_auxiliary): Allocate as lrecord.
* extents.c (allocate_extent_info): Allocate as lrecord.
* extents.c (copy_extent): Allocate as lrecord.
* faces.c (allocate_face): Allocate as lrecord.
* faces.h (struct Lisp_Face): Add lrecord_header.
* file-coding.c (allocate_coding_system): Allocate as lrecord.
* file-coding.c (Fcopy_coding_system): Allocate as lrecord.
* file-coding.h (struct Lisp_Coding_System): Add lrecord_header.
* fns.c (Ffillarray): Allocate as lrecord.
* frame-impl.h (struct frame): Add lrecord_header.
* frame.c (nuke_all_frame_slots): Use lrecord functions.
* frame.c (allocate_frame_core): Allocate as lrecord.
* glyphs.c (allocate_image_instance): Allocate as lrecord.
* glyphs.c (Fcolorize_image_instance): Allocate as lrecord.
* glyphs.c (allocate_glyph): Allocate as lrecord.
* glyphs.h (struct Lisp_Image_Instance): Add lrecord_header.
* glyphs.h (struct Lisp_Glyph): Add lrecord_header.
* gui.c (allocate_gui_item): Allocate as lrecord.
* gui.h (struct Lisp_Gui_Item): Add lrecord_header.
* keymap.c (struct Lisp_Keymap): Add lrecord_header.
* keymap.c (make_keymap): Allocate as lrecord.
* lisp.h (struct Lisp_Vector): Add lrecord_header.
* lisp.h (struct Lisp_Bit_Vector): Add lrecord_header.
* lisp.h (struct weak_box): Add lrecord_header.
* lisp.h (struct ephemeron): Add lrecord_header.
* lisp.h (struct weak_list): Add lrecord_header.
* lrecord.h (struct lcrecord_header): Not used, remove.
* lrecord.h (struct free_lcrecord_header): Not used, remove.
* lrecord.h (struct lcrecord_list): Not needed anymore, remove.
* lrecord.h (lcrecord_list): Not needed anymore, remove.
* lrecord.h: (enum data_description_entry_flags): Remove
XD_FLAG_FREE_LISP_OBJECT.
* lstream.c: Lrecord list Vlstream_free_list remove.
* lstream.c (Lstream_new): Allocate as lrecord.
* lstream.c (Lstream_delete): Use lrecod functions.
* lstream.c (reinit_vars_of_lstream): Vlstream_free_list
initialization remove.
* lstream.h (struct lstream): Add lrecord_header.
* emacs.c (main_1): Remove lstream initialization.
* mule-charset.c (make_charset): Allocate as lrecord.
* objects-impl.h (struct Lisp_Color_Instance): Add
lrecord_header.
* objects-impl.h (struct Lisp_Font_Instance): Add lrecord_header.
* objects.c (Fmake_color_instance): Allocate as lrecord.
* objects.c (Fmake_font_instance): Allocate as lrecord.
* objects.c (reinit_vars_of_objects): Allocate as lrecord.
* opaque.c: Lcreord list Vopaque_ptr_list remove.
* opaque.c (make_opaque): Allocate as lrecord.
* opaque.c (make_opaque_ptr): Allocate as lrecord.
* opaque.c (free_opaque_ptr): Use lrecord functions.
* opaque.c (reinit_opaque_early):
* opaque.c (init_opaque_once_early): Vopaque_ptr_list
initialization remove.
* opaque.h (Lisp_Opaque): Add lrecord_header.
* opaque.h (Lisp_Opaque_Ptr): Add lrecord_header.
* emacs.c (main_1): Remove opaque variable initialization.
* print.c (default_object_printer): Use new lrecord_header.
* print.c (print_internal): Use new lrecord_header.
* print.c (debug_p4): Use new lrecord_header.
* process.c (make_process_internal): Allocate as lrecord.
* procimpl.h (struct Lisp_Process): Add lrecord_header.
* rangetab.c (Fmake_range_table): Allocate as lrecord.
* rangetab.c (Fcopy_range_table): Allocate as lrecord.
* rangetab.h (struct Lisp_Range_Table): Add lrecord_header.
* scrollbar.c (create_scrollbar_instance): Allocate as lrecord.
* scrollbar.h (struct scrollbar_instance): Add lrecord_header.
* specifier.c (make_specifier_internal): Allocate as lrecord.
* specifier.h (struct Lisp_Specifier): Add lrecord_header.
* symbols.c:
* symbols.c (Fmake_variable_buffer_local): Allocate as lrecord.
* symbols.c (Fdontusethis_set_symbol_value_handler): Allocate
as lrecord.
* symbols.c (Fdefvaralias): Allocate as lrecord.
* symeval.h (struct symbol_value_magic): Add lrecord_header.
* toolbar.c (update_toolbar_button): Allocate as lrecord.
* toolbar.h (struct toolbar_button): Add lrecord_header.
* tooltalk.c (struct Lisp_Tooltalk_Message): Add lrecord_header.
* tooltalk.c (make_tooltalk_message): Allocate as lrecord.
* tooltalk.c (struct Lisp_Tooltalk_Pattern): Add lrecord_header.
* tooltalk.c (make_tooltalk_pattern): Allocate as lrecord.
* ui-gtk.c (allocate_ffi_data): Allocate as lrecord.
* ui-gtk.c (allocate_emacs_gtk_object_data): Allocate as lrecord.
* ui-gtk.c (allocate_emacs_gtk_boxed_data): Allocate as lrecord.
* ui-gtk.h (structs): Add lrecord_header.
* window-impl.h (struct window): Add lrecord_header.
* window-impl.h (struct window_mirror): Add lrecord_header.
* window.c (allocate_window): Allocate as lrecord.
* window.c (new_window_mirror): Allocate as lrecord.
* window.c (make_dummy_parent): Allocate as lrecord.
MEMORY_USAGE_STATS
* alloc.c (fixed_type_block_overhead): Not used anymore, remove.
* buffer.c (compute_buffer_usage): Get storage size from new
allocator.
* marker.c (compute_buffer_marker_usage): Get storage size from
new allocator.
* mule-charset.c (compute_charset_usage): Get storage size from
new allocator.
* scrollbar-gtk.c (gtk_compute_scrollbar_instance_usage): Get
storage size from new allocator.
* scrollbar-msw.c (mswindows_compute_scrollbar_instance_usage):
Get storage size from new allocator.
* scrollbar-x.c (x_compute_scrollbar_instance_usage): Get
storage size from new allocator.
* scrollbar.c (compute_scrollbar_instance_usage): Get storage
size from new allocator.
* unicode.c (compute_from_unicode_table_size_1): Get storage
size from new allocator.
* unicode.c (compute_to_unicode_table_size_1): Get storage size
from new allocator.
* window.c (compute_window_mirror_usage): Get storage size from
new allocator.
* window.c (compute_window_usage): Get storage size from new
allocator.
MC_ALLOC_TYPE_STATS:
* alloc.c (alloc_lrecord): Bump lrecord count.
* alloc.c (noseeum_alloc_lrecord): Bump lrecord count.
* alloc.c (struct lrecord_stats): Storage for counts.
* alloc.c (init_lrecord_stats): Zero statistics.
* alloc.c (inc_lrecord_stats): Increase the statistic.
* alloc.c (dec_lrecord_stats): Decrease the statistic.
* alloc.c (gc_plist_hack): Used to print the information.
* alloc.c (Fgarbage_collect): Return the collected information.
* mc-alloc.c (remove_cell): Decrease lrecord count.
* mc-alloc.h: Set flag MC_ALLOC_TYPE_STATS.
* emacs.c (main_1): Init lrecord statistics.
* lrecord.h: Add prototypes for *_lrecord_stats.
Strings:
* alloc.c (Fmake_string): Initialize ascii_begin to zero.
* alloc.c (gc_count_num_short_string_in_use): Remove.
* alloc.c (gc_count_string_total_size): Remove.
* alloc.c (gc_count_short_string_total_size): Remove.
* alloc.c (debug_string_purity): Remove.
* alloc.c (debug_string_purity_print): Remove.
* alloc.c (sweep_strings): Remove.
Remove static C-readonly Lisp objects:
* alloc.c (c_readonly): Not needed anymore, remove.
* alloc.c (GC_CHECK_LHEADER_INVARIANTS): Remove some obsolete
lheader invariants assertions.
* buffer.c (DEFVAR_BUFFER_LOCAL_1): Allocate dynamically.
* console.c (DEFVAR_CONSOLE_LOCAL_1): Allocate dynamically.
* gpmevent.c: Indirection via MC_ALLOC_Freceive_gpm_event.
* gpmevent.c (Fgpm_enable): Allocate dynamically.
* gpmevent.c (syms_of_gpmevent): Allocate dynamically.
* lisp.h (C_READONLY): Not needed anymore, remove.
* lisp.h (DEFUN): Allocate dynamically.
* lrecord.h (C_READONLY_RECORD_HEADER_P): Not needed anymore,
remove.
* lrecord.h (SET_C_READONLY_RECORD_HEADER): Not needed anymore,
remove.
* symbols.c (guts_of_unbound_marker):
* symeval.h (defsubr): Allocate dynamically.
* symeval.h (DEFSUBR_MACRO): Allocate dynamically.
* symeval.h (DEFVAR_ SYMVAL_FWD): Allocate dynamically.
* tests.c (TESTS_DEFSUBR): Allocate dynamically.
Definition of mcpro:
* lisp.h: Add mcpro prototypes.
* alloc.c (common_init_alloc_early): Add initialization for
mcpros.
* alloc.c (mcpro_description_1): New.
* alloc.c (mcpro_description): New.
* alloc.c (mcpros_description_1): New.
* alloc.c (mcpros_description): New.
* alloc.c (mcpro_one_name_description_1): New.
* alloc.c (mcpro_one_name_description): New.
* alloc.c (mcpro_names_description_1): New.
* alloc.c (mcpro_names_description): New.
* alloc.c (mcpros): New.
* alloc.c (mcpro_names): New.
* alloc.c (mcpro_1): New.
* alloc.c (mc_pro): New.
* alloc.c (garbage_collect_1): Add mcpros to root set.
Usage of mcpro:
* alloc.c (make_string_nocopy): Add string to root set.
* symbols.c (init_symbols_once_early): Add Qunbound to root set.
Changes to the Portable Dumper:
* alloc.c (FREE_OR_REALLOC_BEGIN): Since dumped objects can be
freed with the new allocator, remove assertion for !DUMPEDP.
* dumper.c: Adjust comments, increase PDUMP_HASHSIZE.
* dumper.c (pdump_make_hash): Shift address only 2 bytes, to
avoid collisions.
* dumper.c (pdump_objects_unmark): No more mark bits within
the object, remove.
* dumper.c (mc_addr_elt): New. Element data structure for mc
hash table.
* dumper.c (pdump_mc_hash): New hash table: `lookup table'.
* dumper.c (pdump_get_mc_addr): New. Lookup for hash table.
* dumper.c (pdump_get_indirect_mc_addr): New. Lookup for
convertibles.
* dumper.c (pdump_put_mc_addr): New. Putter for hash table.
* dumper.c (pdump_dump_mc_data): New. Writes the table for
relocation at load time to the dump file.
* dumper.c (pdump_scan_lisp_objects_by_alignment): New.
Visits all dumped Lisp objects.
* dumper.c (pdump_scan_non_lisp_objects_by_alignment): New.
Visits all other dumped objects.
* dumper.c (pdump_reloc_one_mc): New. Updates all pointers
of an object by using the hash table pdump_mc_hash.
* dumper.c (pdump_reloc_one): Replaced by pdump_reloc_one_mc.
* dumper.c (pdump): Change the structure of the dump file, add
the mc post dump relocation table to dump file.
* dumper.c (pdump_load_finish): Hand all dumped objects to the
new allocator and use the mc post dump relocation table for
relocating the dumped objects at dump file load time, free not
longer used data structures.
* dumper.c (pdump_load): Free the dump file.
* dumper.h: Remove pdump_objects_unmark.
* lrecord.h (DUMPEDP): Dumped objects can be freed, remove.
DUMP_IN_EXEC:
* Makefile.in.in: Linking for and with dump in executable only if
DUMP_IN_EXEC is defined.
* config.h.in: Add new flag `DUMP_IN_EXEC'
* emacs.c: Condition dump-data.h on DUMP_IN_EXEC.
* emacs.c (main_1): Flag `-si' only works if dump image is
written into executable.
Miscellanious
* lrecord.h (enum lrecord_type): Added numbers to all types,
very handy for debugging.
* xemacs.def.in.in: Add mc-alloc functions to make them visible
to the modules.
| author | crestani |
|---|---|
| date | Fri, 08 Apr 2005 23:11:35 +0000 |
| parents | eb65d362090f |
| children | fd1acd2f457a |
line wrap: on
line source
;;; derived.el --- allow inheritance of major modes ;;; (formerly mode-clone.el) ;; Copyright (C) 1993, 1994, 1999, 2003 Free Software Foundation, Inc. ;; Author: David Megginson (dmeggins@aix1.uottawa.ca) ;; Maintainer: XEmacs Development Team ;; Keywords: extensions, dumped ;; This file is part of XEmacs. ;; XEmacs is free software; you can redistribute it and/or modify it ;; under the terms of the GNU General Public License as published by ;; the Free Software Foundation; either version 2, or (at your option) ;; any later version. ;; XEmacs is distributed in the hope that it will be useful, but ;; WITHOUT ANY WARRANTY; without even the implied warranty of ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ;; General Public License for more details. ;; You should have received a copy of the GNU General Public License ;; along with XEmacs; see the file COPYING. If not, write to the Free ;; Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA ;; 02111-1307, USA. ;;; Synched up with: FSF 21.3. ;;; Commentary: ;; This file is dumped with XEmacs. ;; XEmacs is already, in a sense, object oriented -- each object ;; (buffer) belongs to a class (major mode), and that class defines ;; the relationship between messages (input events) and methods ;; (commands) by means of a keymap. ;; ;; The only thing missing is a good scheme of inheritance. It is ;; possible to simulate a single level of inheritance with generous ;; use of hooks and a bit of work -- sgml-mode, for example, also runs ;; the hooks for text-mode, and keymaps can inherit from other keymaps ;; -- but generally, each major mode ends up reinventing the wheel. ;; Ideally, someone should redesign all of Emacs's major modes to ;; follow a more conventional object-oriented system: when defining a ;; new major mode, the user should need only to name the existing mode ;; it is most similar to, then list the (few) differences. ;; ;; In the mean time, this package offers most of the advantages of ;; full inheritance with the existing major modes. The macro ;; `define-derived-mode' allows the user to make a variant of an existing ;; major mode, with its own keymap. The new mode will inherit the key ;; bindings of its parent, and will, in fact, run its parent first ;; every time it is called. For example, the commands ;; ;; (define-derived-mode hypertext-mode text-mode "Hypertext" ;; "Major mode for hypertext.\n\n\\{hypertext-mode-map}" ;; (setq case-fold-search nil)) ;; ;; (define-key hypertext-mode-map [down-mouse-3] 'do-hyper-link) ;; ;; will create a function `hypertext-mode' with its own (sparse) ;; keymap `hypertext-mode-map.' The command M-x hypertext-mode will ;; perform the following actions: ;; ;; - run the command (text-mode) to get its default setup ;; - replace the current keymap with 'hypertext-mode-map,' which will ;; inherit from 'text-mode-map'. ;; - replace the current syntax table with ;; 'hypertext-mode-syntax-table', which will borrow its defaults ;; from the current text-mode-syntax-table. ;; - replace the current abbrev table with ;; 'hypertext-mode-abbrev-table', which will borrow its defaults ;; from the current text-mode-abbrev table ;; - change the mode line to read "Hypertext" ;; - assign the value 'hypertext-mode' to the 'major-mode' variable ;; - run the body of commands provided in the macro -- in this case, ;; set the local variable `case-fold-search' to nil. ;; ;; The advantages of this system are threefold. First, text mode is ;; untouched -- if you had added the new keystroke to `text-mode-map,' ;; possibly using hooks, you would have added it to all text buffers ;; -- here, it appears only in hypertext buffers, where it makes ;; sense. Second, it is possible to build even further, and make ;; a derived mode from a derived mode. The commands ;; ;; (define-derived-mode html-mode hypertext-mode "HTML") ;; [various key definitions] ;; ;; will add a new major mode for HTML with very little fuss. ;; ;; Note also the function `derived-mode-p' which can tell if the current ;; mode derives from another. In a hypertext-mode, buffer, for example, ;; (derived-mode-p 'text-mode) would return non-nil. This should always ;; be used in place of (eq major-mode 'text-mode). ;;; Code: ;;; PRIVATE: defsubst must be defined before they are first used (defsubst derived-mode-hook-name (mode) "Construct the mode hook name based on mode name MODE." (intern (concat (symbol-name mode) "-hook"))) (defsubst derived-mode-map-name (mode) "Construct a map name based on a MODE name." (intern (concat (symbol-name mode) "-map"))) (defsubst derived-mode-syntax-table-name (mode) "Construct a syntax-table name based on a MODE name." (intern (concat (symbol-name mode) "-syntax-table"))) (defsubst derived-mode-abbrev-table-name (mode) "Construct an abbrev-table name based on a MODE name." (intern (concat (symbol-name mode) "-abbrev-table"))) ;; PUBLIC: define a new major mode which inherits from an existing one. ;; XEmacs -- no autoload (defmacro define-derived-mode (child parent name &optional docstring &rest body) "Create a new mode as a variant of an existing mode. The arguments to this command are as follow: CHILD: the name of the command for the derived mode. PARENT: the name of the command for the parent mode (e.g. `text-mode') or nil if there is no parent. NAME: a string which will appear in the status line (e.g. \"Hypertext\") DOCSTRING: an optional documentation string--if you do not supply one, the function will attempt to invent something useful. BODY: forms to execute just before running the hooks for the new mode. Do not use `interactive' here. BODY can start with a bunch of keyword arguments. The following keyword arguments are currently understood: :group GROUP Declare the customization group that corresponds to this mode. :syntax-table TABLE Use TABLE instead of the default. A nil value means to simply use the same syntax-table as the parent. :abbrev-table TABLE Use TABLE instead of the default. A nil value means to simply use the same abbrev-table as the parent. Here is how you could define LaTeX-Thesis mode as a variant of LaTeX mode: (define-derived-mode LaTeX-thesis-mode LaTeX-mode \"LaTeX-Thesis\") You could then make new key bindings for `LaTeX-thesis-mode-map' without changing regular LaTeX mode. In this example, BODY is empty, and DOCSTRING is generated by default. On a more complicated level, the following command uses `sgml-mode' as the parent, and then sets the variable `case-fold-search' to nil: (define-derived-mode article-mode sgml-mode \"Article\" \"Major mode for editing technical articles.\" (setq case-fold-search nil)) Note that if the documentation string had been left out, it would have been generated automatically, with a reference to the keymap. The new mode runs the hook constructed by the function `derived-mode-hook-name'." (declare (debug (&define name symbolp sexp [&optional stringp] [&rest keywordp sexp] def-body))) (when (and docstring (not (stringp docstring))) ;; Some trickiness, since what appears to be the docstring may really be ;; the first element of the body. (push docstring body) (setq docstring nil)) (when (eq parent 'fundamental-mode) (setq parent nil)) (let ((map (derived-mode-map-name child)) (syntax (derived-mode-syntax-table-name child)) (abbrev (derived-mode-abbrev-table-name child)) (declare-abbrev t) (declare-syntax t) (hook (derived-mode-hook-name child)) (group nil)) ;; Process the keyword args. (while (keywordp (car body)) (case (pop body) (:group (setq group (pop body))) (:abbrev-table (setq abbrev (pop body)) (setq declare-abbrev nil)) (:syntax-table (setq syntax (pop body)) (setq declare-syntax nil)) (t (pop body)))) (setq docstring (derived-mode-make-docstring parent child docstring syntax abbrev)) `(progn (defvar ,hook nil ,(format "Hook run when entering %s mode." name)) (defvar ,map (make-sparse-keymap)) ,(if declare-syntax `(defvar ,syntax (make-syntax-table))) ,(if declare-abbrev `(defvar ,abbrev (progn (define-abbrev-table ',abbrev nil) ,abbrev))) (put ',child 'derived-mode-parent ',parent) ,(if group `(put ',child 'custom-mode-group ,group)) (defun ,child () ,docstring (interactive) ; Run the parent. (delay-mode-hooks (,(or parent 'kill-all-local-variables)) ; Identify the child mode. (setq major-mode (quote ,child)) (setq mode-name ,name) ; Identify special modes. ,(when parent `(progn (if (get (quote ,parent) 'mode-class) (put (quote ,child) 'mode-class (get (quote ,parent) 'mode-class))) ; Set up maps and tables. (unless (keymap-parent ,map) (set-keymap-parents ,map (list (current-local-map)))) ,(when declare-syntax ;; XEmacs change: we do not have char-table-parent `(derived-mode-merge-syntax-tables (syntax-table) ,syntax)))) (use-local-map ,map) ,(when syntax `(set-syntax-table ,syntax)) ,(when abbrev `(setq local-abbrev-table ,abbrev)) ; Splice in the body (if any). ,@body ) ;; Run the hooks, if any. ;; Make the generated code work in older Emacs versions ;; that do not yet have run-mode-hooks. (if (fboundp 'run-mode-hooks) (run-mode-hooks ',hook) (run-hooks ',hook)))))) ;; PUBLIC: find the ultimate class of a derived mode. (defun derived-mode-class (mode) "Find the class of a major MODE. A mode's class is the first ancestor which is NOT a derived mode. Use the `derived-mode-parent' property of the symbol to trace backwards. Since major-modes might all derive from `fundamental-mode', this function is not very useful." (while (get mode 'derived-mode-parent) (setq mode (get mode 'derived-mode-parent))) mode) (make-obsolete 'derived-mode-class 'derived-mode-p) ;; PUBLIC: find if the current mode derives from another. ;; from GNU Emacs 21 subr.el (defun derived-mode-p (&rest modes) "Non-nil if the current major mode is derived from one of MODES. Uses the `derived-mode-parent' property of the symbol to trace backwards." (let ((parent major-mode)) (while (and (not (memq parent modes)) (setq parent (get parent 'derived-mode-parent)))) parent)) ;;; PRIVATE (defun derived-mode-make-docstring (parent child &optional docstring syntax abbrev) "Construct a docstring for a new mode if none is provided." (let ((map (derived-mode-map-name child)) (hook (derived-mode-hook-name child))) (unless (stringp docstring) ;; Use a default docstring. (setq docstring (if (null parent) (format "Major-mode. Uses keymap `%s', abbrev table `%s' and syntax-table `%s'." map abbrev syntax) (format "Major mode derived from `%s' by `define-derived-mode'. It inherits all of the parent's attributes, but has its own keymap, abbrev table and syntax table: `%s', `%s' and `%s' which more-or-less shadow %s's corresponding tables." parent map abbrev syntax parent)))) (unless (string-match (regexp-quote (symbol-name hook)) docstring) ;; Make sure the docstring mentions the mode's hook. (setq docstring (concat docstring (if (null parent) "\n\nThis mode " (concat "\n\nIn addition to any hooks its parent mode " (if (string-match (regexp-quote (format "`%s'" parent)) docstring) nil (format "`%s' " parent)) "might have run,\nthis mode ")) (format "runs the hook `%s'" hook) ", as the final step\nduring initialization."))) (unless (string-match "\\\\[{[]" docstring) ;; And don't forget to put the mode's keymap. (setq docstring (concat docstring "\n\n\\{" (symbol-name map) "}"))) docstring)) ;;; OBSOLETE ;; The functions below are only provided for backward compatibility with ;; code byte-compiled with versions of derived.el prior to Emacs-21. (defsubst derived-mode-setup-function-name (mode) "Construct a setup-function name based on a MODE name." (intern (concat (symbol-name mode) "-setup"))) ;; Utility functions for defining a derived mode. ;; XEmacs -- don't autoload (defun derived-mode-init-mode-variables (mode) "Initialise variables for a new MODE. Right now, if they don't already exist, set up a blank keymap, an empty syntax table, and an empty abbrev table -- these will be merged the first time the mode is used." (if (boundp (derived-mode-map-name mode)) t (eval `(defvar ,(derived-mode-map-name mode) ;; XEmacs change (make-sparse-keymap (derived-mode-map-name mode)) ,(format "Keymap for %s." mode))) (put (derived-mode-map-name mode) 'derived-mode-unmerged t)) (if (boundp (derived-mode-syntax-table-name mode)) t (eval `(defvar ,(derived-mode-syntax-table-name mode) ;; XEmacs change ;; Make a syntax table which doesn't specify anything ;; for any char. Valid data will be merged in by ;; derived-mode-merge-syntax-tables. ;; (make-char-table 'syntax-table nil) (make-syntax-table) ,(format "Syntax table for %s." mode))) (put (derived-mode-syntax-table-name mode) 'derived-mode-unmerged t)) (if (boundp (derived-mode-abbrev-table-name mode)) t (eval `(defvar ,(derived-mode-abbrev-table-name mode) (progn (define-abbrev-table (derived-mode-abbrev-table-name mode) nil) (make-abbrev-table)) ,(format "Abbrev table for %s." mode))))) ;; Utility functions for running a derived mode. (defun derived-mode-set-keymap (mode) "Set the keymap of the new MODE, maybe merging with the parent." (let* ((map-name (derived-mode-map-name mode)) (new-map (eval map-name)) (old-map (current-local-map))) (and old-map (get map-name 'derived-mode-unmerged) (derived-mode-merge-keymaps old-map new-map)) (put map-name 'derived-mode-unmerged nil) (use-local-map new-map))) (defun derived-mode-set-syntax-table (mode) "Set the syntax table of the new MODE, maybe merging with the parent." (let* ((table-name (derived-mode-syntax-table-name mode)) (old-table (syntax-table)) (new-table (eval table-name))) (if (get table-name 'derived-mode-unmerged) (derived-mode-merge-syntax-tables old-table new-table)) (put table-name 'derived-mode-unmerged nil) (set-syntax-table new-table))) (defun derived-mode-set-abbrev-table (mode) "Set the abbrev table for MODE if it exists. Always merge its parent into it, since the merge is non-destructive." (let* ((table-name (derived-mode-abbrev-table-name mode)) (old-table local-abbrev-table) (new-table (eval table-name))) (derived-mode-merge-abbrev-tables old-table new-table) (setq local-abbrev-table new-table))) ;;;(defun derived-mode-run-setup-function (mode) ;;; "Run the setup function if it exists." ;;; (let ((fname (derived-mode-setup-function-name mode))) ;;; (if (fboundp fname) ;;; (funcall fname)))) (defun derived-mode-run-hooks (mode) "Run the mode hook for MODE." (let ((hooks-name (derived-mode-hook-name mode))) (if (boundp hooks-name) (run-hooks hooks-name)))) ;; Functions to merge maps and tables. (defun derived-mode-merge-keymaps (old new) "Merge an OLD keymap into a NEW one. The old keymap is set to be the last cdr of the new one, so that there will be automatic inheritance." ;; XEmacs change. FSF 19.30 to 21.3 has a whole bunch of weird crap here ;; for merging prefix keys and such. Hopefully none of this is ;; necessary in XEmacs. (set-keymap-parents new (list old))) (defun derived-mode-merge-syntax-tables (old new) "Merge an OLD syntax table into a NEW one. Where the new table already has an entry, nothing is copied from the old one." ;; XEmacs change: on the other hand, Emacs 21.3 just has ;; (set-char-table-parent new old) here. ;; We use map-char-table, not map-syntax-table, so we can explicitly ;; check for inheritance. (map-char-table #'(lambda (key value) (let ((newval (get-range-char-table key new 'multi))) (cond ((eq newval 'multi) ; OK, dive into the class hierarchy (map-char-table #'(lambda (key1 value1) (when (eq ?@ (char-syntax-from-code (get-range-char-table key new ?@))) (put-char-table key1 value new)) nil) new key)) ((eq ?@ (char-syntax-from-code newval)) ;; class at once (put-char-table key value new)))) nil) old)) ;; Merge an old abbrev table into a new one. ;; This function requires internal knowledge of how abbrev tables work, ;; presuming that they are obarrays with the abbrev as the symbol, the expansion ;; as the value of the symbol, and the hook as the function definition. (defun derived-mode-merge-abbrev-tables (old new) (if old (mapatoms #'(lambda (symbol) (or (intern-soft (symbol-name symbol) new) (define-abbrev new (symbol-name symbol) (symbol-value symbol) (symbol-function symbol)))) old))) (provide 'derived) ;;; arch-tag: 630be248-47d1-4f02-afa0-8207de0ebea0 ;;; derived.el ends here