Mercurial > hg > xemacs-beta
annotate lisp/code-process.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 | 0b060ef35789 |
| children | 91b3aa59f49b |
| rev | line source |
|---|---|
| 428 | 1 ;;; code-process.el --- Process coding functions for XEmacs. |
| 2 | |
| 3 ;; Copyright (C) 1985-1987, 1993, 1994, 1997 Free Software Foundation, Inc. | |
| 853 | 4 ;; Copyright (C) 1995, 2000, 2002 Ben Wing |
| 428 | 5 ;; Copyright (C) 1997 MORIOKA Tomohiko |
| 6 | |
| 7 ;; Author: Ben Wing | |
| 8 ;; MORIOKA Tomohiko | |
| 9 ;; Maintainer: XEmacs Development Team | |
| 10 ;; Keywords: mule, multilingual, coding system, process | |
| 11 | |
| 12 ;; This file is part of XEmacs. | |
| 13 | |
| 14 ;; This file is very similar to code-process.el | |
| 15 | |
| 16 ;; XEmacs is free software; you can redistribute it and/or modify it | |
| 17 ;; under the terms of the GNU General Public License as published by | |
| 18 ;; the Free Software Foundation; either version 2, or (at your option) | |
| 19 ;; any later version. | |
| 20 | |
| 21 ;; XEmacs is distributed in the hope that it will be useful, but | |
| 22 ;; WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 23 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
| 24 ;; General Public License for more details. | |
| 25 | |
| 26 ;; You should have received a copy of the GNU General Public License | |
| 27 ;; along with XEmacs; see the file COPYING. If not, write to the Free | |
| 28 ;; Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA | |
| 29 ;; 02111-1307, USA. | |
| 30 | |
| 31 ;;; Code: | |
| 32 | |
| 33 (defvar process-coding-system-alist nil | |
| 34 "Alist to decide a coding system to use for a process I/O operation. | |
| 35 The format is ((PATTERN . VAL) ...), | |
| 36 where PATTERN is a regular expression matching a program name, | |
| 37 VAL is a coding system, a cons of coding systems, or a function symbol. | |
| 38 If VAL is a coding system, it is used for both decoding what received | |
| 39 from the program and encoding what sent to the program. | |
| 40 If VAL is a cons of coding systems, the car part is used for decoding, | |
| 41 and the cdr part is used for encoding. | |
| 853 | 42 If VAL is a function symbol, it is called with two arguments, a symbol |
| 43 indicating the operation being performed (one of `start-process', | |
| 44 `call-process', `open-network-stream', or `open-multicast-group') and the | |
| 45 program name. The function must return a coding system or a cons of | |
| 46 coding systems which are used as above.") | |
| 428 | 47 |
| 48 (defun call-process (program &optional infile buffer displayp &rest args) | |
| 49 "Call PROGRAM synchronously in separate process. | |
| 853 | 50 |
| 428 | 51 The program's input comes from file INFILE (nil means `/dev/null'). |
| 853 | 52 XEmacs feature: INFILE can also be a list of (BUFFER [START [END]]), i.e. |
| 53 a list of one to three elements, consisting of a buffer and optionally | |
| 54 a start position or start and end position. In this case, input comes | |
| 55 from the buffer, starting from START (defaults to the beginning of the | |
| 56 buffer) and ending at END (defaults to the end of the buffer). | |
| 57 | |
| 428 | 58 Insert output in BUFFER before point; t means current buffer; |
| 59 nil for BUFFER means discard it; 0 means discard and don't wait. | |
| 837 | 60 If BUFFER is a string, then find or create a buffer with that name, |
| 61 then insert the output in that buffer, before point. | |
| 428 | 62 BUFFER can also have the form (REAL-BUFFER STDERR-FILE); in that case, |
| 63 REAL-BUFFER says what to do with standard output, as above, | |
| 64 while STDERR-FILE says what to do with standard error in the child. | |
| 65 STDERR-FILE may be nil (discard standard error output), | |
| 853 | 66 t (mix it with ordinary output), a file name string, or (XEmacs feature) |
| 67 a buffer object. If STDERR-FILE is a buffer object (but not the name of | |
| 68 a buffer, since that would be interpreted as a file), the standard error | |
| 69 output will be inserted into the buffer before point. | |
| 428 | 70 |
| 853 | 71 Fourth arg DISPLAY non-nil means redisplay buffer as output is inserted. |
| 428 | 72 Remaining arguments are strings passed as command arguments to PROGRAM. |
| 73 | |
| 853 | 74 If BUFFER is 0, returns immediately with value nil. |
| 75 Otherwise waits for PROGRAM to terminate and returns a numeric exit status | |
| 76 or a signal description string. If you quit, the process is first killed | |
| 77 with SIGINT, then with SIGKILL if you quit again before the process exits. | |
| 78 | |
| 79 If INFILE is a file, we transfer its exact contents to the process without | |
| 80 any encoding/decoding. (#### This policy might change.) | |
| 81 | |
| 82 Otherwise, the read/write coding systems used for process I/O on the | |
| 83 process are determined as follows: | |
| 84 | |
| 85 1. `coding-system-for-read', `coding-system-for-write', if non-nil. | |
| 86 (Intended as a temporary overriding mechanism for use by Lisp | |
| 87 code.) | |
| 88 2. The matching value for the process name from `process-coding-system-alist', | |
| 89 if any, and if non-nil. The value may be either a single coding | |
| 90 system, used for both read and write; or a cons of read/write; or a | |
| 91 function, called to get one of the other two values. | |
| 92 3. For writing: If a buffer was given in INFILE, the value of | |
| 93 `buffer-file-coding-system' in that buffer. | |
| 94 For reading: if a buffer was given in BUFFER, the value of | |
| 95 `buffer-file-coding-system-for-read' in that buffer. | |
| 96 4. The value of `default-process-coding-system', which should be a cons | |
| 97 of read/write coding systems, if the values are non-nil. | |
| 98 5. The coding system `undecided' for read, and `raw-text' for write. | |
| 99 | |
| 100 Note that the processes of determining the read and write coding systems | |
| 101 proceed essentially independently one from the other, as in `start-process'." | |
| 102 (let (cs-r cs-w) | |
| 103 (let (ret) | |
| 104 (catch 'found | |
| 105 (let ((alist process-coding-system-alist) | |
| 106 (case-fold-search nil)) | |
| 107 (while alist | |
| 108 (if (string-match (car (car alist)) program) | |
| 109 (throw 'found (setq ret (cdr (car alist))))) | |
| 110 (setq alist (cdr alist)) | |
| 111 ))) | |
| 112 (if (functionp ret) | |
| 113 (setq ret (funcall ret 'call-process program))) | |
| 114 (cond ((consp ret) | |
| 115 (setq cs-r (car ret) | |
| 116 cs-w (cdr ret))) | |
| 117 ((and ret (find-coding-system ret)) | |
| 118 (setq cs-r ret | |
| 119 cs-w ret)))) | |
| 120 (let ((coding-system-for-read | |
| 121 (or coding-system-for-read cs-r | |
| 122 (let ((thebuf (if (consp buffer) (car buffer) buffer))) | |
| 123 (and (or (bufferp thebuf) (stringp thebuf)) | |
| 124 (get-buffer thebuf) | |
| 125 (symbol-value-in-buffer | |
| 857 | 126 'buffer-file-coding-system-for-read (get-buffer thebuf)))) |
| 853 | 127 (car default-process-coding-system) |
| 128 'undecided)) | |
| 129 (coding-system-for-write | |
| 130 (or coding-system-for-write cs-w | |
| 131 (and (consp infile) | |
| 132 (symbol-value-in-buffer | |
| 133 'buffer-file-coding-system | |
| 134 (get-buffer (car infile)))) | |
| 135 (cdr default-process-coding-system) | |
| 136 'raw-text))) | |
| 137 (apply 'call-process-internal program infile buffer displayp args)))) | |
| 428 | 138 |
| 139 (defun call-process-region (start end program | |
| 140 &optional deletep buffer displayp | |
| 141 &rest args) | |
| 142 "Send text from START to END to a synchronous process running PROGRAM. | |
| 143 Delete the text if fourth arg DELETEP is non-nil. | |
| 144 | |
| 145 Insert output in BUFFER before point; t means current buffer; | |
| 146 nil for BUFFER means discard it; 0 means discard and don't wait. | |
| 837 | 147 If BUFFER is a string, then find or create a buffer with that name, |
| 148 then insert the output in that buffer, before point. | |
| 428 | 149 BUFFER can also have the form (REAL-BUFFER STDERR-FILE); in that case, |
| 150 REAL-BUFFER says what to do with standard output, as above, | |
| 151 while STDERR-FILE says what to do with standard error in the child. | |
| 152 STDERR-FILE may be nil (discard standard error output), | |
| 853 | 153 t (mix it with ordinary output), a file name string, or (XEmacs feature) |
| 154 a buffer object. If STDERR-FILE is a buffer object (but not the name of | |
| 155 a buffer, since that would be interpreted as a file), the standard error | |
| 156 output will be inserted into the buffer before point. | |
| 428 | 157 |
| 158 Sixth arg DISPLAYP non-nil means redisplay buffer as output is inserted. | |
| 159 Remaining args are passed to PROGRAM at startup as command args. | |
| 160 | |
| 161 If BUFFER is 0, returns immediately with value nil. | |
| 853 | 162 Otherwise waits for PROGRAM to terminate and returns a numeric exit status |
| 163 or a signal description string. If you quit, the process is first killed | |
| 164 with SIGINT, then with SIGKILL if you quit again before the process exits. | |
| 165 | |
| 166 The read/write coding systems used for process I/O on the process are | |
| 167 the same as for `call-process'." | |
| 2356 | 168 |
| 169 ;; We can't delete the region before feeding it to `call-process', so we | |
| 170 ;; take care not to delete the insertion when we delete the region. START | |
| 171 ;; and END may not be markers; copy them. (point) will end up after the | |
| 172 ;; insertion. A copy of (point) tracks the beginning of the insertion. | |
| 173 | |
| 174 (let ((s (and deletep (copy-marker start))) ; Only YOU can | |
| 175 (e (and deletep (copy-marker end t))) ; prevent | |
| 176 (p (and deletep (copy-marker (point)))) ; excess consing! | |
| 177 (retval | |
| 178 (apply #'call-process program (list (current-buffer) start end) | |
| 179 buffer displayp args))) | |
| 180 (when deletep | |
| 181 (if (<= s p e) | |
| 182 ;; region was split by insertion | |
| 183 ;; the order checks are gilt lilies | |
| 184 (progn (when (< (point) e) (delete-region (point) e)) | |
| 185 (when (< s p) (delete-region s p))) | |
| 186 ;; insertion was outside of region | |
| 187 (delete-region s e))) | |
| 188 retval)) | |
| 428 | 189 |
| 190 (defun start-process (name buffer program &rest program-args) | |
| 191 "Start a program in a subprocess. Return the process object for it. | |
| 192 NAME is name for process. It is modified if necessary to make it unique. | |
| 193 BUFFER is the buffer or (buffer-name) to associate with the process. | |
| 853 | 194 Process output goes at end of that buffer, unless you specify |
| 195 an output stream or filter function to handle the output. | |
| 196 BUFFER may be also nil, meaning that this process is not associated | |
| 197 with any buffer. | |
| 198 BUFFER can also have the form (REAL-BUFFER STDERR-BUFFER); in that case, | |
| 199 REAL-BUFFER says what to do with standard output, as above, | |
| 200 while STDERR-BUFFER says what to do with standard error in the child. | |
| 201 STDERR-BUFFER may be nil (discard standard error output, unless a stderr | |
| 202 filter is set). Note that if you do not use this form at process creation, | |
| 203 stdout and stderr will be mixed in the output buffer, and this cannot be | |
| 204 changed, even by setting a stderr filter. | |
| 428 | 205 Third arg is program file name. It is searched for as in the shell. |
| 206 Remaining arguments are strings to give program as arguments. | |
| 853 | 207 |
| 208 The read/write coding systems used for process I/O on the process are | |
| 209 determined as follows: | |
| 210 | |
| 211 1. `coding-system-for-read', `coding-system-for-write', if non-nil. | |
| 212 (Intended as a temporary overriding mechanism for use by Lisp | |
| 213 code.) | |
| 214 2. The matching value for the process name from `process-coding-system-alist', | |
| 215 if any, and if non-nil. The value may be either a single coding | |
| 216 system, used for both read and write; or a cons of read/write; or a | |
| 217 function, called to get one of the other two values. | |
| 218 3. The value of `default-process-coding-system', which should be a cons | |
| 219 of read/write coding systems, if the values are non-nil. | |
| 220 4. The coding system `undecided' for read, and `raw-text' for write. | |
| 221 | |
| 222 Note that the processes of determining the read and write coding systems | |
| 223 proceed essentially independently one from the other. For example, a value | |
| 224 determined from `process-coding-system-alist' might specify a read coding | |
| 225 system but not a write coding system, in which the read coding system is as | |
| 226 specified and the write coding system comes from proceeding to step 3 (and | |
| 227 looking in `default-process-coding-system'). | |
| 228 | |
| 229 You can change the coding systems later on using | |
| 230 `set-process-coding-system', `set-process-input-coding-system', or | |
| 231 `set-process-output-coding-system'. | |
| 232 | |
| 233 See also `set-process-filter' and `set-process-stderr-filter'." | |
| 428 | 234 (let (cs-r cs-w) |
| 235 (let (ret) | |
| 236 (catch 'found | |
| 237 (let ((alist process-coding-system-alist) | |
| 238 (case-fold-search nil)) | |
| 239 (while alist | |
| 240 (if (string-match (car (car alist)) program) | |
| 241 (throw 'found (setq ret (cdr (car alist))))) | |
| 242 (setq alist (cdr alist)) | |
| 243 ))) | |
| 244 (if (functionp ret) | |
| 245 (setq ret (funcall ret 'start-process program))) | |
| 246 (cond ((consp ret) | |
| 247 (setq cs-r (car ret) | |
| 248 cs-w (cdr ret))) | |
| 853 | 249 ((and ret (find-coding-system ret)) |
| 428 | 250 (setq cs-r ret |
| 251 cs-w ret)))) | |
| 252 (let ((coding-system-for-read | |
| 771 | 253 (or coding-system-for-read cs-r |
| 853 | 254 (car default-process-coding-system) 'undecided)) |
| 428 | 255 (coding-system-for-write |
| 771 | 256 (or coding-system-for-write cs-w |
| 853 | 257 (cdr default-process-coding-system) 'raw-text))) |
| 428 | 258 (apply 'start-process-internal name buffer program program-args) |
| 259 ))) | |
| 260 | |
| 261 (defvar network-coding-system-alist nil | |
| 262 "Alist to decide a coding system to use for a network I/O operation. | |
| 263 The format is ((PATTERN . VAL) ...), | |
| 264 where PATTERN is a regular expression matching a network service name | |
| 265 or is a port number to connect to, | |
| 266 VAL is a coding system, a cons of coding systems, or a function symbol. | |
| 267 If VAL is a coding system, it is used for both decoding what received | |
| 268 from the network stream and encoding what sent to the network stream. | |
| 269 If VAL is a cons of coding systems, the car part is used for decoding, | |
| 270 and the cdr part is used for encoding. | |
| 271 If VAL is a function symbol, the function must return a coding system | |
| 272 or a cons of coding systems which are used as above. | |
| 273 | |
| 274 See also the function `find-operation-coding-system'.") | |
| 275 | |
| 276 (defun open-network-stream (name buffer host service &optional protocol) | |
| 277 "Open a TCP connection for a service to a host. | |
| 444 | 278 Return a process object to represent the connection. |
| 428 | 279 Input and output work as for subprocesses; `delete-process' closes it. |
| 280 NAME is name for process. It is modified if necessary to make it unique. | |
| 281 BUFFER is the buffer (or buffer-name) to associate with the process. | |
| 282 Process output goes at end of that buffer, unless you specify | |
| 283 an output stream or filter function to handle the output. | |
| 284 BUFFER may be also nil, meaning that this process is not associated | |
| 853 | 285 with any buffer. |
| 428 | 286 Third arg is name of the host to connect to, or its IP address. |
| 287 Fourth arg SERVICE is name of the service desired, or an integer | |
| 288 specifying a port number to connect to. | |
| 289 Fifth argument PROTOCOL is a network protocol. Currently 'tcp | |
| 290 (Transmission Control Protocol) and 'udp (User Datagram Protocol) are | |
| 291 supported. When omitted, 'tcp is assumed. | |
| 292 | |
| 442 | 293 Output via `process-send-string' and input via buffer or filter (see |
| 428 | 294 `set-process-filter') are stream-oriented. That means UDP datagrams are |
| 295 not guaranteed to be sent and received in discrete packets. (But small | |
| 296 datagrams around 500 bytes that are not truncated by `process-send-string' | |
| 297 are usually fine.) Note further that UDP protocol does not guard against | |
| 853 | 298 lost packets. |
| 299 | |
| 300 The read/write coding systems used for process I/O on the process are | |
| 301 determined as follows: | |
| 302 | |
| 303 1. `coding-system-for-read', `coding-system-for-write', if non-nil. | |
| 304 (Intended as a temporary overriding mechanism for use by Lisp | |
| 305 code.) | |
| 306 2. The matching value for the service from `network-coding-system-alist', | |
| 307 if any, and if non-nil. The value may be either a single coding | |
| 308 system, used for both read and write; or a cons of read/write; or a | |
| 309 function, called to get one of the other two values. | |
| 310 3. The value of `default-network-coding-system', which should be a cons | |
| 311 of read/write coding systems, if the values are non-nil. | |
| 312 4. The coding system `undecided' for read, and `raw-text' for write. | |
| 313 | |
| 314 Note that the processes of determining the read and write coding systems | |
| 315 proceed essentially independently one from the other, as in `start-process'. | |
| 316 | |
| 317 You can change the coding systems later on using | |
| 318 `set-process-coding-system', `set-process-input-coding-system', or | |
| 319 `set-process-output-coding-system'." | |
| 428 | 320 (let (cs-r cs-w) |
| 321 (let (ret) | |
| 322 (catch 'found | |
| 323 (let ((alist network-coding-system-alist) | |
| 324 (case-fold-search nil) | |
| 325 pattern) | |
| 326 (while alist | |
| 327 (setq pattern (car (car alist))) | |
| 328 (and | |
| 329 (cond ((numberp pattern) | |
| 330 (and (numberp service) | |
| 331 (eq pattern service))) | |
| 332 ((stringp pattern) | |
| 333 (or (and (stringp service) | |
| 334 (string-match pattern service)) | |
| 335 (and (numberp service) | |
| 336 (string-match pattern | |
| 337 (number-to-string service)))))) | |
| 338 (throw 'found (setq ret (cdr (car alist))))) | |
| 339 (setq alist (cdr alist)) | |
| 340 ))) | |
| 341 (if (functionp ret) | |
| 342 (setq ret (funcall ret 'open-network-stream service))) | |
| 343 (cond ((consp ret) | |
| 344 (setq cs-r (car ret) | |
| 345 cs-w (cdr ret))) | |
| 853 | 346 ((and ret (find-coding-system ret)) |
| 428 | 347 (setq cs-r ret |
| 348 cs-w ret)))) | |
| 349 (let ((coding-system-for-read | |
| 853 | 350 (or coding-system-for-read cs-r |
| 351 (car default-network-coding-system) | |
| 352 'undecided)) | |
| 428 | 353 (coding-system-for-write |
| 853 | 354 (or coding-system-for-write cs-w |
| 355 (cdr default-network-coding-system) | |
| 356 'raw-text))) | |
| 428 | 357 (open-network-stream-internal name buffer host service protocol)))) |
| 358 | |
| 771 | 359 (defun set-buffer-process-coding-system (decoding encoding) |
| 360 "Set coding systems for the process associated with the current buffer. | |
| 361 DECODING is the coding system to be used to decode input from the process, | |
| 362 ENCODING is the coding system to be used to encode output to the process. | |
| 363 | |
| 853 | 364 For a list of possible values of CODING-SYSTEM, use \\[coding-system-list]." |
| 771 | 365 (interactive |
| 366 "zCoding-system for process input: \nzCoding-system for process output: ") | |
| 367 (let ((proc (get-buffer-process (current-buffer)))) | |
| 368 (if (null proc) | |
| 369 (error "no process") | |
| 370 (get-coding-system decoding) | |
| 371 (get-coding-system encoding) | |
| 372 (set-process-coding-system proc decoding encoding))) | |
| 373 (force-mode-line-update)) | |
| 374 | |
| 440 | 375 ;;; code-process.el ends here |