Mercurial > hg > xemacs-beta
comparison src/gc.c @ 5168:cf900a2f1fa3
extract gap array from extents.c, use in range tables
-------------------- ChangeLog entries follow: --------------------
src/ChangeLog addition:
2010-03-22 Ben Wing <ben@xemacs.org>
* Makefile.in.in (objs):
* array.c:
* array.c (gap_array_adjust_markers):
* array.c (gap_array_move_gap):
* array.c (gap_array_make_gap):
* array.c (gap_array_insert_els):
* array.c (gap_array_delete_els):
* array.c (gap_array_make_marker):
* array.c (gap_array_delete_marker):
* array.c (gap_array_delete_all_markers):
* array.c (gap_array_clone):
* array.h:
* depend:
* emacs.c (main_1):
* extents.c:
* extents.c (EXTENT_GAP_ARRAY_AT):
* extents.c (extent_list_num_els):
* extents.c (extent_list_locate):
* extents.c (extent_list_at):
* extents.c (extent_list_delete_all):
* extents.c (allocate_extent_list):
* extents.c (syms_of_extents):
* extents.h:
* extents.h (XEXTENT_LIST_MARKER):
* lisp.h:
* rangetab.c:
* rangetab.c (mark_range_table):
* rangetab.c (print_range_table):
* rangetab.c (range_table_equal):
* rangetab.c (range_table_hash):
* rangetab.c (verify_range_table):
* rangetab.c (get_range_table_pos):
* rangetab.c (Fmake_range_table):
* rangetab.c (Fcopy_range_table):
* rangetab.c (Fget_range_table):
* rangetab.c (put_range_table):
* rangetab.c (Fclear_range_table):
* rangetab.c (Fmap_range_table):
* rangetab.c (unified_range_table_bytes_needed):
* rangetab.c (unified_range_table_copy_data):
* rangetab.c (unified_range_table_lookup):
* rangetab.h:
* rangetab.h (struct range_table_entry):
* rangetab.h (struct Lisp_Range_Table):
* rangetab.h (rangetab_gap_array_at):
* symsinit.h:
Rename dynarr.c to array.c. Move gap array from extents.c to array.c.
Extract dynarr, gap array and stack-like malloc into new file array.h.
Rename GAP_ARRAY_NUM_ELS -> gap_array_length(). Add gap_array_at(),
gap_array_atp().
Rewrite range table code to use gap arrays. Make put_range_table()
smarter so that its operation is O(log n) for adding a localized
range.
* gc.c (lispdesc_block_size_1):
Don't ABORT() when two elements are located at the same place.
This will happen with a size-0 gap array -- both parts of the array
(before and after gap) are in the same place.
| author | Ben Wing <ben@xemacs.org> |
|---|---|
| date | Mon, 22 Mar 2010 19:12:15 -0500 |
| parents | 9e0b43d3095c |
| children | 6c6d78781d59 |
comparison
equal
deleted
inserted
replaced
| 5167:e374ea766cc1 | 5168:cf900a2f1fa3 |
|---|---|
| 555 for (pos = 0; desc[pos].type != XD_END; pos++) | 555 for (pos = 0; desc[pos].type != XD_END; pos++) |
| 556 { | 556 { |
| 557 EMACS_INT offset = lispdesc_indirect_count (desc[pos].offset, desc, obj); | 557 EMACS_INT offset = lispdesc_indirect_count (desc[pos].offset, desc, obj); |
| 558 if (offset == max_offset) | 558 if (offset == max_offset) |
| 559 { | 559 { |
| 560 #if 0 | |
| 561 /* This can legitimately happen with gap arrays -- if there are | |
| 562 no elements in the array, and the gap size is 0, then both | |
| 563 parts of the array will be of size 0 and in the same place. */ | |
| 560 stderr_out ("Two relocatable elements at same offset?\n"); | 564 stderr_out ("Two relocatable elements at same offset?\n"); |
| 561 ABORT (); | 565 ABORT (); |
| 566 #endif | |
| 562 } | 567 } |
| 563 else if (offset > max_offset) | 568 else if (offset > max_offset) |
| 564 { | 569 { |
| 565 max_offset = offset; | 570 max_offset = offset; |
| 566 max_offset_pos = pos; | 571 max_offset_pos = pos; |