Merge from emacs--devo--0

[gnu-emacs] / src / ChangeLog.unicode

2007-06-06  Jason Rumney  <jasonr@gnu.org>

        * w32font.c (w32font_draw): Delete brush after using it.

2007-06-04  Jason Rumney  <jasonr@gnu.org>

        * w32font.c (w32font_open): Don't set font_idx.
        (w32font_text_extents): Try GetTextExtentPoint32W before defaulting
        to font settings.
        (w32font_draw): Fill background explicitly.

2007-06-03  Jason Rumney  <jasonr@gnu.org>

        * w32term.c (w32_initialize): Don't call w32font_initialize.

        * w32font.c (w32font_info): Remove subranges.
        (QCsubranges, Qmodern, Qswiss, Qroman): Remove.
        (QCfamily, Qmonospace, Qsans_serif, Qmono, Qsans, Qsans__serif)
        (Qraster, Qoutline, Qlatin, Qgreek, Qcoptic, Qcyrillic, Qarmenian)
        (Qhebrew, Qarabic, Qsyriac, Qnko, Qthaana, Qdevanagari, Qbengali)
        (Qgurmukhi, Qgujarati, Qoriya, Qtamil, Qtelugu, Qkannada)
        (Qmalayalam, Qsinhala, Qthai, Qlao, Qtibetan, Qmyanmar, Qgeorgian)
        (Qhangul, Qethiopic, Qcherokee, Qcanadian_aboriginal, Qogham)
        (Qrunic, Qkhmer, Qmongolian, Qsymbol, Qbraille, Qhan)
        (Qideographic_description, Qcjk_misc, Qkana, Qbopomofo, Qkanbun)
        (Qyi, Qbyzantine_musical_symbol, Qmusical_symbol, Qmathematical):
        New symbols.
        (font_callback_data): New struct.
        (w32font_list, w32font_match): Use it.
        (w32font_open): Don't populate subranges.
        (w32font_has_char): Use script Lisp symbols, not subrange bitmask.
        (w32font_encode_char): Always return unicode code-point as-is.
        (w32font_text_extents): Supply a tranformation matrix to
        GetGlyphOutline. Never look up by glyph index. Avoid looping
        twice. Use unicode version of GetTexExtentPoint32 instead of
        glyph index version.
        (set_fonts_frame): Remove
        (w32_enumfont_pattern_entity): Add frame parameter, use it to
        set frame parameter. Use backward compatible fake foundries.
        Save generic family in extra slot under QCfamily. Make width slot
        constant. Save QCspacing value. Save list of scripts instead of
        binary subranges.
        (w32_generic_family, logfonts_match, font_matches_spec): New functions.
        (add_font_entity_to_list): Use font_callback_data struct. Filter
        unwanted fonts.
        (add_one_font_entity_to_list): Use font_callback_data struct.
        (w32_registry): Default to iso10646_1;
        (fill_in_logfont): Use dpi from extra slot. Don't bother with
        string font registries. Don't fill in font name if it is a generic
        family name, fill family instead. Use spacing, family and script
        extra info to fill pitch, family and charset fields.
        (list_all_matching_fonts): Use font_callback_data struct.
        (unicode_range_for_char): Remove.
        (font_supported_scripts): New function.
        (w32font_initialize): Remove.
        (syms_of_w32font): Update which symbols are defined.

2007-06-02  Jason Rumney  <jasonr@gnu.org>

        * font.c (font_pixel_size): Reverse assq_no_quit args.

        * w32term.h (FONT_WIDTH): Report max width, not average.
        (FONT_MAX_WIDTH): Remove.
        (FONT_AVG_WIDTH): New macro.

        * xfaces.c (Fx_list_fonts) [WINDOWSNT]: Remove Windows only
        redefinition of FONT_WIDTH.

        * w32term.c (x_font_min_bounds): Use FONT_AVG_WIDTH.
        (w32_cache_char_metrics): Use FONT_WIDTH.

        * w32fns.c (w32_load_system_font, w32_list_fonts): Use FONT_AVG_WIDTH.

2007-06-01  Jason Rumney  <jasonr@gnu.org>

        * w32font.c (w32font_open): Make lfHeight negative.

        * w32fns.c (x_default_font_parameter): Use new style font name.
        (Fx_create_frame, x_create_tip_frame): Initialize resx and resy.

2007-05-31  Jason Rumney  <jasonr@gnu.org>

        * w32font.c (QCsubranges): New symbol.
        (w32font_open, w32font_has_char): Get subranges from subproperty
        of extra.
        (w32_enumfont_pattern_entity): Set subranges as subproperty of extra.
        (syms_of_w32font): Define :subranges symbol.

        * font.c (font_put_extra): Expose externally.

        * font.h (font_put_extra): Moved declaration from font.c.

        * font.c (Ffont_get): Use font driver to determine otf capability.
        (adjust_anchor): Check if driver defines anchor_point before using.

        * w32font.c (w32font_open): Handle size, height and pixel_size better.
        (w32font_draw): Use options.
        (w32_enumfont_pattern_entity): Set size to 0 for scalable fonts.
        Fix detection of truetype fonts.
        (registry_to_w32_charset): Handle charsets other than iso8859-1
        expressed as lisp symbols.
        (w32_registry): Express charset as lisp symbol.
        (fill_in_logfont): Reverse pixel and point height logic.
        Don't set width here. Set quality to default.

        * w32fns.c (w32_load_system_font): Fix detecting FIXED_PITCH fonts.
        (x_to_w32_font): Fill in lfPitchAndFamily correctly.

        * xterm.c (x_draw_glyph_string_foreground) [USE_FONT_BACKEND]:
        Remove redundant loop and allocation.

        * makefile.w32-in (font.o, w32font.o): New objects.
        (fontset.o, xdisp.o, xfaces.o, w32fns.o, w32term.o): Depend on font.h
        (FONTOBJ): New group of objects conditioned on USE_FONT_BACKEND.

        * xdisp.c (fill_composite_glyph_string): Make the first arg to
        STORE_XCHARB a valid l-value.

        * w32term.c (w32_native_per_char_metric): Swap width and rbearing
        calculations for non-Truetype fonts.
        (x_draw_glyph_string): Sync with xterm.c.
        (x_draw_glyph_string_foreground) [USE_FONT_BACKEND]: Remove
        redundant code.
        (w32_initialize) [USE_FONT_BACKEND]: Call w32font_initialize.

        * w32term.h (w32_output_data) [USE_FONT_BACKEND]: Add fontp member.
        (FRAME_FONT_OBJECT) [USE_FONT_BACKEND]: New macro from xterm.h.

        * w32fns.c [USE_FONT_BACKEND]: Port font backend changes from xfns.c.
        (x_to_w32_charset, w32_to_x_charset): Expose externally.

        * w32font.c: New file for w32 font backend.     

2007-05-28  Kenichi Handa  <handa@m17n.org>

        * term.c: Don't include "buffer.h" twice.

2007-05-21  Kenichi Handa  <handa@m17n.org>

        * syntax.c (skip_syntaxes): Synch with the trunk.

2007-05-14  Kenichi Handa  <handa@m17n.org>

        * character.c (Funibyte_string): New function.
        (syms_of_character): Defsubr it.

2007-05-01  Jason Rumney  <jasonr@gnu.org>

        * w32term.c [USE_FONT_BACKEND]:
        (x_get_font_repertory, note_mouse_movement, x_set_mouse_face_gc):
        (x_set_glyph_string_clipping, x_set_glyph_string_clipping_exactly):
        (x_draw_glyph_string, x_draw_glyph_string_foreground):
        (x_draw_composite_glyph_string_foreground, x_new_fontset2):
        (x_free_frame_resources): Sync with xterm.c.

2007-04-29  Andreas Schwab  <schwab@suse.de>

        * lread.c (read1): Use CHAR_TABLE_STANDARD_SLOTS to validate
        char-table size.

2007-04-26  Kenichi Handa  <handa@m17n.org>

        * font.c (check_otf_features): Define it regardless of
        HAVE_LIBOTF.

2007-04-25  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_driver): Delete font_otf_gsub and
        font_otf_gpos, add font_drive_otf.

        * fontset.c (fontset_find_font): Pay attention to font size
        specified for a font.
        (reorder_font_vector): Check contents of font_def.

        * font.c (struct otf_list): Delete it.
        (otf_list): Make it a lisp variable..
        (otf_open): Use lispy otf_list.
        (generate_otf_features): Renamed from parse_gsub_gpos_spec.
        (check_otf_features): New function.
        (font_otf_DeviceTable, font_otf_ValueRecord, font_otf_Anchor): New
        functinos.
        (font_drive_otf): New function merging font_otf_gsub and
        font_otf_gpos.
        (font_open_for_lface): New arg spec.  Change argument order.
        (font_load_for_face): Adjusted for the change of
        font_open_for_lface.
        (Ffont_drive_otf): New function merging Ffont_otf_gsub and
        Ffont_otf_gpos.
        (syms_of_font): Staticpro otf_list.  Delete defsubr of
        Sfont_otf_gsub and Sfont_otf_gpos.  Defsubr Sfont_drive_otf.

        * xfaces.c (set_font_frame_param): Adjusted for the change of
        font_open_for_lface.

        * font.h (font_open_for_lface): Adjust prototype.
        (struct font_driver): Delete members otf_gsub and otf_gpos, add
        member otf_drive.
        (font_otf_gsub, font_otf_gpos): Delete externs.
        (font_drive_otf): Extern it.

2007-04-24  Kenichi Handa  <handa@m17n.org>

        * font.c (font_at): If the window W is not on a window system,
        return Qnil.

        * coding.c (produce_chars): Don't call insert_from_gap if no
        characters to produce.
        (encode_coding): Likewise.

2007-04-17  Kenichi Handa  <handa@m17n.org>

        * fontset.c (free_realized_fontsets): Avoid unnecessary call of
        Fclear_face_cache.

        * xfaces.c (face_for_font): Check also face->font==font->font.font.

2007-04-16  Miles Bader  <miles@gnu.org>

        * emacs.c (main): Change default value of `enable_font_backend' to 1.
        Parse "--disable-font-backend" option.
        (standard_args): Add "--disable-font-backend" option.

2007-04-13  Kenichi Handa  <handa@m17n.org>

        * fontset.c (fontset_find_font): New function.
        (fontset_font): Use fontset_find_font.
        (make_fontset_for_ascii_face): Don't set face ID in rfont_def.
        Register the specified font for all Latin characters.
        (new_fontset_from_font): Register the specified font for all Latin
        characters.
        (dump_fontset): For a realized fontset, include the base fontset
        name in the returned vector.

2007-04-11  Kenichi Handa  <handa@m17n.org>

        * character.h (CHAR_STRING): Cast C to unsigned on calling
        char_string.

        * character.c (char_string): Type of arg C changed to unsigned.
        Signal an error if C is an invalid character code.

        * editfns.c (general_insert_function): Use CHARACTERP, not INTEGERP.
        (Fchar_to_string): Likewise.

2007-03-23  Kenichi Handa  <handa@m17n.org>

        * character.h (MIN_MULTIBYTE_LEADING_CODE)
        (MAX_MULTIBYTE_LEADING_CODE): New macros.

        * regex.c (analyse_first): Fix for multibyte characters in "case
        charset:" and "case categoryspec:".

2007-03-04  Andreas Schwab  <schwab@suse.de>

        * Makefile.in (LIBES): Move standard libraries to the end.

2007-02-27  Kenichi Handa  <handa@m17n.org>

        * process.c: Cancel the change done by sync with HEAD.

2007-02-23  Kenichi Handa  <handa@m17n.org>

        * alloc.c (Fgarbage_collect): If nextb->text->inhibit_shrinking is
        nonzero, don't shrink the buffer nextb.

        * buffer.h (struct buffer_text): New member inhibit_shrinking.

        * coding.c (coding_alloc_by_making_gap): New arg offset.
        (alloc_destination): Call coding_alloc_by_making_gap with the arg
        offset.
        (decode_coding_iso_2022): Update coding->safe_charsets.
        (decode_coding_gap): Temporarily set
        current_buffer->text->inhibit_shrinking to 1.

2007-02-20  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_draw_composite_glyph_string_foreground): Fix 
        indexing into elements of s->cmp and s->char2b.

2007-02-16  Juanma Barranquero  <lekktu@gmail.com>

        * regex.c (RE_STRING_CHAR_AND_LENGTH) [! emacs]: Add missing arg `len'.

2007-02-16  Kenichi Handa  <handa@m17n.org>

        * regex.c (GET_CHAR_BEFORE_2, GET_CHAR_AFTER): Check the variable
        target_multibyte instead of multibyte.
        (re_match_2_internal): Call bcmp_translate with target_multibyte.
        (bcmp_translate): Change the argument name from multibyte to
        target_multibyte.

2007-02-15  Kenichi Handa  <handa@m17n.org>

        These changes are to compile a regexp into a pattern that can be
        used both for multibyte and unibyte targets.

        * Makefile.in (search.o): Depend on charset.h.

        * character.c (multibyte_char_to_unibyte_safe): New function.

        * search.c: Include "charset.h".
        (compile_pattern_1): Delete argument multibyte.  Don't set
        cp->buf.target_multibyte here.  Set cp->buf.charset_unibyte.
        (compile_pattern): Don't compare cp->buf.target_multibyte.
        Compare cp->buf.charset_unibyte.
        (compile_pattern): Set cp->buf.target_multibyte.

        * lisp.h (multibyte_char_to_unibyte_safe): Extern it.

        * regex.h (struct re_pattern_buffer): New member charset_unibyte.

        * regex.c (RE_STRING_CHAR, RE_STRING_CHAR_AND_LENGTH): New arg
        multibyte.  Callers changed.
        (RE_CHAR_TO_MULTIBYTE, RE_CHAR_TO_UNIBYTE): New macros.
        (MAKE_CHAR_MULTIBYTE, MAKE_CHAR_UNIBYTE): Deleted.  Callers
        changed to use RE_CHAR_TO_MULTIBYTE and RE_CHAR_TO_UNIBYTE
        respectively.
        (SETUP_ASCII_RANGE, SETUP_UNIBYTE_RANGE): New macros.
        (SETUP_MULTIBYTE_RANGE): Generate a more compact range_table.
        (regex_compile): Make the compiled pattern usable both for
        multibyte and unibyte targets.
        (analyse_first): Make the fastmap usable both for multibyte and
        unibyte targets.
        (TRANSLATE_VIA_MULTIBYTE): Deleted.
        (re_match_2_internal): Pay attention to the case that the
        multibyteness of bufp and target may be different.

2007-02-14  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (x_produce_glyphs): When a font is not found, make the
        empty box occupy at least one column width.

2007-01-30  Miles Bader  <miles@gnu.org>

        * Makefile.in: Remove redundant HAVE_XFT clause.

2006-12-20  Kenichi Handa  <handa@m17n.org>

        * xrdb.c (x_load_resources): Setup the default fontSet X reource.

2006-12-15  Kenichi Handa  <handa@m17n.org>

        * regex.c (regex_compile): Synch with HEAD.

2006-12-12  Kenichi Handa  <handa@m17n.org>

        * fontset.c (Finternal_char_font): Fix previous change.

2006-12-07  Kenichi Handa  <handa@m17n.org>

        * fontset.c (Finternal_char_font): Fix for the case of POSITION
        being nil.

2006-12-06  Kenichi Handa  <handa@m17n.org>

        * xftfont.c (xftfont_open): Call FcConfigSubstitute.

2006-12-05  Kenichi Handa  <handa@m17n.org>

        * xftfont.c (xftfont_open): Don't enable antialias explicitly.

2006-11-30  Kenichi Handa  <handa@m17n.org>

        * search.c (simple_search): Fix previous change.

2006-11-29  Kenichi Handa  <handa@m17n.org>

        * xftfont.c (ftfont_font_format): Extern declaration.

        * frame.c (x_set_font): Fix the second arg to fs_query_fontset.

        * xfont.c (xfont_driver): Initialize ftfont_driver.type by 0.
        (xfont_list): Don't directly use Lisp_Object as an operand of &&.

        * ftfont.c (ftfont_driver): Initialize ftfont_driver.type by 0.
        (ftfont_font_format): Fix previous change.

        * font.h (Ffont_xlfd_name): EXFUN it.

        * font.c (font_parse_xlfd): Fix the array size of `f'.
        (register_font_driver): Use EQ to compare driver->type.

        * xfns.c (xic_create_xfontset2) [USE_FONT_BACKEND]: New function.
        (create_frame_xic) [USE_FONT_BACKEND]: Call xic_create_xfontset2.
        (xic_set_xfontset) [USE_FONT_BACKEND]: Likewise.

2006-11-27  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_pattern_entity): Check if FC_FONTFORMAT is
        defined.
        (ftfont_list_generic_family, ftfont_list, ftfont_font_format):
        Likewise.

2006-11-24  Kenichi Handa  <handa@m17n.org>

        * xfont.c (xfont_open): Set font->format.

        * xftfont.c (xftfont_open): Set font->format.

        * ftfont.c (ftfont_pattern_entity): Add fontformat in a pattern.
        (ftfont_list): Include FC_FONTFORMAT in FcObject.
        (ftfont_open): Set font->format.
        (ftfont_font_format): New function.

        * font.h (struct font): New memeber format.

        * font.c (Qopentype): New variable.
        (syms_of_font): Defsym it.
        (Fquery_font): Change the format of the last element of the return
        value.

2006-11-22  Kenichi Handa  <handa@m17n.org>

        * search.c (simple_search): Fix sync with HEAD.

2006-11-17  Kenichi Handa  <handa@m17n.org>

        * xfns.c (xic_create_xfontset): Try the default fontset name as a
        last resort.

2006-11-08  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding_charset): Fix detection of multi-byte
        charset.

2006-11-07  Bob Halley  <halley@play-bow.org>  (tiny change)

        * ccl.c (ccl_driver): If DST is NULL, set ccl->produced to 0.

2006-10-31  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (get_next_display_element): Set it->face_id for the
        first component of a composition.
        (x_produce_glyphs): Check if the font is changed or not for
        composition.

2006-10-31  Kenichi Handa  <handa@m17n.org>

        (get_next_display_element): Set it->face_id for the
        first component of a composition.
        (x_produce_glyphs): Check if the font is changed or not for
        composition.

2006-10-30  Kenichi Handa  <handa@m17n.org>

        * fontset.c (Qlatin): New variable.
        (syms_of_fontset): Define it as a lisp symbol.
        (Fset_fontset_font): If TARGET is `latin', use FONT_SPEC for
        ASCII.

2006-10-27  Kenichi Handa  <handa@m17n.org>

        * font.c (font_unparse_fcname): Pay attention to the case that
        some of font property is a null string.

2006-10-26  Kenichi Handa  <handa@m17n.org>

        * term.c: Include "composite.h".
        (encode_terminal_code): Output all components of composition.
        Check the size of encode_terminal_src.
        (produce_glyphs): For compostion, call produce_composite_glyph.
        (append_composite_glyph, produce_composite_glyph): New functions.

        * xdisp.c (x_produce_glyphs): In handling composition, if a font
        is not found, get font_info from the current ascii face.

2006-10-23  Kenichi Handa  <handa@m17n.org>

        * fileio.c (Finsert_file_contents): On replacing, temporarily bind
        buffer-file-name to Qnil before calling insert_from_buffer.

        * font.c (font_unparse_fcname): Pay attention to the case that
        foundry is a null string.

2006-10-17  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_list): Allow registry "unicode-sip".

        * font.c (Qunicode_sip): New variable.
        (syms_of_font): Declare it as a Lisp symbol.

        * font.h (Qunicode_sip): Extern it.

2006-10-16  Kenichi Handa  <handa@m17n.org>

        * composite.c (get_composition_id): Pay attention to TAB
        component.

        * xterm.c (x_draw_composite_glyph_string_foreground): Don't draw
        TAB.  Adjusted for the change of s->char2b which always points to
        the first elememnt of allocated memory.

        * xftfont.c (xftfont_text_extents): Fix calculation of descent
        value.

        * xdisp.c (handle_composition_prop): Set it->c to the first
        non-TAB component.
        (fill_composite_glyph_string): Argument changed.
        (BUILD_COMPOSITE_GLYPH_STRING): Adjusted for the above change.
        (x_produce_glyphs): Fix handling of left/right padding.

2006-10-04  Kenichi Handa  <handa@m17n.org>

        * regex.c (analyse_first): Cancel the change for synching with
        HEAD.

2006-08-28  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding_system): Fix for handling off
        inhibit_iso_escape_detection.  Fix for the case that no coding
        system is defined for a specific coding category.

2006-08-21  Kenichi Handa  <handa@m17n.org>

        * font.c (font_matching_entity): Dolete unused local var.

        * xftfont.c (xftfont_open): Call XftDefaultSubstitute before
        opening a font.

        * fileio.c (Finsert_file_contents): On recovering a file, assume
        Unix-like eol.
        (choose_write_coding_system): On auto-saving a file, force
        Unix-like eol.

        * coding.c (setup_coding_system): Fix setting of
        coding->common_flags based on eol_type.
        (coding_inherit_eol_type): If PARENT is not nil, be sure to
        inherit from it.

2006-08-04  Kenichi Handa  <handa@m17n.org>

        * alloc.c (NSTATICS): Increased to 0x600.

2006-08-01  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_driver): Set ftfont_driver.match to
        ftfont_match.
        (ftfont_list): Don't check :name property.
        (ftfont_match): New function.
        (ftfont_pattern_entity): If the pattern doesn't contain
        FC_SPACING, don't assuce FC_MONO.

        * font.h (struct font_driver): New member `match'.
        (font_update_drivers): Prototype adjusted.

        * font.c (font_parse_fcname): Don't change :name property of FONT.
        (font_parse_name): Likewise.
        (LGSTRING_HEADER_SIZE, LGSTRING_GLYPH_SIZE, check_gstring): Define
        them unconditionally.
        (font_matching_entity): New function.
        (font_open_by_name): Try font_matching_entity if exact match is
        not found.
        (font_update_drivers): Delete the arg FONT.  Return a list of
        actually used backends.  Don't free faces, font caches here.
        Don't store data in frame parameters.  Don't call x_set_font.
        (Ffont_spec): Store :name property as is.
        (Ffont_get): Check HAVE_LIBOTF before calling font_otf_capability.
        (Ffont_otf_gsub): Call font->driver->otf_gsub instead of
        font_otf_gsub.
        (Ffont_otf_gpos): Call font->driver->otf_gpos instead of
        font_otf_gpos.
        (Ffont_otf_alternates): Check if the driver has otf_gsub function.
        Call font->driver->otf_gsub instead of font_otf_gsub.
        
        * frame.c (x_set_font_backend): Do more works that were done in
        font_update_drivers before.

        * xfont.c (xfont_match): New function.
        (xfont_driver): Set xfont_driver.match to xfont_match.
        (xfont_draw): Set font in GC if necessary.

        * ftxfont.c (ftxfont_match): New function.
        (syms_of_ftxfont): Set ftxfont_driver.match to ftxfont_match.

        * xftfont.c (xftfont_match): New function.
        (syms_of_xftfont): Set xftfont_driver.match to xftfont_match.

2006-07-28  Kenichi Handa  <handa@m17n.org>

        * font.h (struct font): New member scalable.
        (struct font_driver): New arg ALTERANTE_SUBST to otf_gsub.
        (font_otf_gsub): Prototype adjusted.

        * font.c (font_otf_capability): Fix handling of the default
        langsys.
        (parse_gsub_gpos_spec): Type changed to void.  New arg nbytes.
        Check the contents of SPEC.
        (LGSTRING_HEADER_SIZE, LGSTRING_GLYPH_SIZE): New macros.
        (check_gstring): New function.
        (REPLACEMENT_CHARACTER): New macro.
        (font_otf_gsub): New arg alternate_subst.  Be sure to set all
        glyph codes of GSTRING.
        (font_otf_gpos): Be sure to set all glyph codes of GSTRING.
        (font_prepare_composition): Set cmp->glyph_len.
        (font_open_entity): Set font->scalable.
        (Ffont_get): Handle :otf property.
        (Ffont_otf_gsub, Ffont_otf_gpos, Ffont_otf_alternates): New
        functions.
        (Fquery_font): Use font->font.full_name.
        (syms_of_font): Defsubr Sfont_otf_gsub, Sfont_otf_gpos, and
        Sfont_otf_alternates.

        * ftfont.c (ftfont_open): Set font->font.full_name and
        font->font.name properly.  Fix calculation of font->font.height
        and font->min_width.

        * ftxfont.c (ftxfont_create_gcs): New function.
        (ftxfont_draw_bitmap): Fix arg to ftfont_driver.get_bitmap.
        (ftxfont_draw_backgrond): Fix filling region.
        (ftxfont_default_fid): New function.
        (ftxfont_open): Set xfotn->fid to the return value of
        ftxfont_default_fid.
        (ftxfont_prepare_face): Use ftxfont_create_gcs to create GCs.
        (ftxfont_done_face): Free only GCs that are created by
        ftxfont_create_gcs.
        (ftxfont_draw): If face->gc != s->gc, create proper GCs.

        * xterm.c (x_set_glyph_string_clipping_exactly) [USE_FONT_BACKEND]:
        Clip to src->width, etc (not src->clip_XXX).

        * xfns.c (x_create_tip_frame) [USE_FONT_BACKEND]: Handle
        FontBackend frame parameter.

2006-07-26  Kenichi Handa  <handa@m17n.org>

        * font.h (struct font_driver_list): New member `on'.
        (Fclear_font_cache): EXFUN it.
        (font_update_drivers): Extern it.

        * font.c (font_unparse_fcname): Fix typo (swidth->width).
        (font_list_entities): Check driver_list->on.
        (register_font_driver): Initalize `on' member to 0.
        (font_update_drivers): New function.
        (Fclear_font_cache): Check driver_list->on.

        * frame.h (Qfont_backend): Extern it.
        (x_set_font_backend): Extern it.

        * frame.c (Qfont_backend): New variable.
        (frame_parms): New element for font-backend.
        (x_set_font_backend): New function.

        * xfns.c (Fx_create_frame) [USE_FONT_BACKEND]: Handle
        FontBackend frame parameter.
        (x_frame_parm_handlers) [USE_FONT_BACKEND]: New element
        x_set_font_backend.

        * xfont.c (xfont_list): Don't try listing by :name property if the
        name is not for XLFD.

2006-07-24  Kenichi Handa  <handa@m17n.org>

        * font.h (LGLYPH_FROM, LGLYPH_TO, LGLYPH_SET_FROM)
        (LGLYPH_SET_TO): New macros.
        (LGLYPH_XOFF, LGLYPH_YOFF, LGLYPH_WADJUST): Check if adjustment
        element of G is vector or not.
        (font_at): Extern it.

        * font.c: Include window.h.
        (font_lispy_object): New function.
        (font_prepare_composition): Check LGLYPH_FORM (g) to detect the
        end of valid glyph.
        (font_close_object): Fix getting (struct font *).
        (font_at): New function.
        (Ffont_get): If FONT is a font-object, get entity from it.
        (Ffont_make_gstring): Initialize elements of glyphs with nil.
        (Ffont_fill_gstring): Use macro LGSTRING_XXX and LGLYPH_XXX.  Fix
        range check.
        (Ffont_at): New function.
        (syms_of_font): Defsubr Sfont_at.

        * xdisp.c (it_props): Move the entry for Qauto_composed to just
        before the entry for Qcompostion.
        (handle_auto_composed_prop): Call auto-composition-function with 4
        args.
        (handle_composition_prop) [USE_FONT_BACKEND]: Set it->face_id from
        the font in gstring.
        (fill_composite_glyph_string) [USE_FONT_BACKEND]: Check
        LGLYPH_FORM (g) to detect the end of valid glyph.
        (x_produce_glyphs) [USE_FONT_BACKEND]: Don't update it->face_id if
        we are composing with gstring.

        * xterm.c (x_draw_composite_glyph_string_foreground) [USE_FONT_BACKEND]:
        Check if adjustment is vector or not.

        * Makefile.in (font.o): Make it depends on window.h.

2006-07-24  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_draw_composite_glyph_string_foreground): Check if
        adjustment is vector or not.

2006-07-20  Kenichi Handa  <handa@m17n.org>

        * casefiddle.c (casify_object): Sync with HEAD.

2006-07-18  Miles Bader  <miles@gnu.org>

        * character.h (CHECK_CHARACTER): Redefine in terms of CHECK_TYPE.

2006-07-14  Kenichi Handa  <handa@m17n.org>

        * font.h (LGLYPH_XOFF, LGLYPH_YOFF, LGLYPH_WIDTH, LGLYPH_WADJUST)
        (LGLYPH_SET_WIDTH): Adjusted for the change of LGLYPH format.
        (LGLYPH_ADJUSTMENT, LGLYPH_SET_ADJUSTMENT): New macros.

        * font.c (font_merge_old_spec): Treat '*' in foundry as a wild
        card.
        (DEVICE_DELTA): Fix typo.
        (font_otf_gpos): Adjusted for the change of LGLYPH format.
        (font_prepare_composition): Likewise.

        * xterm.c (x_draw_composite_glyph_string_foreground): Adjusted for
        the change of LGLYPH format.

2006-07-07  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_list): Fix typo.
        (ftfont_build_basic_charsets): Don't include letters with
        diactrics.

2006-07-09  Jan Dj\e,Ad\e(Brv  <jan.h.d@swipnet.se>

        * xfaces.c (realize_non_ascii_face): Set face->extra to NULL.

        * xftfont.c (xftfont_done_face): Call XftDrawDestroy only if
        xftface_info is non-NULL.

2006-07-07  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_list): Fix typo.
        (ftfont_build_basic_charsets): Don't include letters with
        diactrics.

2006-07-05  Jan Dj\e,Ad\e(Brv  <jan.h.d@swipnet.se>

        * ftfont.c (ftfont_list): Move misplaced #endif

2006-07-05  Kenichi Handa  <handa@m17n.org>

        * ftfont.c (ftfont_list): Pay attention to the case that
        FC_CAPABILITY is not defined.

2006-07-03  Kenichi Handa  <handa@m17n.org>

        * xftfont.c (xftfont_open): Set charset related members to -1.

        * ftfont.c (ftfont_list): Handle QCotf property.  Handling of
        QCname fixed.
        (ftfont_open): Set charset related members to -1.

        * fontset.c (Votf_script_alist): New variable.
        (syms_of_fontset): Initialize it.
        (fontset_font): Delete unused variable.

        * fontset.h (Votf_script_alist): Extern it.

        * font.c (font_find_for_lface): Code optimized.

        * font.h (font_close_object, font_merge_old_spec): Extern them.

2006-06-28  Kenichi Handa  <handa@m17n.org>

        * font.c (QCscalable, Qc, Qm, Qp, Qd): New variables.
        (syms_of_font): Initialize them.
        (font_pixel_size): Allow float value in dpi.
        (font_prop_validate_type): Deleted.
        (font_prop_validate_symbol, font_prop_validate_style): Argument
        changed.  Caller changed.
        (font_prop_validate_non_neg): Renamed from
        font_prop_validate_size.
        (font_prop_validate_extra): Deleted.
        (font_prop_validate_spacing): New function.
        (font_property_table): Add elements for all known properties.
        (get_font_prop_index): Renamed from check_font_prop_name.  New
        argument FROM.  Caller changed.
        (font_prop_validate): Validate all known properties.
        (font_put_extra): Argument force deleted.  Caller changed.
        (font_expand_wildcards): Make it static.  Fix the way of shrinking
        the possible range.
        (font_parse_xlfd): Arguemnt merge deleted.  Fix handling of RESX,
        RESY, SPACING, and AVGWIDTH.  Don't validate property values here.
        Caller changed.
        (font_unparse_xlfd): Handle dpi, spacing, and scalable properties.
        (font_parse_fcname): Arguemnt merge deleted.  Fix parsing of point
        size.  Don't validate properties values here.  Caller changed.
        (font_unparse_fcname): Handle dpi, spacing, and scalable
        properties.
        (font_open_by_name): Delete unused variable.
        (Ffont_spec): Likewise.  Validate property values.
        (Ffont_match_p): New function.

        * font.h (QCscalable): Extern it.
        (font_parse_xlfd, font_parse_fcname): Prototype adjusted.

        * ftfont.c (ftfont_list): Handle properties dpi, spacing, and
        scalable.

        * xfont.c (xfont_query_font): Adjusted for the change of
        font_parse_xlfd.
        (xfont_list_pattern): New function.
        (xfont_list): Use xfont_list_pattern.

        * xftfont.c (xftfont_prepare_face): Cancel previous change.
        (xftfont_done_face): Likewise.

2006-06-26  Kenichi Handa  <handa@m17n.org>

        * font.h (Flist_fonts): EXFUN it.

2006-06-25  Jason Rumney  <jasonr@gnu.org>

        * w32term.c (w32_initialize): Add back smoothing_type and
        smoothing_enabled definitions.

2006-06-23  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_draw_glyph_string) [USE_FONT_BACKEND]: Check
        s->face->font on determining underline position.

2006-06-21  Kenichi Handa  <handa@m17n.org>

        * font.c (font_parse_xlfd): Fix generating of CHARSET_REGISTRY field.
        (font_has_char): Accept font-object too.
        (font_find_for_lface): Try at first with a size specified in face.

        * xftfont.c (xftfont_prepare_face): Make non-ascii face share
        face->extra with ascii face.
        (xftfont_done_face): Don't free face->extra of non-ascii face.

2006-06-20  Kenichi Handa  <handa@m17n.org>

        * frame.c (x_set_font) [USE_FONT_BACKEND]: Fix argument to
        font_open_by_name.

2006-06-19  Kenichi Handa  <handa@m17n.org>

        * font.h (QCspacing, QCdpi): Extern them.
        (enum font_spacing): New enum.
        (FONT_PIXEL_SIZE_QUANTUM): New macro.

        * font.c (POINT_TO_PIXEL): Don't divice POINT by 10.
        (QCspacing, QCdpi): New variables.
        (syms_of_font): Initialize them.
        (font_pixel_size): New function.
        (font_put_extra): New function.
        (font_parse_xlfd): Fix handling of font size.  Add QCdpi property
        in FONT_EXTRA.
        (font_parse_fcname): Handle enumenrated values (e.g. bold).  Fix
        handling font size.  Add QCname property that contains only
        unknown properties.
        (font_score): Change argument.  Caller changed.  Pay attention to
        FONT_PIXEL_SIZE_QUANTUM.
        (font_sort_entites): Fix handling of font size.
        (font_list_entities): Likewise.
        (font_find_for_lface): Likewise.
        (font_open_for_lface): Likewise.
        (font_open_by_name): Likewise.
        (Ffont_spec): Add QCname property that contains only unknown
        properties.

        * ftfont.c (ftfont_list): Use assq_no_quit, not Fassq.  Don't
        include weight in listing pattern, instead check weight of each
        listed font.  Don't include scalable in pattern.  Pay attention to
        FONT_PIXEL_SIZE_QUANTUM.

2006-06-19  Kenichi Handa  <handa@m17n.org>

        * lread.c (read_escape): Fix the code synched with HEAD.

        * font.c (font_parse_fcname): Fix parsing of point-size.
        (font_unparse_fcname): Produce symbolic names for style
        properties.
        (font_list_entities): Handle float size correctly.
        (font_open_by_name): Prefer `normal' property values if the name
        doesn't specify them.

        * fontset.c (Finternal_char_font): Use font_get_name, not
        Ffont_xlfd_name.

        * ftfont.c (ftfont_pattern_entity): Use the numeric value 100 for
        FC_WEIGHT_REGULAR.  Exclude FC_SIZE and FC_PIXEL_SIZE from listing
        pattern.  Don't force scalable.

        * xftfont.c (xftfont_open): For generating a name, start from
        96-byte buffer.

2006-06-16  Jan Dj\e,Ad\e(Brv  <jan.h.d@swipnet.se>

        * frame.h (x_new_fontset2): Fix prototype.

2006-06-16  Kenichi Handa  <handa@m17n.org>

        * font.h (struct font_driver): Member parse_name deleted.
        (font_match_p, font_get_spec, font_parse_fcname)
        (font_unparse_fcname): Extern them.
        (font_get_name): Prototype adjusted.

        * font.c (XLFD_SMALLNUM_MASK): Delete this macro.
        (XLFD_LARGENUM_MASK): Delete XLFD_ENCODING_MASK from it.
        (font_expand_wildcards): Fix handling ENCODING field.  Avoid
        unnecessary checks for weight, slant, and swidth.
        (font_parse_fcname): New function.
        (font_unparse_fcname): New function.
        (font_parse_name): New function.
        (font_match_p): New function.
        (font_get_name): Return value changed to Lisp string.
        (font_get_spec): New function.
        (Qunspecified, Qignore_defface): Don't extern them.
        (font_find_for_lface): Assume that LFACE is fully specified.
        (font_load_for_face): If lface[LFACE_FONT_INDEX] is an font
        object, use it for FACE.
        (font_open_by_name): Call Ffont_spec with QCname prop.  Don't call
        driver->parse_name.
        (Ffont_spec): Call font_parse_name, not font_parse_xlfd.

        * fontset.h (new_fontset_from_font) [USE_FONT_BACKEND]: Prototype
        adjusted.

        * fontset.c (new_fontset_from_font) [USE_FONT_BACKEND]: Argument F
        deleted.  Don't call Fnew_fontset.  Instead, directly call
        make_fontset.

        * frame.h (x_new_fontset2) [USE_FONT_BACKEND]: Prototype adjusted.

        * frame.c (x_set_font) [USE_FONT_BACKEND]: Adjusted for the change
        of x_new_fontset2.

        * ftfont.c (Qmonospace, Qsans_serif, Qserif, Qmono, Qsans)
        (Qsans__serif): New variables.
        (ftfont_generic_family_list): New variable.
        (syms_of_ftfont): Initialize the above variables.
        (ftfont_pattern_entity): Argument NAME deleted.
        (ftfont_list_generic_family): New function.
        (ftfont_parse_name): Delete this function.
        (ftfont_list): Try generic family only when FcFontList found no
        font.
        (ftfont_list_family): Fix args to FcObjectSetBuild.

        * xfaces.c (check_lface_attrs) [USE_FONT_BACKEND]: Accept font
        object in attrs[LFACE_FONT_INDEX].
        (set_lface_from_font_name): Cancel all changes for font-backend.
        (set_lface_from_font_and_fontset) [USE_FONT_BACKEND]: New
        function.
        (Finternal_set_lisp_face_attribute) [USE_FONT_BACKEND]: Accept a
        font object in QCfont attribute.
        (set_font_frame_param) [USE_FONT_BACKEND]: Likewise.
        (realize_default_face) [USE_FONT_BACKEND]: Call
        set_lface_from_font_and_fontset.

        * xfns.c (x_default_font_parameter) [USE_FONT_BACKEND]: Try also
        "fixed", and signal error here if no suitable font was found.

        * xfont.c (xfont_parse_name): Delete this function.

        * xftfont.c (xftfont_open): Change coding style of error
        handling.  Generate fontconfig's fontname pattern.

        * xterm.h (struct x_output) [USE_FONT_BACKEND]: New member fontp.
        (FRAME_FONT_OBJECT) [USE_FONT_BACKEND]: New macro.

        * xterm.c (x_new_fontset2) [USE_FONT_BACKEND]: Change arguments.
        Both args FONTSET and FONT_OBJECT must be existing ones.

2006-06-16  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

        * macterm.c (mac_set_unicode_keystroke_event): Don't use MAKE_CHAR.

2006-06-14  Kenichi Handa  <handa@m17n.org>

        * xfont.c (xfont_open, xfont_encode_char): Fix typo.

        * font.h (struct font): Fix typo.

        * font.c (enum xlfd_field_index): Rename XLFD_XXX_SIZE_INDEX to
        XLFD_XXX_INDEX.
        (enum xlfd_field_mask): New enum.
        (intern_font_field): Argument changed.  Caller changed.  If digits
        are followed by non-digits, return a symbol.
        (font_expand_wildcards): New function.
        (font_parse_xlfd): Fix wildcard handling.
        (Ffont_spec): If :name is specified, reflect the info in the other
        properties.

        * ftfont.c (ftfont_pattern_entity): Fix typo.
        (ftfont_list): Enforce FC_LANG in PATTERN to cancel the effect of
        locale.

2006-06-09  Kenichi Handa  <handa@m17n.org>

        * font.h (Qiso8859_1, Qiso10646_1, Qunicode_bmp): Extern them.

        * font.c (Qiso8859_1, Qiso10646_1, Qunicode_bmp): Moved from
        ftfont.c.
        (font_unparse_xlfd): Fix argument type declaration.  Append "*" if
        registry doesn't specify encoding part.
        (font_find_for_lface): Pay attention to LFACE_FONT_INDEX.
        (font_open_by_name): At first try parsing the name.
        (syms_of_font): Declare Qiso8859_1, Qiso10646_1, and Qunicode_bmp
        as Lisp symbols.

        * fontset.c (reorder_font_vector): Pay attention to the case that
        the 3rd element of font_def is nil.
        (fontset_font): For the default fontset, append one more fontset
        elements for a script-based font specification.  Don't add script
        attribute on finding a font.
        (new_fontset_from_font): Unconditionally set FONTSET_ASCII to the
        font name.
        (fontset_ascii_font): If a font can't be opened, return nil.

        * ftfont.c (Qiso8859_1, Qiso10646_1, Qunicode_bmp): Moved to
        font.c.
        (ftfont_pattern_entity): New function.
        (ftfont_get_cache): Assume that freetype_font_cache is already
        initialized.
        (ftfont_list): Handle the case that a file is specified in font
        name.  Use ftfont_pattern_entity to generate entities.
        (ftfont_has_char): Check if the pattern contains FC_CHARSET.
        (syms_of_ftfont): Initialize freetype_font_cache.

        * xftfont.c (xftfont_open): Make the font name fontconfig's
        style.  Add BLOCK_INPUT and UNBLOCK_INPUT.
        (xftfont_close): Free font->font.name if not NULL.

        * xfont.c (xfont_list): If script is specified for a font, return
        null_vector.
        (xfont_list_family): Declare argument type.

        * xfaces.c (set_lface_from_font_name): If a font doesn't have a
        name, set LFACE_FONT (lface) to nil.

        * xterm.c (x_new_fontset2): If an ASCII font couldn't be loaded,
        return Qnil.

2006-06-08  Jason Rumney  <jasonr@gnu.org>

        * w32term.c (w32_initialize): Manually sync 2006-06-05 change from
        HEAD.

2006-06-08  Kenichi Handa  <handa@m17n.org>

        * emacs.c (main): Check -enable-font-backend arg after the check
        of -nl.
        (standard_args): Add "-enable-font-backend".

        * coding.c (Ffind_operation_coding_system): Sync with HEAD.

        * callproc.c (Fcall_process): Sync with HEAD.

        * coding.h (CODING_REQUIRE_ENCODING): Comment sync with HEAD.

2006-06-07  Kenichi Handa  <handa@m17n.org>

        * xftfont.c (xftfont_default_fid): Set fid_known to 1.
        (struct xftdraw_list, xftdraw_list): Delete them.
        (register_xftdraw, check_xftdraw): Delete them.
        (xftfont_prepare_face): Don't call register_xftdraw.
        (xftfont_done_face): Don't call check_xftdraw.
        (xftfont_draw): Get backroudn color only when with_background is
        nonzero.

        * xfont.c (xfont_encode_char): Fix calculation of char2b.

2006-06-06  Kenichi Handa  <handa@m17n.org>

        These changes are for the new font handling codes.

        * Makefile.in (ALL_CFLAGS): Add @FREETYPE_CFLAGS@,
        @FONTCONFIG_CFLAGS@, and @LIBOTF_CFLAGS@.
        (LIB_X11_LIB): If HAVE_XFT is defined, set to @XFT_LIBS@.
        (FONTSRC, FONTOBJ): New variables.
        (obj): Add $(FONTOBJ).
        (SOME_MACHINE_OBJECTS): Lib_X11_Lib.
        (LIBES): Add @FREETYPE_LIBS@, @FONTCONFIG_LIBS@, and
        @LIBOTF_LIBS@.
        (font.o, ftfont.o, xfont.o, xftfont.o, ftxfont.o): New targets.
        (fontset.o, xdisp.o, xfaces.o, xfns.o, xterm.o): Depends on
        $(FONTSRC).

        * font.h, font.c, xfont.c, ftfont.c, xftfont.c, ftxfont.c: New
        files.

        * character.h (Vscript_representative_chars): Extern it.

        * character.c (Vscript_representative_chars): New variable.
        (syms_of_character): Declare it as a Lisp variable.

        * composite.c (get_composition_id) [USE_FONT_BACKEND]: If
        enable_font_backend is nonzero, accept the composition method
        COMPOSITION_WITH_GLYPH_STRING.

        * composite.h (enum composition_method) [USE_FONT_BACKEND]: New
        enumeration COMPOSITION_WITH_GLYPH_STRING.

        * config.in: Re-generated.

        * dispextern.h (struct glyph_string) [USE_FONT_BACKEND]: New
        members clip_x, clip_y, clip_width, and clip_height.
        (struct face) [USE_FONT_BACKEND]: New members font_info and extra.

        * emacs.c (main) [USE_FONT_BACKEND]: Handle arg
        --enable-font-backend.  Call syms_of_font.

        * fns.c (assoc_no_quit): New function.

        * fontset.h (FONT_INFO_FROM_FACE): New macro.
        (face_for_font, new_fontset_from_font)
        (fontset_ascii_font) [USE_FONT_BACKEND]: Extern them.

        * fontset.c [USE_FONT_BACKEND]: Include "font.h".
        (fontset_font, fontset_ascii, face_for_char)
        (make_fontset_for_ascii_face, Ffont_info)
        (Finternal_char_font) [USE_FONT_BACKEND]: If enable_font_backend
        is nonzero, use font-backend mechanism.
        (find_font_encoding): Make it non-static.
        (new_fontset_from_font, fontset_ascii_font) [USE_FONT_BACKEND]:
        New functions.

        * frame.h (struct frame): New members resx and resy.
        (struct frame) [USE_FONT_BACKEND]: New member font_driver_list.
        (x_new_fontset2) [USE_FONT_BACKEND]: Extern it.

        * frame.c [USE_FONT_BACKEND]: Include "font.h".
        (make_frame, x_set_font) [USE_FONT_BACKEND]: Use font-backend
        mechanism.

        * lisp.h (assoc_no_quit): Extern it.

        * xdisp.c: If USE_FONT_BACKEND is defined, include "font.h".
        Through out the file, use FONT_INFO_FROM_FACE instead of
        FONT_INFO_FROM_ID, use get_per_char_metric instead of
        rif->per_char_metric.
        (handle_composition_prop) [USE_FONT_BACKEND]: If the composition
        method is COMPOSITION_WITH_GLYPH_STRING, just set it->c to ' '.
        (get_glyph_face_and_encoding, fill_composite_glyph_string)
        (get_char_face_and_encoding, BUILD_COMPOSITE_GLYPH_STRING)
        (x_produce_glyphs) [USE_FONT_BACKEND]: If enable_font_backend is
        nonzero, use font-backend mechanism.
        (get_per_char_metric): New function.

        * xfaces.c [USE_FONT_BACKEND]: Include "font.h".
        (set_lface_from_font_name)
        (set_font_frame_param, free_realized_face)
        (prepare_face_for_display, clear_face_gcs)
        (Finternal_set_font_selection_order, realize_x_face)
        [USE_FONT_BACKEND]: If enable_font_backend is nonzero, use
        font-backend mechanism.
        (clear_face_cache) [USE_FONT_BACKEND]: Don't call
        clear_font_table.
        (load_face_font) [USE_FONT_BACKEND]: Abort.
        (face_symbolic_value, face_symbolic_weight, face_symbolic_slant)
        (face_symbolic_swidth, face_for_font) [USE_FONT_BACKEND]: New
        functions.

        * xfns.c [USE_FONT_BACKEND]: Include "font.h".
        (x_default_font_parameter) [USE_FONT_BACKEND]: New function.
        (Fx_create_frame) [USE_FONT_BACKEND]: If enable_font_backend is
        nonzero, register all available font drivers.  Call
        x_default_font_parameter for deciding a font.
        (x_create_tip_frame) [USE_FONT_BACKEND]: Likewise.

        * xterm.c [USE_FONT_BACKEND]: Include "font.h".
        (x_set_mouse_face_gc, x_set_glyph_string_clipping)
        (x_set_glyph_string_clipping_exactly)
        (x_compute_glyph_string_overhangs)
        (x_draw_glyph_string_foreground)
        (x_draw_composite_glyph_string_foreground, x_draw_glyph_string)
        (x_free_frame_resources) [USE_FONT_BACKEND]: If
        enable_font_backend is nonzero, use font-backend mechanism.
        (x_new_fontset2) [USE_FONT_BACKEND]: New function.

2006-05-15  Kenichi Handa  <handa@m17n.org>

        * coding.h (system_eol_type): Fix synching with HEAD.

        * coding.c (system_eol_type): Sync with HEAD.
        (coding_inherit_eol_type): If PARENT is nil, inherit from
        system_eol_type.
        (syms_of_coding): Initialize system_eol_type.

        * callproc.c (Fcall_process): Sync with HEAD.

        * process.c (setup_process_coding_systems): Fix synching with
        HEAD.
        (read_process_output): Likewise.
        (Fset_process_coding_system): Inherit system's eol format if
        necessary.

        * fileio.c (choose_write_coding_system): Fix synching with HEAD.

        * keymap.c (push_key_description): Fix synching with HEAD.

2006-05-02  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

        * macgui.h (USE_ATSUI): Don't enable on emacs-unicode-2 branch.

2006-04-07  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_eol): Pay attention to buffer relocation in
        del_range_2.
        (decode_coding): Call decode_eol before restoring undo_list.

2006-03-20  Kenichi Handa  <handa@m17n.org>

        * charset.c (Fdefine_charset_internal): Fix setting of
        emacs_mule_bytes.

2006-03-14  Kenichi Handa  <handa@m17n.org>

        * keyboard.c (read_char): Check if C is a character or not before
        looking up Vkeyboard_translate_table.

2006-03-10  Kenichi Handa  <handa@m17n.org>

        * coding.c (DECODE_EMACS_MULE_20_RELATIVE_COMPOSITION): Fix
        condition to terminate the loop.

2006-03-09  Kenichi Handa  <handa@m17n.org>

        * coding.c (produce_composition): Compare charbuf[i] instead of
        args[i] against 0.
        (Fterminal_coding_system): Use EQ to compare Lisp objects.

2006-03-07  Kenichi Handa  <handa@m17n.org>

        * coding.c (DECODE_COMPOSITION_START): If the source is short, set
        coding->result to CODING_RESULT_INSUFFICIENT_SRC.
        (decode_coding_gap): Set CODING_MODE_LAST_BLOCK after the call of
        detect_coding.
        (emacs_mule_char): Handle old style (Emacs 20) component character
        of a composition.
        (DECODE_EMACS_MULE_COMPOSITION_RULE_20): Fix parsing a composition
        rule.
        (DECODE_EMACS_MULE_20_RULEBASE_COMPOSITION): Likewise.
        (decode_coding_emacs_mule): Handle invalid bytes correctly.

2006-03-04  Kenichi Handa  <handa@m17n.org>

        * coding.c (encode_coding_ccl): Allocate destination dynamically
        when necessary.

2006-03-03  Kenichi Handa  <handa@m17n.org>

        * ccl.c (Fccl_execute_on_string): Fix the condition of terminating
        the loop.  When quitted, show a proper error message.

2006-03-02  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding): Fix previous change.

        * xterm.c (x_set_glyph_string_clipping_exactly): Set
        src->clip_head and src->clip_tail temporarily instead of src->hl.

        * ccl.c (CCL_WRITE_STRING): Handle a flag bit for multibyte
        character sequence.
        (Fccl_execute_on_string): Use ASET, not XSET.

2006-03-01  Kenichi Handa  <handa@m17n.org>

        * search.c (search_buffer): Fix handling of "\\" in a trivial
        regexp.

2006-02-28  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding): Fix the condition of terminating the
        decoding loop.

2006-02-27  Kenichi Handa  <handa@m17n.org>

        * data.c (Faset): On setting a character bigger than 255 in a
        unibyte string, signal an error instead of make the string
        multibyte.

2006-02-22  Kenichi Handa  <handa@m17n.org>

        * charset.c (map_charset_chars): Fix for ascii-compatible charset
        made by a mapping table.

2006-02-21  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (fill_composite_glyph_string): Check s->face is NULL or
        not.
        (BUILD_COMPOSITE_GLYPH_STRING): If C is TAB, set s->face to NULL.
        (x_produce_glyphs): If CH is TAB, set cmp->offsets properly.

        * xterm.c (x_draw_composite_glyph_string_foreground): Check
        s->face is NULL or not.

2006-02-20  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_set_glyph_string_clipping_exactly): New function.
        (x_draw_glyph_string): Fix drawing of right_overhang and
        left_overhang around/on cursor.

        * xdisp.c (draw_glyphs): Fix inclusion of right_overwriting
        glyphs.

        * term.c (produce_glyphs): Sync to HEAD.

2006-02-15  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (x_produce_glyphs): Handle composition with TAB.

2006-02-05  Kenichi Handa  <handa@m17n.org>

        * coding.c: Cancel incorrect synching with HEAD.

2006-02-03  Kenichi Handa  <handa@m17n.org>

        * coding.c (Fdefine_coding_system_internal): Avoid a duplicated
        element in Vcoding_system_alist.
        (Fdefine_coding_system_alias): Likewise.

2006-01-19  Kenichi Handa  <handa@m17n.org>

        * xterm.c (handle_one_xevent): Handle keysyms 0x1000000..0x10000FF.

        * coding.c: Sync to HEAD for handling autoload-coding-system.
        (Qcoding_system_define_form): New variable.
        (syms_of_coding): Intern and staticpro it.
        (Fcoding_system_p): Check Qcoding_system_define_form.
        (Fcheck_coding_system): Try to autoload the definition of
        CODING-SYSTEM.
        
        * coding.h (CODING_SYSTEM_P): If ID is not available, call
        Fcoding_system_p.
        (CHECK_CODING_SYSTEM): If ID is not available, call
        Fcheck_coding_system.
        (CHECK_CODING_SYSTEM_GET_SPEC): Try also Fcheck_coding_system.
        (CHECK_CODING_SYSTEM_GET_ID): Likewise.

2006-01-17  Kenichi Handa  <handa@m17n.org>

        * xterm.c (handle_one_xevent): Delete unnecessary code inserted by
        sync with HEAD.

        * coding.c (code_conversion_restore): GCPRO arg.

2005-12-28  Kenichi Handa  <handa@m17n.org>

        * character.c (lisp_string_width): Check multibyteness of STRING.

2005-10-18  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

        * macterm.c (mac_encode_char): Call ccl_driver with the last arg
        Qnil.  Use JIS_TO_SJIS instead of ENCODE_SJIS.
        (decode_mac_font_name): Use decode_coding_c_string instead of
        decode_coding.
        (x_load_font): Initialize fontp->fontset to -1.  Set
        fontp->encoding_type.

2005-10-17  Kenichi Handa  <handa@m17n.org>

        * search.c (search_buffer): Give up BM search on case-fold-search
        if one of a target character has a case-equivalence of different
        byte length even if that target charcter is an ASCII.
        (simple_search): Fix culculation of byte length of matched text.
        (boyer_moore): Fix handling of case-equivalent multibyte
        characters.

2005-10-15  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding): Fix handling of invalid bytes.

2005-10-06  Kenichi Handa  <handa@m17n.org>

        * xterm.c (handle_one_xevent): Handle keysyms directly mapped to
        Unicode characters.

2005-09-23  Kenichi Handa  <handa@m17n.org>

        * coding.c (encode_coding_object): If a pre-write-conversion
        function makes a new buffer, kill it.

2005-07-29  Kenichi Handa  <handa@m17n.org>

        * coding.c (QCascii_compatible_p): New variable. 
        (syms_of_coding): Initialize it.
        (ONE_MORE_BYTE): Decrement `src' before calling string_char.
        (ONE_MORE_BYTE_NO_CHECK): Likewise.
        (record_conversion_result): Add `default:' case.
        (coding_charset_list): Delete unused variable `coding_type'.
        (Fdefine_coding_system_internal): Add `ascii-compatible-p'
        property in the plist of the coding system.
        (Fcoding_system_put): Check QCascii_compatible_p.

2005-06-09  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (get_next_display_element): Sync with the change in
        HEAD (2005-06-08).

2005-06-06  Kenichi Handa  <handa@m17n.org>

        * callproc.c (Fcall_process): Sync with the change in
        HEAD (2005-06-04).

2005-06-05  Miles Bader  <miles@gnu.org>

        * xfaces.c (Finternal_lisp_face_equal_p): Restore previously
        removed calculation of frame `f', as it's now used.

2005-05-22  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

        * macterm.c (x_font_name_to_mac_font_name): Sync with trunk
        for the case that does not require code conversion.

2005-05-11  Kenichi Handa  <handa@m17n.org>

        * Makefile.in (shortlisp): Cancel previous change.
        (RUN_TEMACS): Include "-nl" if HAVE_SHM is defined.
        (emacs${EXEEXT}): Run $(RUN_TEMACS) unconditionally.
        (UNIDATA): New variable.
        (${lispsource}international/charprop.el): Depends on ${UNIDATA}.
        (bootstrap-emacs${EXEEXT}): Depends on charprop.el.  Run
        $(RUN_TEMACS) unconditionally.

2005-05-10  Kenichi Handa  <handa@m17n.org>

        * Makefile.in (shortlisp): Add ../lisp/international/charprop.el.
        (temacs${EXEEXT}): Build charprop.el if necessary.
        (admindir): New variable.
        ($(lispsource)international/charprop.el): New target.

2005-05-04  Miles Bader  <miles@gnu.org>

        * character.c (chars-in-region): Obsolete function removed.
        (syms_of_character): Remove its initialization.

2005-04-28  Benjamin Riefenstahl  <b.riefenstahl@turtle-trading.net>

        * w32select.c (validate_coding_system)
        (setup_windows_coding_system): New functions.
        (convert_to_handle_as_coded, Fw32_get_clipboard_data): Use
        setup_windows_coding_system.
        (setup_config, Fw32_get_clipboard_data): Use
        validate_coding_system.
        (Fx_selection_exists): Move call to setup_config to a place
        were signals are allowed. 

        * lisp.h (Fcoding_system_base, Fcoding_system_eol_type)
        (Fcheck_coding_system): Add declarations.

2005-04-28  Kenichi Handa  <handa@m17n.org>

        * s/ms-w32.h (STDC_HEADERS): Sync with the change in
        HEAD (2005-04-23).

2005-04-25  Kenichi Handa  <handa@m17n.org>

        * charset.c (load_charset_map_from_vector): Fix for the first
        iteration.

2005-04-22  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

        * macfns.c (Fx_create_frame, x_create_tip_frame): Pass Lisp
        string as the second argument for x_new_fontset.

2005-04-18  Kenichi Handa  <handa@m17n.org>

        * fns.c (Fstring_as_multibyte): Fix the change for syncing with
        CVS head.

2005-04-09  Kenichi Handa  <handa@m17n.org>

        * search.c (search_buffer): Fix the change for syncing with CVS
        head.
        (search_buffer): Likewise.

2005-03-31  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (get_next_display_element): Sync with CVS head.

2005-03-29  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding_object): Use safe_call1 instead of call1.
        (encode_coding_object): Use safe_call instead of call2.

2005-03-14  Kenichi Handa  <handa@m17n.org>

        * fontset.c (set_default_ascii_font): Fix the change for
        syncing with CVS head.

2005-01-30  Kenichi Handa  <handa@m17n.org>

        * fontset.c (Fset_fontset_font): Check family element of a given
        vector.

        * Makefile.in (lisp): Include charprop.el.

2005-01-17  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

        * macfns.c (Fx_create_frame, x_create_tip_frame): Fix crash.
        Not sure if it's unnecessary.

2005-01-16  Steven Tamm  <steventamm@mac.com>

        * macfns.c (Fx_create_frame, x_create_tip_frame): ifdef'd out
        some possibly unnecessary fontset checking code that crashed
        when creating a new frame

2005-01-17  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (merge_faces): Fix argument to lookup_derived_face and
        lookup_face.

        * xdisp.c (Fformat_mode_line): Fix argument to lookup_named_face.

        * fringe.c (draw_fringe_bitmap_1): Fix argument to
        lookup_named_face.

2004-12-25  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (get_next_display_element): Sync to the change in HEAD
        on 2004-12-21.

2004-12-11  Kenichi Handa  <handa@m17n.org>

        * search.c: Sync to the change in HEAD on 2004-11-19, 20.

        * w32console.c: Sync to the change in HEAD on 2004-12-01.

        * coding.c: Cancel the change done in HEAD on 2004-11-30.
        (coding_charset_list): New function.

        * coding.h (coding_charset_list): Extern it.

        * term.c: Sync to the change in HEAD on 2004-11-30.

2004-12-09  Kenichi Handa  <handa@m17n.org>

        * fontset.c (Fset_fontset_font): Call find_font_encoding with
        concatenation of family and registry.

2004-12-06  Kenichi Handa  <handa@m17n.org>

        * character.h (BYTE8_STRING): Fix typo.

        * editfns.c (Ftranslate_region_internal): Don't convert unibyte
        string to multibyte (sync to HEAD).

        * casefiddle.c (casify_region): Handle changes in byte-length
        using replace_range_2 (sync to HEAD).

2004-11-24  Andreas Schwab  <schwab@suse.de>

        * chartab.c (map_char_table): GCPRO table and arg.

2004-10-29  Kenichi Handa  <handa@m17n.org>

        * syntax.c (skip_syntaxes): Return lispy 0 (not nil) if point is
        already at limit.

2004-10-23  Kenichi Handa  <handa@m17n.org>

        * fontset.c (fs_load_font): Use fast_string_match_ignore_case
        instead of fast_c_string_match_ignore_case.
        (find_font_encoding): Argument changed to Lisp_Object.  Use
        fast_string_match_ignore_case instead of
        fast_c_string_match_ignore_case.  Caller changed.

2004-10-15  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (get_next_display_element): In unibyte case, decide to
        display in octal form by checking a chacter by
        UNIBYTE_CHAR_HAS_MULTIBYTE_P.

        * charset.c (Fset_unibyte_charset): Setup
        unibyte_has_multibyte_table.

        * character.c (unibyte_has_multibyte_table): New variable.

        * character.h (unibyte_has_multibyte_table): Extern it.
        (UNIBYTE_CHAR_HAS_MULTIBYTE_P): New macro.

2004-10-14  Kenichi Handa  <handa@m17n.org>

        * callproc.c (Fcall_process): Fix merging of 2004-10-13 change.

2004-10-13  Kenichi Handa  <handa@m17n.org>

        * coding.c (encode_coding_iso_2022): Fix handling of charset
        annotation.

2004-10-12  Kenichi Handa  <handa@m17n.org>

        * coding.c (setup_coding_system): If coding_system is nil, use
        Qundecided.
        (Fterminal_coding_system): Return nil if terminal coding system is
        `undecided'.
        (syms_of_coding): Define coding-system `undecided' here.  Setup
        terminal_coding as `undecided'.

2004-10-04  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (message_dolog, set_message_1): Call
        unibyte_char_to_multibyte with arg type int.

        * fileio.c (Fsubstitute_in_file_name): Fix previous change.

        * lread.c (read1): Fix reading of a char-table.

        * print.c (print_object): Include sub char-table in cicularities
        detection.

2004-10-01  Kenichi Handa  <handa@m17n.org>

        * keymap.c (where_is_internal_2): Fix for the case that KEY is a
        cons.  Append the found sequences in car of ARGS instead of
        prepending.

2004-09-28  Kenichi Handa  <handa@m17n.org>

        * fileio.c (report_file_error): Make a unibyte string from
        strerror (errorno).
        (Fsubstitute_in_file_name): Fix the arg to
        unibyte_char_to_multibyte.  It is evaluated twice.

2004-09-19  Kenichi Handa  <handa@m17n.org>

        * charset.h (CHAR_CHARSET): Shortcut for ASCII case.

2004-09-14  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding): Fix previous change.

2004-09-13  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding_utf_16): Don't set detect_info->found if
        BOM is not found.
        (detect_coding): Optimization for ISO-2022 when no 8-bit data is
        found.
        (detect_coding_system): Likewise.
        
2004-09-01  Jason Rumney  <jasonr@gnu.org>

        * w32fns.c (x_to_w32_font): Update to use new coding struct.

2004-08-17  Kenichi Handa  <handa@m17n.org>

        * charset.c (Fdeclare_equiv_charset): Fix handing of CHARS.
        (Fiso_charset): Likewise.

2004-08-03  Steven Tamm  <steventamm@mac.com>

        * macterm.c (mac_encode_char): Add charset argument and update
        to use encoding_type    
        (x_new_font,x_new_fontset): Merge in changes from xterm.c;
        switch to pure fontset
        (decode_mac_font_name): Temporarily remove decoding
        (x_font_name_to_mac_font_name): Temporarily remove encoding
        (x_load_font): Temporarily remove encoding

2004-06-30  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (Fface_font): If frame is not on a window system,
        ignore CHARACTER arg.  If HAVE_WINDOW_SYSTEM is not defined, don't
        refer to face->font.
        (split_font_name_into_vector, build_font_name_from_vector)
        (lookup_non_ascii_face, realize_non_ascii_face): Define them only
        whne HAVE_WINDOW_SYSTEM is defined.

2004-05-29  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (BUILD_GLYPH_STRINGS): Check if s is NULL.
        (x_produce_glyphs): Fix setting of members of cmp in case
        cmp->glyph_len is zero,

        * fontset.c (Fset_fontset_font): Docstring fixed.
        (Ffontset_info): Make it backward compatible.  New arg ALL.

2004-05-11  Kim F. Storm  <storm@cua.dk>

        * process.c (read_process_output): Grow decoding_buf when needed;
        this could cause a crash in allocate_string and compact_small_strings.

2004-04-29  Kenichi Handa  <handa@m17n.org>

        * fileio.c (WRITE_BUF_SIZE): This macro deleted.
        (e_write): Fix previous change.

2004-04-28  Kenichi Handa  <handa@m17n.org>

        * coding.c (setup_coding_system): Set coding->common_flags
        correctly for raw-text.
        (consume_chars): On encoding unibyte text by raw-text, don't check
        multibyte form.
        (encode_coding): On encoding by raw-text, never use translation
        tables.

        * fileio.c (e_write): Short cut for the case of no encoding.

2004-04-20  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding): Delete unused variables.
        (detect_coding_system): Likewise.

2004-04-18  Kenichi Handa  <handa@m17n.org>

        * coding.c (encode_coding_utf_8): Fix handling of raw-byte char.
        (consume_chars): Fix handling of 8-bit bytes in unibyte source.

2004-04-14  Kenichi Handa  <handa@m17n.org>

        Sync all files to HEAD.

2004-04-14  Kenichi Handa  <handa@m17n.org>

        * coding.c (Ffind_coding_systems_region_internal): Include
        raw-text and no-conversion in the result.

        * fontset.h: Sync to HEAD.

        * fontset.c: Sync to HEAD.

2004-04-14  Kenichi Handa  <handa@m17n.org>

        * fontset.c (find_font_encoding): Return `ascii' for unknown
        encoding.
        (load_font_get_repertory): Delete unnecessary check of ENCODING of
        FONT_DEF.
        (font_def_arg, add_arg, from_arg, to_arg): New args.
        (set_fontset_font): Argument changed.
        (Fset_fontset_font): Fix for the case that TARGET is a script
        name and charset name.
        (new_fontset_from_font_name): Fix argument to Fnew_fontset.

2004-04-13  Kenichi Handa  <handa@m17n.org>

        * fontset.c (fontset_font): Renamed from fontset_face.  Return
        value changed.
        (face_suitable_for_char_p): Adjusted for the change of
        fontset_font.
        (face_for_char): Likewise.
        (make_fontset_for_ascii_face): Fix setting of the fontset element
        for ASCII.
        (Finternal_char_font): Use fontset_font instead of FACE_FOR_CHAR
        to get a font name.
        (Ffontset_info): Adjusted for the change of fontset_font.

        * composite.c: Sync to HEAD.

        * search.c: Sync to HEAD.

        * coding.c: Sync to HEAD.
        (emacs_mule_char): Check invalid code more regidly.
        
        * coding.h: Sync to HEAD.

        * charset.c: Sync to HEAD.

        * charset.h: Sync to HEAD.

        * character.h (LEADING_CODE_LATIN_1_MIN)
        (LEADING_CODE_LATIN_1_MAX): Delete these macros.

2004-04-08  Kenichi Handa  <handa@m17n.org>

        * category.h: Sync to HEAD.

        * category.c: Sync to HEAD.

        * syntax.h: Sync to HEAD.

        * syntax.c: Sync to HEAD.

        * regex.h: Sync to HEAD.

        * regex.c: Sync to HEAD.

2004-04-07  Kenichi Handa  <handa@m17n.org>

        * editfns.c: Sync to HEAD.
        (check_translation): New function.
        (Ftranslate_region_internal): Handle M:N mapping.

2004-04-06  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (xlfd_point_size): Set font->numeric[XLFD_PIXEL_SIZE].

2004-03-30  Kenichi Handa  <handa@m17n.org>

        * coding.c (DECODE_DESIGNATION): Set chars_96 to -1 instead of
        goto invalid_code.
        (decode_coding_iso_2022): Fix handling of invalid designation.

        * fileio.c (Finsert_file_contents): Be sure to call unbind_to
        after calling code_conversion_save.

2004-03-11  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (handle_auto_composed_prop): Fix Lisp_Object/int mixup.

        * print.c (print_prune_string_charset): Fix Lisp_Object/int mixup.

        * fontset.c: Include "intervals.h".
        (fontset_face): Fix comparing of Lisp_Objects.
        (free_face_fontset): Fix Lisp_Object/int mixup.
        (new_fontset_from_font_name): Likewise.

        * editfns.c (Ftranslate_region_internal): Fix Lisp_Object/int mixup.

        * coding.c: Add many prototypes for static functions.
        (get_translation_table): Allow max_lookup to be NULL.
        (decode_coding): Call get_translation_table with max_lookup NULL.
        (Ffind_coding_systems_region_internal): Likewise.
        (Funencodable_char_position, Fcheck_coding_systems_region):
        Likewise.

2004-03-11  Kenichi Handa  <handa@m17n.org>

        * coding.c (get_translation_table): Declare it as Lisp_Object.
        (LOOKUP_TRANSLATION_TABLE): New macro.
        (produce_chars): Use LOOKUP_TRANSLATION_TABLE instead of
        CHAR_TABLE_REF.
        (consume_chars): Likewise.

2004-03-11  Kenichi Handa  <handa@m17n.org>

        * coding.c (MAX_ANNOTATION_LENGTH): Adjusted for the change of
        annotation data format.
        (ADD_ANNOTATION_DATA, ADD_COMPOSITION_DATA, ADD_CHARSET_DATA):
        Change arguments FROM and TO to single argument NCHARS.  Caller
        changed.
        (decode_coding_utf_8): Pay attention to coding->charbuf_used.
        (decode_coding_utf_16, decode_coding_emacs_mule)
        (decode_coding_iso_2022, decode_coding_sjis, decode_coding_big5)
        (decode_coding_ccl, decode_coding_charset): Likewise.
        (get_translation): New function.
        (produce_chars): New arguments translation_table and last_block.
        Translate characters here.  Return number of carryover chars.
        Caller changed.
        (produce_composition): New argument pos.  Caller changed.
        Adjusted for the change of annotation data format.
        (produce_charset, produce_annotation): Likewise.
        (decode_coding, encode_coding): Don't call translate_chars.
        (consume_chars): New arg translation_table.  Caller changed.
        (translate_chars): Deleted.
        (syms_of_coding): Make translation-table's number of extra slots
        2.

2004-03-09  Kenichi Handa  <handa@m17n.org>

        * search.c (simple_search): Fix setting this_pos_byte in backward
        search.

        * coding.c (detect_coding_emacs_mule): Fix counting of encoded
        byte sequence.
        (detect_coding_ccl): Fix setting of the variable valids.

2004-03-04  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_list_fonts): Fix the detection of an auto-scaled font.

        * coding.c (decode_coding_utf_16): Fix handling of surrogate pair.

        * editfns.c (Ftranslate_region_internal): Renamed from
        Ftranslate_region.  Accept a char-table in TABLE.
        (syms_of_editfns): Defsubr Stranslate_region_internal.

        * xfaces.c (set_lface_from_font_name): If a font is specified for
        a frame, generate a fontset from the font.
        (build_scalable_font_name): If the scalable font is requested for
        a specific size, don't change that size.
        (try_font_list): Try a scalable font also in the case that a
        pattern string is specified,
        

2004-03-03  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (Fface_font): New optional arg CHARACTER.

2004-02-17  Kenichi Handa  <handa@m17n.org>

        * charset.h (CHARSET_OFFSET): New macro.

2004-02-13  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_get_font_repertory): Fix for non-Unicode-bmp charset.

        * fontset.c (fontset_face): Handle the case that repertory is a
        char-table.
        (find_font_encoding): Return nil for unknown encoding.
        (Fset_fontset_font): Ignore a font of unknown encoding.

2004-02-09  Kenichi Handa  <handa@m17n.org>

        * keymap.c (describe_vector): Handle default value of a char
        table.

        * fontset.c (fontset_face): Handle fallback fonts correctly.
        (Ffontset_info): Return infomation about fallback fonts.

2004-02-06  Kenichi Handa  <handa@m17n.org>

        * fontset.c (FONTSET_DEFAULT): New macro.
        (FONTSET_ADD): Handle the case that range is nil.
        (fontset_add): Likewise.
        (Fset_fontset_font): Change the 2nd arg name to TARGET, and handle
        the case that it is nil.
        (dump_fontset): Call FONTSET_DEFAULT, not FONTSET_FALLBACK.
        (syms_of_fontset): Set char-table-extra-slots property of fontset
        to 9.

        * charset.h (CHAR_CHARSET_P): Fix for the case that the method is
        subset or superset.

2004-01-30  Kenichi Handa  <handa@m17n.org>

        * emacs.c (main): Call init_charset after syms_of_XXX.

        * charset.c (Vcharset_map_directory): Deleted.
        (Vcharset_map_path): New variable
        (load_charset_map_from_file): Use Vcharset_map_path instead.
        (init_charset): Initialize Vcharset_map_path.
        (syms_of_charset): Delete declaration of "charset-map-directory",
        add declaration of "charset-map-path".

2004-01-29  Kenichi Handa  <handa@m17n.org>

        * fns.c (string_char_to_byte): Optimize for ASCII only string.
        (string_byte_to_char): Likewise.

        * fileio.c (Finsert_file_contents): Avoid detecting a code twice.

        * coding.c (detect_coding_iso_2022): Fix handling of SS2 and SS3.
        (detect_coding): Treat '\0' as normal ASCII byte..
        (detect_coding_system): Likewise.

2004-01-27  Kenichi Handa  <handa@m17n.org>

        * coding.h (SJIS_TO_JIS2, JIS_TO_SJIS2): New macros.

        * coding.c (QCmnemonic, QCdefalut_char)
        (QCdecode_translation_table, QCencode_translation_table)
        (QCpost_read_conversion, QCpre_write_conversion): New variables.
        (get_translation_table): Return a list of translation tables if
        necessary.
        (decode_coding): Call get_translation_table with ENCODEP 0.
        (char_encodable_p): If translation_table is non-nil, always call
        translate_char.
        (Fdefine_coding_system_internal): Accept list of translation
        tables as :encode-translation-table and :decode-translation-table.
        (Fcoding_system_put): New function.
        (syms_of_coding): Declare new symbols.  Defsubr
        Scoding_system_put.
        (decode_coding_sjis): Handle 4th charset (typically JISX0212).
        (encode_coding_sjis): Likewise.

        * charset.c (map_charset_chars): Fix arg to map_charset_chars in
        when the charset is superset type.

        * character.c (translate_char): Accept list of translation tables.

2004-01-25  Kenichi Handa  <handa@m17n.org>

        * coding.h (enum coding_attr_index): New member
        coding_attr_trans_tbl.
        (CODING_ATTR_TRANS_TBL): New macro.

        * coding.c (get_translation_table): New function.
        (translate_chars): Fix the bug of skipping annotation data.
        (decode_coding): Utilze get_translation_table.
        (encode_coding): Likewise.
        (char_encodable_p): Translate char if necessary.
        (Funencodable_char_position): Likewise.
        (Ffind_coding_systems_region_internal): Setup translation table
        for encode in a coding system attribute vector in advance.
        (Fcheck_coding_systems_region): Likewise.
        (Fdefine_coding_system_internal): Allow a symbol as translation
        table.  For shift-jis type coding system, allow 4th charset.

2004-01-24  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding_sjis): Check the first byte rigidly.

        * xdisp.c (get_next_display_element): Pass -1 as POS to
        FACE_FOR_CHAR if displaying a C-string.

2004-01-23  Kenichi Handa  <handa@m17n.org>

        * composite.c (get_composition_id): Handle xoff and yoff in a
        composition rule.

        * composite.h (COMPOSITION_DECODE_RULE): New arg xoff and yoff.
        (struct composition): New member lbearing and rbearing.

        * xdisp.c (move_it_to): Optimize for the case (op & MOVE_TO_Y).
        (x_get_glyph_overhangs): Handle a composition glyph.
        (x_produce_glyphs): Setup lbearing and rbreaing for a composition
        glyph.

        * xterm.c (x_compute_glyph_string_overhangs): Handle also a
        composition glyph.

2004-01-18  Kenichi Handa  <handa@m17n.org>

        * print.c: Include charset.h.
        (Vprint_charset_text_property): New variable.
        (Qdefault): Extern it.
        (PRINT_STRING_NON_CHARSET_FOUND)
        (PRINT_STRING_UNSAFE_CHARSET_FOUND): New macros.
        (print_check_string_result): New variable.
        (print_check_string_charset_prop): New function.
        (print_prune_charset_plist): New variable.
        (print_prune_string_charset): New function.
        (print_object): Call print_prune_string_charset if
        Vprint_charset_text_property is not t.
        (print_interval): Print nothing if itnerval->plist is nil.
        (syms_of_print): Declare Vprint_charset_text_property as a lisp
        variable.  Init and staticpro print_prune_charset_plist.

2004-01-15  Kenichi Handa  <handa@m17n.org>

        * fontset.c (new_fontset_from_font_name): Use the specified font
        for all characters in the new fontset.

        * macterm.c (x_set_mouse_face_gc): Call FACE_FOR_CHAR with POS and
        OBJECT args.

        * xdisp.c (x_produce_glyphs): Call FACE_FOR_CHAR with POS and
        OBJECT args for composition too.

        * w32term.c (x_set_mouse_face_gc): Call FACE_FOR_CHAR with POS and
        OBJECT args.

2004-01-13  Kenichi Handa  <handa@m17n.org>

        * dispextern.h (FACE_FOR_CHAR): New args POS and OBJECT.

        * fontset.c (reorder_font_vector): Adjusted for the change of
        FONT_DEF format.
        (fontset_face): New arg id.  Caller changed.
        (face_for_char): New args pos and object.
        (make_fontset_for_ascii_face): Adjusted for the change of FONT_DEF
        format.n
        (fs_query_fontset): Check NAME by Fassoc too.
        (Fset_fontset_font): Allow non-XLFD font name.
        (Ffontset_info): Adjusted for the change of FONT_DEF format.

        * fontset.h (face_for_char): Prototype adjusted.

        * xdisp.c (face_before_or_after_it_pos): Call FACE_FOR_CHAR with
        POS and OBJECT args.
        (get_next_display_element): Likewise.
        (append_space): Likewise.
        (extend_face_to_end_of_line): Likewise.
        (get_char_face_and_encoding): Likewise.
        (BUILD_COMPOSITE_GLYPH_STRING): Likewise.
        (x_produce_glyphs): Likewise.

        * xfaces.c (compute_char_face): Call FACE_FOR_CHAR with
        POS and OBJECT args.

        * xterm.c (x_set_mouse_face_gc): Call FACE_FOR_CHAR with
        POS and OBJECT args.

2004-01-03  Jason Rumney  <jasonr@gnu.org>

        * w32select.c (Fw32_set_clipboard_data): Avoid potential realloc
        of GlobalAlloc'ed memory.

2003-12-29  Kenichi Handa  <handa@m17n.org>

        * ccl.c (Fccl_execute_on_string): Fix the condition of loop.

        * charset.h (charset_table_used): Delete extern.

        * charset.c (charset_table_used): Make it static.
        (map_charset_chars): Fix args to c_function with.

        * chartab.c (map_sub_char_table_for_charset): Fix args to
        c_function with.

        * coding.h (enum coding_result_code): Delete
        CODING_RESULT_INSUFFICIENT_CMP, add CODING_RESULT_INVALID_SRC.

        * coding.c (Qinsufficient_source, Qinconsistent_eol)
        (Qinvalid_source, Qinterrupted, Qinsufficient_memory): New
        variables.
        (Vlast_code_conversion_error): New variables.
        (syms_of_coding): DEFSYM or DEFVAR_LISP them. 
        (ONE_MORE_BYTE): Record error if any instead of signaling an
        error.  If non-ASCII multibyte char is found, return the negative
        value of the code.  All callers changed to check it.
        (ONE_MORE_BYTE_NO_CHECK): Likewise.
        (record_conversion_result): New function.  All codes setting
        coding->result are changed to call this function.
        (detect_coding_utf_8): Don't use the local variable incomplete.
        (decode_coding_utf_8): Likewise.
        (emacs_mule_char): Change the second arg to `const'.
        (detect_coding_emacs_mule): Don't use the local variable
        incomplete.
        (detect_coding_sjis): Likewise.
        (detect_coding_big5): Likewise.
        (decode_coding): Fix of flushing out unprocessed data.
        (make_conversion_work_buffer): Fix making of a work buffer.
        (decode_coding_object): Return coding->dst_object;

        * fontset.c (set_fontset_font): Fix args.

        * lisp.h (CHARACTERBITS): Define as 22.

        * process.c (send_process): Be sure to set coding->src_multibyte.

        * xdisp.c (handle_auto_composed_prop): Fix setting of limit.

2003-12-02  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (handle_auto_composed_prop): Give limit to
        Fnext_single_char_property_change.

2003-12-02  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding): Fix previous change.
        (detect_coding_system): Likewise.

2003-12-02  Kenichi Handa  <handa@m17n.org>

        * composite.c (syms_of_composite): Don't make the compostion hash
        table week.

        * fontset.c (Fset_fontset_font): Fix docstring.

        * lisp.h (detect_coding_system): Adjust prototype.

        * fileio.c (kill_workbuf_unwind): Delete this function.
        (Finsert_file_contents): Adjust the call of detect_coding_system.
        Get conversion_buffer by code_conversion_save.  Use the macor
        CODING_MAY_REQUIRE_DECODING.  After decoding, update
        coding_system.

        * coding.h (make_conversion_work_buffer): Delete extern.
        (code_conversion_save): Extern it.

        * coding.c (enum iso_code_class_type): Delete ISO_carriage_return.
        (CODING_GET_INFO): Delete argument eol_type.  Callers changed.
        (decode_coding_utf_8): Don't do eol converion.
        (detect_coding_utf_16): Check coding->src_chars, not
        coding->src_bytes.  Add heuristics for those that have no
        signature.
        (decode_coding_emacs_mule): Don't do eol converion.
        (decode_coding_iso_2022): Likewise.
        (decode_coding_sjis): Likewise.
        (decode_coding_big5): Likewise.
        (decode_coding_charset): Likewise.
        (adjust_coding_eol_type): Return a new coding system.
        (detect_coding): Don't detect eol.  Fix for utf-16 detection.
        (decode_eol): In case of CRLF->LF conversion, use del_range_2 on
        each change.
        (decode_coding): Pay attention to undo_list.  Do eol convesion for
        all types of coding-systems (if necessary).
        (Vcode_conversion_work_buf_list): Delete it.
        (Vcode_conversion_reused_workbuf): Renamed from
        Vcode_conversion_reused_work_buf.
        (Vcode_conversion_workbuf_name): New variable.
        (reused_workbuf_in_use): New variable.
        (make_conversion_work_buffer): Delete the arg DEPTH.
        (code_conversion_restore): Argument changed to cons.
        (code_conversion_save): Delete the argument BUFFER.  Callers
        changed.
        (detect_coding_system): New argument src_chars.  Callers changed.
        Fix for utf-16 detection.
        (init_coding_once): Don't use ISO_carriage_return.
        (syms_of_coding): Initialized Vcode_conversion_workbuf_name and
        reused_workbuf_in_use.

2003-11-24  Kenichi Handa  <handa@m17n.org>

        * keymap.c (store_in_keymap): Pay attention to the case that idx
        is a cons specifying a character range.

        * coding.c (Fdefine_coding_system_internal): Fix previous change.

2003-11-23  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (handle_auto_composed_prop): Fix the case of returning
        HANDLED_RECOMPUTE_PROPS.

        * coding.c (Fdefine_coding_system_internal): Fix checking of
        ascii compatibility.

2003-11-22  Kenichi Handa  <handa@m17n.org>

        * charset.c (find_charsets_in_text): Delete unused locale
        variable.
        (Fset_charset_priority): Update Vemacs_mule_charset_list too.

        * coding.c (encode_coding_emacs_mule): Emit bytes with MSB.
        Resync charset_list to Vemacs_mule_charset_list.

        * keymap.c (store_in_keymap): Pay attention to the case that idx
        is a cons specifying a character range.

2003-11-18  Kenichi Handa  <handa@m17n.org>

        * composite.c (update_compositions): Bind inhibit-read-only, etc
        to t before calling remove-list-of-text-properties.

        * print.c (print_object): Always print ASCII chars as is.

2003-11-17  Kenichi Handa  <handa@m17n.org>

        * keymap.c (Fdefine_key): Fix handling of Lucid style event type
        list.

        * fns.c (Fmapconcat): Signal an error if SEQUENCE is a char table.
        (Fmapcar): Likewise.
        (Fmapc): Likewise.

2003-11-15  Kenichi Handa  <handa@m17n.org>

        * syntax.c (skip_chars): Be sure to alloca char_ranges when
        necessary.

2003-11-14  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (set_lface_from_font_name): Fix for the case that
        FONTNAME is not fontset name.

2003-11-13  Kenichi Handa  <handa@m17n.org>

        * fns.c (base64_encode_1): Fix previous change.

2003-11-08  Kenichi Handa  <handa@m17n.org>

        * fontset.c (set_fontset_font): New function.
        (Fset_fontset_font): If a font is specified for a charset, use
        map_charset_chars to store the font spec in a fontset.

2003-10-29  Kenichi Handa  <handa@m17n.org>

        * fontset.c (fontset_face): Create a fallback fontset on demand
        (make_fontset): Don't create a fallback fontset here.
        (free_face_fontset): Free a fallback fontset (if any) too.
        (n_auto_fontsets): Delete this variable.
        (auto_fontset_alist): New variable.
        (new_fontset_from_font_name): Check auto_fontset_alist.
        (dump_fontset) [FONTSET_DEBUG]: Fully re-written.
        (Ffontset_list_all) [FONTSET_DEBUG]: New function.
        (syms_of_fontset): Initialize and staticpro auto_fontset_alist.
        Defsubr Sfontset_list_all.

2003-10-24  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_list_fonts): Fix excluding of auto-scaled fonts.

2003-10-23  Kenichi Handa  <handa@m17n.org>

        * fontset.c (Fnew_fontset): Check NAME more rigidly.

2003-10-17  Kenichi Handa  <handa@m17n.org>

        * editfns.c (Fgoto_char): Fix docstring.

2003-10-16  Kenichi Handa  <handa@m17n.org>

        * insdel.c (insert_from_gap): Adjust intervals correctly.

2003-10-12  Jason Rumney  <jasonr@gnu.org>

        * w32term.c (GLYPHSET, WCRANGE): Define if system headers don't.
        (pfnGetFontUnicodeRanges): New dynamically loaded function.
        (w32_initialize): Try to load it.
        (x_get_font_repertory): Use it if available.
        (w32_encode_char): Add shortcut for unicode output.

        * w32fns.c (w32_load_system_font): Default charset to -1.
        (x_to_w32_charset): Match all fonts for unicode.
        (w32_to_x_charset): New parameter matching. Don't return partial
        or wildcard charsets.
        (w32_to_all_x_charsets): Don't return partial or wildcard charsets.
        (w32_codepage_for_font): Return CP_UNICODE for unicode.
        (w32_to_x_font): Match charset to real charset.
        (enum_font_cb2): Always list unicode versions.

        * makefile.w32-in (temacs): Increase EMHEAP.

2003-10-11  Jason Rumney  <jasonr@gnu.org>

        * w32term.c (w32_encode_char): New charset parameter.
        font_info.encoding becomes encoding_type.
        (x_get_font_repertory): New function. Warning: stub only!
        (x_new_font): Return quickly if font already set.
        (x_new_fontset): fontsetname parameter is Lisp_Object.
        Use new fs_query_fontset. Try new_fontset_from_font_name. Use
        fontset_name for return value.

        * w32term.h: Declare x_get_font_repertory.

        * w32select.c (Fw32_set_clipboard_data): Use string_x_string_p in
        place of find_charset_in_text. Use encode_coding_object in place
        of encode_coding.
        (Fw32_get_clipboard_data): Use decode_coding_c_string in place of
        decode_coding.

        * w32fns.c (Fx_create_frame, x_create_tip_frame): Use new version
        of x_new_fontset.
        (w32_load_system_font): Initialize charset as unicode.
        font_info.encoding becomes encoding_type.
        (w32_to_x_font): Use decode_coding_c_string in place of
        decode_coding.
        (x_to_w32_font): Use encode_coding_object in place of
        encode_coding.
        (syms_of_w32fns): Set get_font_repertory_func.

        * w32console.c: Include character.h. Use terminal_encode_buffer
        from term.c.
        (write_glyphs): Use new version of encode_terminal_code. Use
        encode_coding_object in place of encode_coding.

        * w32bdf.c (w32_load_bdf_font): Clear font_info before filling.
        encoding becomes encoding_type.

        * term.c (terminal_encode_buffer): Make externally visible.

        * makefile.w32-in: Add character.h dependancies.
        (character.o, chartab.o): New targets.

2003-10-10  Kenichi Handa  <handa@m17n.org>

        * fileio.c (Finsert_file_contents) [DOS_NT]: Use the macro
        CODING_ID_EOL_TYPE..

2003-10-07  Andreas Schwab  <schwab@suse.de>

        * coding.c (produce_chars): Revert last change.

2003-10-06  Kenichi Handa  <handa@m17n.org>

        * charset.h (charset_unicode): Extern it.

        * charset.c (string_xstring_p): Check by (C >= 0x100).
        (find_charsets_in_text): Format of the arc CHARSETS changed.  New
        arg MULTIBYTE.
        (Ffind_charset_region, Ffind_charset_string): Adjusted for the
        change of find_charsets_in_text.
        (Fsplit_char): Fix doc.  Never return unknown.

        * chartab.c (char_table_translate): Use CHARACTERP, not INETEGERP.

        * coding.c (Fdefine_coding_system_alias): Update
        Vcoding_system_list.

        * fontset.c (load_font_get_repertory): Pay attention to the case
        that ENCODING of a font is specified by a char-table.

        * xterm.c (x_get_font_repertory): Handle the case that the
        encoding of font is other than Unicode.

2003-10-02  Kenichi Handa  <handa@m17n.org>

        * term.c (encode_terminal_code): Don't handle glyph-table.  Check
        if a character is encodable by the terminal coding system.  If
        not, produces proper number of `?'s.  Update
        terminal_encode_buffer and terminal_encode_buf_size if necessary.
        (produce_glyphs): Check by CHAR_BYTE8_P, not SINGLE_BYTE_CHAR_P.

2003-10-01  Kenichi Handa  <handa@m17n.org>

        * term.c (terminal_encode_buffer, terminal_encode_buf_size): New
        variables.
        (encode_terminal_code): Argument changed.  Encode multiple
        characters at once.  Store the result of encoding in
        terminal_encode_buffer.
        (write_glyphs): Adjusted for the change of encode_terminal_code.
        (insert_glyphs): Likewise.
        (term_init): Initialize terminal_encode_buffer and
        terminal_encode_buf_size.

        * coding.c (consume_chars): If coding->src_object is nil, don't
        check annotation.

2003-09-30  Kenichi Handa  <handa@m17n.org>

        * character.c (char_string): Use ASCII_CHAR_P instead of
        SINGLE_BYTE_CHAR_P.

2003-09-30  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (handle_auto_composed_prop): Check if the last
        characters of auto-composed region is newly composed with the
        following characters.
        (handle_composition_prop): Fix checking of point being inside
        composition.

2003-09-26  Kenichi Handa  <handa@m17n.org>

        * fns.c (concat): Don't change multibyteness of the result by
        concatenating an 8-bit character.

        * data.c (Faset): Check newelt by CHECK_CHARACTER.  Don't change
        multibyteness of the result when newelt is an 8-bit character.

2003-09-29  Dave Love  <fx@gnu.org>

        * xmenu.c (find_and_call_menu_selection): Make menu_bar_items_used
        EMACS_INT.

        * xfns.c (DefaultDepthOfScreen, x_encode_text): Remove unused vars.

        * xfaces.c (face_numeric_value): Declare dim size_t.
        (Finternal_lisp_face_equal_p): Remove unused f.

        * xdisp.c (BUILD_CHAR_GLYPH_STRINGS, display_and_set_cursor)
        (MATRIX_ROW): Remove unused vars.
        (draw_glyphs, x_insert_glyphs, fast_find_position)
        (fast_find_position, fast_find_string_pos): Use EMACS_INT for
        byte/char counts.

        * regex.c (regex_compile): Remove unused var.

        * minibuf.c (Fminibuffer_complete_word): Remove unused var.

        * keymap.c (Fset_keymap_parent, map_keymap, Fcopy_keymap)
        (Faccessible_keymaps, where_is_internal): Remove unused vars.

        * keyboard.c (cancel_hourglass_unwind): Return Qnil.

        * frame.c (frame_name_fnn_p): Make len EMACS_INT.

        * fileio.c (Fwrite_region): Remove unused var.

        * dispnew.c (adjust_frame_glyphs_for_frame_redisplay)
        (adjust_frame_glyphs_for_window_redisplay): Remove unused ch_dim.

        * composite.c (Fremove_list_of_text_properties): Declare.

        * coding.c (inhibit_pre_post_conversion): Removed (unused).
        (alloc_destination, produce_chars): Use EMACS_INT for byte/char
        counts.
        (coding_inherit_eol_type): Remove unused attrs.
        (detect_coding): Cast arg of detect_eol.

        * charset.c (syms_of_charset): Remove unused var p.
        (find_charsets_in_text, Ffind_charset_region): Use EMACS_INT for
        byte/char counts.

        * casetab.c (set_case_table): Remove unused var.

        * window.c (Fdisplay_buffer, Fframe_selected_window): Remove
        unsued vars.

2003-09-26  Dave Love  <fx@gnu.org>

        * xterm.c (x_bitmap_mask): Declare.

2003-09-17  Dave Love  <fx@gnu.org>

        * xterm.c (x_term_init): Fix type error.

        * lisp.h: Add Funibyte_char_to_multibyte.

        * coding.c (Fread_coding_system): Fix arg of XSETSTRING.
        (Fset_coding_system_priority): Doc fix.

        * alloc.c: Sync with HEAD version.

        * ccl.c (ccl_driver): Fix arg of CHARACTERP.

        * indent.c (check_composition): Make start and end EMACS_INT.

        * character.c (lisp_string_width): Make ignore and end EMACS_INT.

        * xdisp.c (handle_composition_prop, check_point_in_composition):
        Make buffer positions EMACS_INT.

        * composite.c (find_composition, run_composition_function)
        (update_compositions, Ffind_composition_internal): Make buffer
        positions EMACS_INT.

        * composite.h (find_composition, update_compositions): Make
        position args EMACS_INT.

        * keyboard.c (adjust_point_for_property): Make beg and end
        EMACS_INT.

        * intervals.c (get_property_and_range)
        * intervals.h (get_property_and_range): Make start and end EMACS_INT.

        * unexalpha.c: Don't include varargs.h.

2003-09-16  Dave Love  <fx@gnu.org>

        * coding.h (ENCODE_UTF_8): New.

        * Makefile.in (gtkutil.o): Depend on coding.h.

        * coding.c (Fset_coding_system_priority): Doc fix.

2003-09-16  Kenichi Handa  <handa@m17n.org>

        * fileio.c (Finsert_file_contents): Call setup_coding_system in
        the case of auto saving.

2003-09-10  Andreas Schwab  <schwab@suse.de>

        * chartab.c (map_char_table): Protect `range' from GC.
        (map_char_table_for_charset): Likewise.

2003-07-09  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding_sjis): Check bytes more rigidly.

2003-06-26  Kenichi Handa  <handa@m17n.org>

        * fileio.c (choose_write_coding_system): Return a decided coding
        system.
        (Fwrite_region): Set Vlast_coding_system_used to the return value
        of choose_write_coding_system.

2003-06-06  Kenichi Handa  <handa@m17n.org>

        * charset.c (Fset_charset_priority): Pay attention to duplicated
        arguments.

        * coding.c (QCcategory): New variable.
        (syms_of_coding): Defsym it.  Set all elements of
        Vcoding_category_table and their symbol values.
        (Fset_coding_system_priority): Doc fix.  Update symbol qvalues of
        coding-category-XXX, and coding-category-list.
        (Fdefine_coding_system_internal): Add category in the plist.

2003-06-05  Kenichi Handa  <handa@m17n.org>

        * callproc.c (Fcall_process): Handle carryover correctly.

        * coding.c (decode_coding_iso_2022): Fix handling of invalid
        bytes.
        (raw_text_coding_system): Check NILP (coding_system).
        (coding_inherit_eol_type): Check NILP (coding_system) and
        NILP (parent).
        (consume_chars): Fix for the case of raw-text. 

        * process.c (read_process_output): Handle carryover correctly.

2003-06-02  Dave Love  <fx@gnu.org>

        * regex.c (re_search_2): Fix last change.

2003-05-30  Kenichi Handa  <handa@m17n.org>

        * regex.c (GET_CHAR_BEFORE_2): Check multibyte, not
        target_multibyte.  Even in a unibyte case, return a converted
        multibyte char.
        (GET_CHAR_AFTER): New macro.
        (PATFETCH): Translate via multibyte char.
        (HANDLE_UNIBYTE_RANGE): Delete this macro.
        (SETUP_MULTIBYTE_RANGE): New macro.
        (regex_compile): Setup compiled code so that its multibyteness
        matches that of a target.  Fix the handling of "[X-YZ]" using
        SETUP_MULTIBYTE_RANGE.
        (analyse_first) <charset>: For filling fastmap for all multibyte
        characters, don't check by BASE_LEADING_CODE_P.
        (re_search_2): Don't check RE_TARGET_MULTIBYTE_P (bufp).  It is
        the same as RE_MULTIBYTE_P (bufp) now.
        (mutually_exclusive_p): Check by (! multibyte ||
        IS_REAL_ASCII (c)).
        (TARGET_CHAR_AND_LENGTH): Delete this macro.
        (TRANSLATE_VIA_MULTIBYTE): New macro.
        (re_match_2_internal): Don't check RE_TARGET_MULTIBYTE_P (bufp).
        It is the same as RE_MULTIBYTE_P (bufp) now.
        <exactn>: Translate via multibyte.
        <anychar>: Fetch a character by RE_STRING_CHAR_AND_LENGTH.  Don't
        translate it.
        <charset, charset_not>: Fetch a character by
        RE_STRING_CHAR_AND_LENGTH.  Translate via multibyte.
        <duplicate>: Call bcmp_translate with the last arg `multibyte'.
        <wordbound, notwordbound, wordbeg, wordend, syntaxspec,
        notsyntaxspec, categoryspec, notcategoryspec> Fetch a character
        by GET_CHAR_AFTER.
        (bcmp_translate):  Likewise.
        
        * search.c (compile_pattern): Check the member target_multibyte,
        not the member multibyte of buf.

        * lread.c (read1): While reading a string, set force_singlebyte
        and force_multibyte correctly.

        * charset.c (Fset_unibyte_charset): Fix setting up of
        unibyte_to_multibyte_table.
        (init_charset_once): Likewise.

2003-05-29  Kenichi Handa  <handa@m17n.org>

        * coding.c (setup_coding_system): If coding has
        post-read-conversion or pre-write-conversion, set
        CODING_REQUIRE_DECODING_MASK and CODING_REQUIRE_ENCODING_MASK
        respectively.
        (decode_coding_gap): Run post-read-conversion if any.

        * fileio.c (Finsert_file_contents): Even if we read into a
        unibyte buffer, check if we must decode the result or not.

2003-05-29  Kenichi Handa  <handa@m17n.org>

        * coding.c (make_conversion_work_buffer): Change the work buffer
        name to the same one as that of Emacs 21.

2003-05-28  Kenichi Handa  <handa@m17n.org>

        * coding.h (make_conversion_work_buffer): Prototype adjusted.
        (code_conversion_restore): Don't extern it.

        * coding.c (detected_mask): Delete unused variable.
        (decode_coding_iso_2022): Pay attention to the byte sequence of
        CTEXT extended segment, and retain those bytes as is.
        (decode_coding_ccl): Delete unused variable `valids'.
        (setup_coding_system): Delete unused variable `category'.
        (consume_chars): Delete unused variable `category'.  Make it work
        for non-multibyte case.
        (make_conversion_work_buffer): Argument changed.
        (saved_coding): Delete unused variable.
        (code_conversion_restore): Don't check saved_coding->destination.
        (code_conversion_save): New function.
        (decode_coding_gap, encode_coding_gap): Call code_conversion_save
        instead of record_unwind_protect.
        (decode_coding_object, encode_coding_object): Likewise.  Recover
        PT.
        (detect_coding_system): Delete unused variable `mask'.
        (Fdefine_coding_system_internal): Delete unsed vaiable id;

        * fileio.c (kill_workbuf_unwind): New function.
        (Finsert_file_contents): On replacing, call
        make_conversion_work_buffer with correct args, and call
        record_unwind_protect with the first arg kill_workbuf_unwind.

        * lisp.h (Fgenerate_new_buffer_name): EXFUN it.

2003-05-20  Kenichi Handa  <handa@m17n.org>

        * fontset.c (BASE_FONTSET_P): Check FONTSET_BASE, not
        FONTSET_NAME.
        (fontset_add): Fix for the case that TO is less than TO1.
        (Ffontset_info): Don't use fallback fontset on checking the
        default fontset.
        (dump_fontset): New function for debugging.

        * coding.c (Fdefine_coding_system_internal): Fix for the case that
        coding_type is Qcharset.

2003-05-07  Kenichi Handa  <handa@m17n.org>

        * chartab.c (map_sub_char_table): New argument DEFAULT_VAL.
        (map_char_table): Don't inherit the value from the parent on
        initializing VAL.  Adjusted for the above change.

2003-05-06  Kenichi Handa  <handa@m17n.org>

        * coding.c (Qsignature, Qendian): Delete these variables.
        (syms_of_coding): Don't initialize them.
        (CATEGORY_MASK_UTF_16_AUTO): New macro.
        (detect_coding_utf_16): Add CATEGORY_MASK_UTF_16_AUTO in
        detect_info->found.
        (decode_coding_utf_16): Don't detect BOM here.
        (encode_coding_utf_16): Produce BOM if CODING_UTF_16_BOM (coding)
        is NOT utf_16_without_bom.
        (setup_coding_system): For a coding system of type utf-16, check
        if the attribute :endian is Qbig or not (not nil or not), and set
        CODING_REQUIRE_DETECTION_MASK if BOM detection is required.
        (detect_coding): If coding type is utf-16 and BOM detection is
        required, detect it.
        (Fdefine_coding_system_internal): For a coding system of type
        utf-16, check if the attribute :endian is Qbig or not (not nil or
        not).

2003-05-06  Kenichi Handa  <handa@m17n.org>

        * coding.c (coding_set_source): Fix for the case that the current
        buffer is different from coding->src_object.
        (decode_coding_object): Don't use the conversion work buffer if
        DST_OBJECT is a buffer.

2003-05-04  Dave Love  <fx@gnu.org>

        * lread.c (read_emacs_mule_char) [len==2]: Index
        emacs_mule_charset correctly.

2003-02-16  Dave Love  <fx@gnu.org>                     

        * coding.c (Qbig5, Vbig5_coding_system, CATEGORY_MASK_BIG5)
        (detect_coding_big5, decode_coding_big5, encode_coding_big5)
        (Fdecode_big5_char, Fencode_big5_char): Deleted.  (Big5 no longer
        treated specially.)
        (setup_coding_system, coding_category, CATEGORY_MASK_ANY)
        (detected_mask): Remove Big5 bits.

2003-04-09  Kenichi Handa  <handa@m17n.org>

        The following changes are to make the font rescaling facility
        compatible with Emacs 21.

        * xfaces.c (Vface_font_rescale_alist): Renamed from
        Vface_resizing_fonts.
        (struct font_name): Rename member resizing_ratio to rescale_ratio.
        (font_rescale_ratio): Renamed from font_resizing_ratio.
        (split_font_name): Set font->rescale_ratio.
        (better_font_p): Pay attention to font->rescale_ratio.
        (build_scalable_font_name): Likewise.  Change RESX, and RESY
        fields.
        (syms_of_xfaces): Declare Vface_font_rescale_alist as a Lisp
        variable.

2003-03-28  Kenichi Handa  <handa@m17n.org>

        * coding.c (Qutf_16_be_nosig, Qutf_16_be, Qutf_16_le_nosig)
        (Qutf_16_le): Remove these variables.
        (syms_of_coding): Don't DEFSYM them.
        (decode_coding_utf_16): Fix handling of BOM.
        (encode_coding_utf_16): Fix handling of BOM.

2003-03-14  Kenichi Handa  <handa@m17n.org>

        * fileio.c (Finsert_file_contents): On replacing, before decoding
        the file into the work buffer, set point of the work buffer to the
        end.

2003-02-13  Dave Love  <fx@gnu.org>

        * coding.c (Fcheck_coding_systems_region): Fix type errors.

2003-02-04  Dave Love  <fx@gnu.org>

        * xterm.c (XTread_socket): Check Lisp types for Vx_keysym_table
        and fix C types.

2003-01-31  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (SKIP_GLYPHS): New macro.
        (set_cursor_from_row): Pay attention to string display properties.

        * category.c (copy_category_entry): Fix for the case that RANGE
        is an integer.

        * xterm.c (x_encode_char): Call ccl_driver with the last arg Qnil.

        * w32term.c (w32_encode_char): Call ccl_driver with the last arg
        Qnil.

2003-01-30  Kenichi Handa  <handa@m17n.org>

        * charset.c (Fcharset_id_internal): New function.
        (syms_of_charset): Defsubr it.

        * coding.c (decode_coding_ccl, encode_coding_ccl): Call ccl_driver
        with the last arg charset_list acquired from coding.
        (Fdefine_coding_system_internal): For ccl-based coding system, fix
        the attribute coding_attr_ccl_valids.

        * coding.h (enum define_coding_ccl_arg_index): Set the first
        member coding_arg_ccl_decoder to coding_arg_max.

        * ccl.h (ccl_driver): Prototype adjusted.

        * ccl.c (CCL_DECODE_CHAR, CCL_ENCODE_CHAR): New macros.
        (ccl_driver): New arg CHARSET_LIST.  Use the above macros instead
        of DECODE_CAHR, ENCODE_CHAR, CHAR_CHARSET.
        (Fccl_execute): Call ccl_driver with the last arg Qnil.
        (Fccl_execute_on_string): Likewise.

2003-01-11  Kenichi Handa  <handa@m17n.org>

        * charset.h (ENCODE_CHAR): If the method is SUBSET or SUPERSET,
        call encode_char.

        * charset.c (encode_char): Fix handling of methods SUBSET and
        SUPERSET.

        * xterm.c (x_new_fontset): Fix previous change.

2003-01-10  Dave Love  <fx@gnu.org>

        * composite.c (syms_of_composite): Make composition_hash_table
        weak.

2003-01-10  Kenichi Handa  <handa@m17n.org>

        * dispextern.h (check_face_attributes, generate_ascii_font_name)
        (font_name_registry): Don't extern them.
        (split_font_name_into_vector, build_font_name_from_vector): Extern
        them.

        * fontset.h (Qfontset): Don't extern it.
        (new_fontset_from_font_name): Extern it.

        * fontset.c: Give 8 extra slots to fontset objects.
        (Qfontset_info): New variable.
        (syms_of_fontset): Defsym it.
        (FONTSET_FALLBACK): New macro.
        (fontset_face): Try also the default fontset.
        (make_fontset): Realize a fallback fontset from the default
        fontset.
        (generate_ascii_font_name): Moved from xfaces.c.  Rewritten by
        using split_font_name_into_vector and build_font_name_from_vector.
        (Fset_fontset_font): Access the elements of font_spec by enum
        FONT_SPEC_INDEX.  If font_spec is a string, extract the registry
        name by using split_font_name_into_vector.
        (Fnew_fontset): If no ASCII font is specified in FONTLIST,
        generate a proper font name from the fontset name.  Update
        Vfontset_alias_alist.
        (n_auto_fontsets): New variable.
        (new_fontset_from_font_name): New function.
        (Ffont_info): Store the information about fonts generated from the
        default fontset in the first extra slot of the returned
        char-table.

        * xfaces.c (generate_ascii_font_name): Moved to fontset.c.
        (font_name_registry): Function deleted.
        (split_font_name_into_vector): New function.
        (build_font_name_from_vector): New function.
        (font_list): The argument REGISTRY is now a list of registry
        names.
        (choose_face_font): If we are choosing an ASCII font, and ATTRS
        specifies an explicit font name, return the name as is.  Make a
        list of registy names.

        * xfns.c (x_set_font, x_create_tip_frame): Adjusted to the change
        of x_new_fontset.
        (Fx_create_frame): Don't call x_new_fontset here.  Just use
        x_list_fonts to check the existence of fonts.

        * xterm.h (x_new_fontset): Prototype adjusted.

        * xterm.c (x_new_fontset): Change the arg FONTSETNAME to Lisp
        string.  Use new_fontset_from_font_name to create a fontset from a
        font name.

2003-01-07  Dave Love  <fx@gnu.org>

        * Makefile.in: Fix some dependencies.

        * keymap.c (Fapropos_internal): Don't gcpro apropos_predicate but
        set it to nil before returning.

        * composite.c (update_compositions): Fix type error.

        * syntax.c (skip_chars, skip_syntaxes): Fix type errors.

2003-01-07  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_new_font): Optimize for the case that the font is
        already set for the frame.

2003-01-06  Kenichi Handa  <handa@m17n.org>

        * chartab.c (char_table_ascii): Check if the char table contents
        is sub-char-table or not.
        (char_table_set): Fix argument to char_table_ascii.
        (char_table_set_range): Likewise.

        * coding.c (CATEGORY_MASK_RAW_TEXT): New macro.
        (detect_coding_utf_8, detect_coding_utf_16)
        (detect_coding_emacs_mule, detect_coding_iso_2022)
        (detect_coding_sjis, detect_coding_big5)
        (detect_coding_ccl, detect_coding_charset): Change argument MASK
        to DETECT_INFO.  Update DETECT_INFO and return 1 if the byte
        sequence is valid in this coding system.  Callers changed.
        (MAX_ANNOTATION_LENGTH): New macro.
        (ADD_ANNOTATION_DATA): New macro.
        (ADD_COMPOSITION_DATA): Argument changed.  Callers changed.  Call
        ADD_ANNOTATION_DATA.  The format of annotation data changed.
        (ADD_CHARSET_DATA): New macro.
        (emacs_mule_char): New argument ID.  Callers changed.
        (decode_coding_emacs_mule, decode_coding_iso_2022)
        (decode_coding_sjis, decode_coding_big5, decode_coding_charset):
        Produce charset annotation data in coding->charbuf.
        (encode_coding_emacs_mule, encode_coding_iso_2022): Pay attention
        to charset annotation data in coding->charbuf.
        (setup_coding_system): Add CODING_ANNOTATE_CHARSET_MASK
        coding->common_flags if the coding system is iso-2022 based and
        uses designation.
        (produce_composition): Adjusted for the new annotation data
        format.
        (produce_charset): New function.
        (produce_annotation): Handle charset annotation.
        (handle_composition_annotation, handle_charset_annotation): New
        functions.
        (consume_chars): Handle charset annotation.  Utilize the above two
        functions.
        (encode_coding_object): If SRC_OBJECT and DST_OBJECT are the same
        buffer, get the deleted text as a string and set
        coding->src_object to that string.
        (detect_coding, detect_coding_system): Use the new struct
        coding_detection_info.

        * coding.h (struct coding_detection_info): New structure.
        (struct coding_system): Prototype of the member `detector'
        adjusted.
        (CODING_ANNOTATE_CHARSET_MASK): New macro.

2003-01-06  Kenichi Handa  <handa@m17n.org>

        * insdel.c (insert_from_gap): Fix argument to offset_intervals.

2003-01-03  Dave Love  <fx@gnu.org>

        * keymap.c (apropos_predicate, apropos_accumulate): Declare
        static.
        (Fapropos_internal): Don't gcpro apropos_accumulate.  Set result
        to new local and nullify apropos_accumulate before returning.
        (syms_of_keymap): Staticpro and initialize apropos_accumulate.

2002-12-05  Kenichi Handa  <handa@m17n.org>

        * charset.c (Fdefine_charset_internal): Setup charset.fast_map
        correctly.

2002-11-26  Dave Love  <fx@gnu.org>

        * fns.c (Flanginfo): Call synchronize_system_time_locale.

2002-11-07  Kenichi Handa  <handa@m17n.org>

        The following changes are to make character composition happen
        automatically on displaying.

        * Makefile.in (lisp, shortlisp): Add composite.elc

        * composite.h (Qauto_composed, Vauto_composition_function,
        Qauto_composition_function): Extern them.

        * composite.c (Vcomposition_function_table,
        Qcomposition_function_table): Delete variables.
        (Qauto_composed, Vauto_composition_function,
        Qauto_composition_function): New variables.
        (run_composition_function): Don't call
        compose-chars-after-function.
        (update_compositions): Clear `auto-composed' text property.
        (compose_chars_in_text): Delete this function.
        (syms_of_composite): Staticpro Qauto_composed and
        Qauto_composition_function.  Declare Vauto_composition_function as
        a Lisp variable.

        * dispextern.h (enum prop_idx): Add member AUTO_COMPOSED_PROP_IDX.

        * xdisp.c (it_props): Add an entry for Qauto_composed.
        (handle_auto_composed_prop): New function.

        * xselect.c (selection_data_to_lisp_data): Don't call
        compose_chars_in_text.

2002-11-06  Dave Love  <fx@gnu.org>

        * keyboard.c (read_char): Modify checking around use of
        Vkeyboard_translate_table.

        * xterm.c (XTread_socket): Check Lisp types for Vx_keysym_table
        and fix C types.

2002-11-06  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding_utf_8): When eol_type is Qdos, handle
        the case that the last byte is '\r' correctly.
        (decode_coding_emacs_mule): Likewise.
        (decode_coding_iso_2022): Likewise.
        (decode_coding_sjis): Likewise.
        (decode_coding_big5): Likewise.
        (decode_coding_charset): Likewise.
        (produce_chars): Likewise.
        (decode_coding): Flushing out the unprocessed data correctly.
        (decode_coding_gap): Set CODING_MODE_LAST_BLOCK bit of
        coding->mode.

2002-10-31  Dave Love  <fx@gnu.org>

        * xterm.c (XTread_socket): Fix changes for defined keysyms.  Add
        XK_ISO... case.
        (xaw_scroll_callback): Revert last change.

2002-10-30  Kenichi Handa  <handa@m17n.org>

        * charset.c (Fset_charset_priority): Update
        Viso_2022_charset_list.

2002-10-29  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (Vface_resizing_fonts): New variable.
        (struct font_name): New member `resizing_ratio'.
        (font_resizing_ratio): New function.
        (split_font_name): Set font->resizing_ratio.
        (better_font_p): Pay attention to font->resizing_ratio.
        (build_scalable_font_name): Likewise.  Don't change POINT_SIZE,
        RESX, and RESY fields.
        (try_alternative_families): Try scalable fonts if
        Vscalable_fonts_allowed is not Qt.
        (syms_of_xfaces): Declare Vface_resizing_fonts as a Lisp variable.

2002-10-29  Dave Love  <fx@gnu.org>

        * xterm.c (xaw_scroll_callback): Cast correctly.

2002-10-28  Dave Love  <fx@gnu.org>

        * keyboard.c (lispy_accent_codes, lispy_accent_keys): Extend.
        (lispy_kana_keys): Comment out.
        (make_lispy_event) [XK_kana_A]: Comment out.

        * xterm.c (xaw_scroll_callback): Cast call_data.
        (XTread_socket): Deal with ASCII keysyms.
        (syms_of_xterm) <Vx_keysym_table>: Fix args of make_hash_table.

2002-10-27  Dave Love  <fx@gnu.org>

        * xterm.c (Vx_keysym_table): New.
        (syms_of_xterm): Initialize it.
        (XTread_socket): Use it.
        From head: Eliminate incorrect optimization that tried to avoid
        decoding the output of X*LookupString.
        (x_get_font_repertory): Delete charset declaration.

2002-10-16  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding): Fix previous change.
        (detect_coding_charset): If only ASCII bytes are found, return 0.
        (detect_coding_system): Fix previous change.
        (Fdefine_coding_system_internal): Setup CODING_ATTR_ASCII_COMPAT
        (attrs) correctly.

2002-10-15  Dave Love  <fx@gnu.org>

        * coding.c (Fcheck_coding_system): Doc fix.

        * editfns.c (Finsert_byte): Return a proper value.

2002-10-14  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding): Fix args to translate_chars.  Pay
        attention to Vstandard_translation_table_for_decode.
        (encode_coding): Fix args to translate_chars.  Pay attention to
        Vstandard_translation_table_for_encode.

        * data.c (Faset): Check NEWELT by ASCII_CHAR_P, not by
        SINGLE_BYTE_CHAR_P.

        * editfns.c (general_insert_function): Check VAL by ASCII_CHAR_P,
        not by SINGLE_BYTE_CHAR_P.

        * fns.c (concat): Check CH by ASCII_CHAR_P, not by
        SINGLE_BYTE_CHAR_P.

        * insdel.c (copy_text): Check C by ASCII_CHAR_P, not by
        SINGLE_BYTE_CHAR_P.

        * keymap.c (Ftext_char_description): Check C by ASCII_CHAR_P, not
        by SINGLE_BYTE_CHAR_P.

        * search.c (Freplace_match): Check C by ASCII_CHAR_P, not by
        SINGLE_BYTE_CHAR_P.

2002-10-14  Dave Love  <fx@gnu.org>

        * fns.c (Fstring_as_multibyte, Fstring_to_multibyte): Doc fix.

2002-10-10  Dave Love  <fx@gnu.org>

        * fns.c (Flanginfo): Fix typo.

        * unexelf.c (unexec): Make last change conditional on Irix 6.5.

2002-10-10  Kenichi Handa  <handa@m17n.org>

        * coding.c (detect_coding_utf_8): Check incomplete byte sequence.
        Don't update *mask when correctly detected.
        (detect_coding_utf_16): Likewise.
        (detect_coding_emacs_mule): Likewise.
        (detect_coding_iso_2022): Likewise.
        (detect_coding_sjis): Likewise.
        (detect_coding_big5): Likewise.
        (detect_coding_ccl): Likewise.
        (decode_coding_sjis): Fix decoding of katakana-jisx0201.
        (detect_eol): Delete the argument CODING, and add the argument
        CATEGORY.
        (detect_coding): Adjusted for the changes above.
        (detect_coding_system): Likewise.

2002-10-09  Kenichi Handa  <handa@m17n.org>

        * character.c (char_string): Renamed from
        char_string_with_unification.  Pay attention to
        CHAR_MODIFIER_MASK.
        (string_char): Renamed from string_char.

        * character.h (CHAR_STRING): Call char_string if C is greater than
        MAX_3_BYTE_CHAR.
        (CHAR_STRING_ADVANCE): Likewise.
        (STRING_CHAR): Call string_char instead of
        string_char_with_unification.
        (STRING_CHAR_AND_LENGTH): Likewise.
        (STRING_CHAR_ADVANCE): Likewise.

2002-10-09  Dave Love  <fx@gnu.org>

        * coding.c (decode_coding_utf_8): Treat surrogates as invalid.

2002-10-07  Kenichi Handa  <handa@m17n.org>

        * keymap.c (push_key_description): Pay attention to
        force_multibyte.

        * regex.c (re_search_2): Fix for the case of unibyte buffer.

2002-10-06  Dave Love  <fx@gnu.org>

        * charset.c (define_charset_internal): Rename `supprementary'.

        * Makefile.in (lisp, shortlisp): Remove latin-N.

2002-10-05  Dave Love  <fx@gnu.org>

        * xfns.c (x_window, x_window): Use use_xim.

        * xterm.c (use_xim): Initialize.
        (xim_open_dpy, xim_initialize, xim_close_dpy): Use use_xim.
        (x_term_init): Maybe set use_xim.

        * xterm.h (use_xim) [HAVE_X_I18N]: Declare.

2002-10-01  Kenichi Handa  <handa@m17n.org>

        * search.c (search_buffer): Fix case-fold-search of multibyte
        characters.
        (boyer_moore): Rename the last argument to char_high_bits.

2002-09-27  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (display_string): Fix for the case of zero width glyph.

        * xfns.c (x_set_font): Change the error message of the case that
        x_new_fontset returns Qt.

        * xfaces.c (set_lface_from_font_name): Reject the default fontset.
        (Finternal_set_lisp_face_attribute): Use signal_error for the
        error of invalid fontset.

        * xterm.c (x_new_fontset): If FONTSETNAME specifies the default
        fontset, return Qt.

2002-09-19  Kenichi Handa  <handa@m17n.org>

        * regex.c (re_search_2): Fix previous change.

2002-09-18  Kenichi Handa  <handa@m17n.org>

        * syntax.c (skip_syntaxes): Fix previous change.

2002-09-13  Kenichi Handa  <handa@m17n.org>

        * syntax.c (skip_chars): Fix previous change.
        (skip_syntaxes): Fix previous change.

2002-09-06  Dave Love  <fx@gnu.org>

        * config.in: Restore it.

2002-09-05  Dave Love  <fx@gnu.org>

        * config.in: Removed (now auto-generated).

        * s/usg5-4.h: Fix last change.

        * unexelf.c (unexec): Make .got handling not SGI-specific.

        * syntax.c (syms_of_syntax) <multibyte-syntax-as-symbol>: Doc fix.

        * regex.c: Use `ifdef HAVE_ALLOCA_H', not `if HAVE_ALLOCA_H'.

        * keyboard.c (read_key_sequence): Fix type error.

        * buffer.c (Fset_buffer_multibyte, Fset_buffer_multibyte): Fix
        type error.

        * fontset.c (fontset_add): Return Lisp_Object.

2002-09-03  Dave Love  <fx@gnu.org>

        * charset.h (charset_ordered_list_tick): Declare extern.

2002-09-03  Kenichi Handa  <handa@m17n.org>

        The following changes (and some of 2002-08-20 changes of mine) are
        for handling syntax, category, and case conversion for unibyte
        characters by converting them to multibyte on the fly.  With these
        changes, we don't have to setup syntax and case tables for unibyte
        characters in each language environment.

        * abbrev.c (Fexpand_abbrev): Convert a unibyte character to
        multibyte if necessary.

        * bytecode.c (Fbyte_code): Likewise.

        * character.h (LEADING_CODE_LATIN_1_MIN) 
        (LEADING_CODE_LATIN_1_MAX): New macros.
        (unibyte_to_multibyte_table): Extern it.
        (unibyte_char_to_multibyte): New macro.
        (MAKE_CHAR_MULTIBYTE): Use unibyte_to_multibyte_table.
        (CHAR_LEADING_CODE): New macro.
        (FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE): New macro.

        * character.c (unibyte_to_multibyte_table): New variable.
        (unibyte_char_to_multibyte): Move to character.h and defined as
        macro.
        (multibyte_char_to_unibyte): If C is an eight-bit character,
        convert it to the corresponding byte value.

        * charset.c (Fset_unibyte_charset): If the dimension of CHARSET is
        not 1, singals an error.  Update the elements of
        unibyte_to_multibyte_table.
        (init_charset_once): Initialize unibyte_to_multibyte_table.
        (syms_of_charset): Define the charset `iso-8859-1'.

        * casefiddle.c (casify_object): Fix previous change.

        * cmds.c (internal_self_insert): In a multibyte buffer, insert C
        as is without converting it to unibyte.  In a unibyte buffer,
        convert C to multibyte before checking the syntax.

        * lisp.h (unibyte_char_to_multibyte): Extern deleted.

        * minibuf.c (Fminibuffer_complete_word): Use the macro
        FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE.

        * regex.h (struct re_pattern_buffer): New member target_multibyte.

        * regex.c (RE_TARGET_MULTIBYTE_P): New macro.
        (GET_CHAR_BEFORE_2): Check target_multibyte, not multibyte.  If
        that is zero, convert an eight-bit char to multibyte.
        (MAKE_CHAR_MULTIBYTE, CHAR_LEADING_CODE): New dummy new macros for
        non-emacs case.
        (PATFETCH): Convert an eight-bit char to multibyte.
        (HANDLE_UNIBYTE_RANGE): New macro.
        (regex_compile): Setup the compiled pattern for multibyte chars
        even if the given regex string is unibyte.  Use PATFETCH_RAW
        instead of PATFETCH in many places.  To handle `charset'
        specification of unibyte, call HANDLE_UNIBYTE_RANGE.   Use bitmap
        only for ASCII chars.
        (analyse_first) <exactn>: Simplified because the compiled pattern
        is multibyte.
        <charset_not>: Setup fastmap from bitmap only for ASCII chars.
        <charset>: Use CHAR_LEADING_CODE to get leading codes.
        <categoryspec>: If multibyte, setup fastmap only for ASCII chars
        here.
        (re_compile_fastmap) [emacs]: Call analyse_first with the arg
        multibyte always 1.
        (re_search_2) In emacs, set the locale variable multibyte to 1,
        otherwise to 0.  New local variable target_multibyte.  Check it
        to decide the multibyteness of STR1 and STR2.  If
        target_multibyte is zero, convert unibyte chars to multibyte
        before translating and checking fastmap.
        (TARGET_CHAR_AND_LENGTH): New macro.
        (re_match_2_internal): In emacs, set the locale variable multibyte
        to 1, otherwise to 0.  New local variable target_multibyte.  Check
        it to decide the multibyteness of STR1 and STR2.  Use
        TARGET_CHAR_AND_LENGTH to fetch a character from D.
        <charset, charset_not>: If multibyte is nonzero, check fastmap
        only for ASCII chars.   Call bcmp_translate with
        target_multibyte, not with multibyte.
        <begline>: Declare the local variable C as `unsigned'.
        (bcmp_translate): Change the last arg name to target_multibyte.

        * search.c (compile_pattern_1): Don't adjust the multibyteness of
        the regexp pattern and the matching target.  Set cp->buf.multibyte
        to the multibyteness of the regexp pattern.  Set
        cp->but.target_multibyte to the multibyteness of the matching
        target.
        (wordify): Use FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE instead of
        FETCH_STRING_CHAR_ADVANCE.
        (Freplace_match): Convert unibyte chars to multibyte.

        * syntax.c (char_quoted): Use FETCH_CHAR_AS_MULTIBYTE to convert
        unibyte chars to multibyte.
        (back_comment): Likewise.
        (scan_words): Likewise.
        (skip_chars): The arg syntaxp is deleted, and the code for
        handling syntaxes is moved to skip_syntaxes.  Callers changed.
        Fix the case that the multibyteness of STRING and the current
        buffer doesn't match.
        (skip_syntaxes): New function.
        (SYNTAX_WITH_MULTIBYTE_CHECK): Check C by ASCII_CHAR_P, not by
        SINGLE_BYTE_CHAR_P.
        (Fforward_comment): Use FETCH_CHAR_AS_MULTIBYTE to convert unibyte
        chars to multibyte.
        (scan_lists): Likewise.
        (Fbackward_prefix_chars): Likewise.
        (scan_sexps_forward): Likewise.

2002-08-23  Kenichi Handa  <handa@m17n.org>

        * xfaces.c (QCfontset): New variable.
        (LFACE_FONTSET): New macro.
        (check_lface_attrs): Check also LFACE_FONTSET_INDEX.
        (set_lface_from_font_name): Setup LFACE_FONTSET (lface).
        (Finternal_set_lisp_face_attribute): Handle QCfontset.
        (Finternal_get_lisp_face_attribute): Likewise.
        (lface_same_font_attributes_p): Fix checking of LFACE_FONT_INDEX,
        check also LFACE_FONTSET_INDEX.
        (face_fontset): Check attrs[LFACE_FONTSET_INDEX], not
        attrs[LFACE_FONT_INDEX].
        (syms_of_xfaces): Intern and staticpro QCfontset.

        * dispextern.h (enum lface_attribute_index): New member
        LFACE_FONTSET_INDEX.

        * fns.c (base64_encode_1): Handle eight-bit chars correctly.

2002-08-21  Kenichi Handa  <handa@m17n.org>

        * coding.c (coding_set_destination): Fix coding->destination for
        the case converting a region.
        (encode_coding_utf_8): Encode eight-bit chars as single byte.
        (encode_coding_object): Fix coding->dst_pos and
        coding->dst_pos_byte for the case converting a region.

        * insdel.c (insert_from_gap): Make it work even if PT != GTP.

        * character.h (BYTE8_STRING): New macro.

        * fns.c (base64_decode_1): Insert eight-bit chars correctly.

2002-08-20  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (get_next_display_element): Don't display unibyte 8-bit
        characters by octal form.

        * abbrev.c (Fexpand_abbrev): Fix for the multibyte case.

        * buffer.h (_fetch_multibyte_char_len): Extern deleted.
        (FETCH_MULTIBYTE_CHAR): Don't use _fetch_multibyte_char_len.
        (BUF_FETCH_MULTIBYTE_CHAR): Likewise.
        (FETCH_CHAR_AS_MULTIBYTE): New macro.

        * casetab.c (set_canon, set_identity, shuffle): Simplified.

        * casefiddle.c (casify_object): Simplified.  Handle the case that
        the case conversion change the byte length.  
        (casify_region): Likewise

        * character.h (MAKE_CHAR_UNIBYTE, MAKE_CHAR_MULTIBYTE): New
        macros.

        * character.c (_fetch_multibyte_char_len): This variable deleted.
        (syms_of_character): Setup Vprintable_chars.

        * editfns.c (Fchar_equal): Fix for the unibyte case.
        (Finsert_byte): New function.
        (syms_of_editfns): Defsubr it.

        * keyboard.c (read_key_sequence): Use ~CHAR_MODIFIER_MASK instead
        of direct code 0x3ffff.

        * search.c (Freplace_match): Fix for the unibyte case.

2002-08-19  Kenichi Handa  <handa@m17n.org>

        * lread.c (safe_to_load_p): Fix the logic.

        * syntax.c (scan_words): Don't treat characters belonging to
        different scripts as constituting a word.

        * editfns.c (Fformat): Use ASCII_CHAR_P, not SINGLE_BYTE_CHAR_P.

        * fontset.c (Fset_fontset_font): Treat `ascii' as charset, not
        script.

        * emacs.c (main): In the case of --unibyte, instead of aborting on
        finding non-empty buffer, make it unibyte.

2002-08-18  Kenichi Handa  <handa@m17n.org>

        * xterm.c (x_new_fontset): Call `create-fontset-from-ascii-font'
        to create a fontset.

2002-08-18  Dave Love  <fx@gnu.org>

        * character.c (Funibyte_char_to_multibyte): Doc fix.

        * xfns.c [HAVE_STDLIB_H]: Fix last change.

2002-08-15  Kenichi Handa  <handa@m17n.org>

        * fontset.c (fontset_add): Make the type `int'.
        (fontset_id_valid_p): Define it if FONTSET_DEBUG is defined.

        * character.c (unibyte_char_to_multibyte): Refer to
        charset_unibyte, not charset_primary.
        (multibyte_char_to_unibyte): Likewise.
        (Funibyte_char_to_multibyte): Likewise.

        * charset.h: (charset_unibyte): Extern it instead of
        charset_primary.

        * charset.c (charset_unibyte): Renamed from charset_primary.
        (Funibyte_charset): Renamed from Fprimary_charset.
        (Fset_unibyte_charset): Renamed from Fset_primary_charset.
        (syms_of_charset): Adjusted for the above changes.

        * w32term.c (x_produce_glyphs): Use ASCII_CHAR_P, not
        SINGLE_BYTE_CHAR_P.  Fix the logic of handling non-ASCII char when
        it->multibyte_p is zero.

        * lisp.h (nonascii_insert_offset, Vnonascii_translation_table):
        Extern deleted.

2002-08-08  Kenichi Handa  <handa@m17n.org>

        * coding.c (Fdefine_coding_system_internal): Fix category setting
        for a coding system of type iso-2022.

2002-08-02  Kenichi Handa  <handa@m17n.org>

        * fontset.h (FS_LOAD_FONT): Call fs_load_font with the arg CHARSET
        -1.

2002-08-01  Kenichi Handa  <handa@m17n.org>

        * syntax.c (Vnext_word_boundary_function_table): New variable.
        (syms_of_syntax): Declare it as a Lisp variable.
        (scan_words): Call functions in Vnext_word_boundary_function_table
        if any.

        * xterm.c (x_load_font): Initialize fontp->fontset to -1.

        * fontset.c (fs_load_font): If fontp->charset is not negative,
        return fontp without setting its members.

2002-07-31  Dave Love  <fx@gnu.org>

        * config.in: Generated with autoheader.

        * xfns.c [HAVE_STDLIB_H]: Change logic (instead of fixing typo).

        * m/sparc.h (HAVE_ALLOCA): Delete.

        * s/irix6-5.h: Don't include strings.h.
        (bcopy, bzero, bcmp): Don't undef.

        * s/irix6-0.h (bcopy, bzero, bcmp): Don't undef.

        * s/usg5-4.h (NO_SIOCTL_H): Don't define.
        (TIOCSIGSEND): Don't test IRIX6.
        (bcopy, bzero, bcmp): Define conditionally.

2002-07-31  Kenichi Handa  <handa@m17n.org>

        * buffer.c (Qas, Qmake, Qto): New variables.
        (Fset_buffer_multibyte): New optional arg METHOD.  Caller changed.
        (syms_of_buffer): Intern and staticpro Qas, Qmake, and Qto.

        * callproc.c (Fcall_process): Don't call insert_1_both directly if
        we are inserting a process output into a multibyte buffer.

        * character.h (CHAR_TO_BYTE8): If C is not eight-bit char, call
        multibyte_char_to_unibyte.

        * character.c (Funibyte_char_to_multibyte): If C can't be decoded
        by the primary charset, make it eight-bit char.
        (Fmultibyte_char_to_unibyte): Call CHAR_TO_BYTE8.

        * charset.c: (charset_eight_bit, Qeight_bit_control): New
        variables.
        (charset_8_bit__control, charset_8_bit_graphic,
        Qeight_bit_control, Qeight_bit_graphic): These variables deleted.
        (define_charset_internal): New function.
        (syms_of_charset): Call define_charset_internal for pre-defined
        charsets.

        * charset.h (charset_8_bit): Extern it.

        * coding.c (make_conversion_work_buffer): Adjusted for the change
        of Fset_buffer_multibyte.
        (encode_coding_raw_text): Increment p0 in the loop.

        * lisp.h (Fset_buffer_multibyte): Prototype adjusted.

        * xdisp.c (setup_echo_area_for_printing, set_message_1): Adjusted
        for the change of Fset_buffer_multibyte.

        * fns.c (Fstring_to_multibyte): New function.
        (syms_of_fns): Declare Fstring_to_multibyte as Lisp subroutine.

2002-07-30  Dave Love  <fx@gnu.org>

        * xfns.c (x_put_x_image): Declare args.

        * xfaces.c (font_name_registry, choose_face_font): Delete unused
        vars.
        (try_font_list): Declare an arg.

        * xdisp.c (message2_nolog, set_message): Declare an arg.

        * terminfo.c (tparam): Declare an arg.  Use P_ to declare tparm.

        * syntax.c (scan_sexps_forward): Declare an arg.

        * scroll.c (calculate_scrolling, calculate_direct_scrolling):
        Declare an arg.

        * lisp.h (Fnew_fontset): Declare.

        * keymap.c (push_key_description): Call CHARACTERP correctly.

        * fontset.c (fontset_add): Declare args.  Call make_number
        correctly.
        (face_for_char): Delete unused vars.
        (Fset_fontset_font): Doc fix.  Delete unused vars.

        * doc.c (Fsubstitute_command_keys): Delete unused vars.

        * composite.c (update_compositions): Declare arg.

        * cm.c (calccost, cmgoto): Declare args.

        * charset.c: Remove `emacs' conditional.  Doc fixes.
        (map_char_table_for_charset): Declare.

        * character.c (syms_of_character) <translation-table-vector>: Doc
        fix.

        * ccl.c: Remove `emacs' conditional.  Include hash table stuff
        from trunk.

2002-07-26  Kenichi Handa  <handa@m17n.org>

        The following changes are to allow specifying multiple font
        patterns for a character range (specified by script or charset).

        * Makefile.in (abbrev.o): Depend on syntax.h.
        (xfaces.o): Depend on charset.h.

        * alloc.c (Fmake_string): Use ASCII_CHAR_P, not
        SINGLE_BYTE_CHAR_P.

        * ccl.c (Fccl_execute_on_string): Add `const' to local variables.

        * character.h (Vchar_script_table): Extern it.

        * character.c (Vscript_alist): This variable deleted.
        (Vchar_script_table, Qchar_script_table): New variable.
        (syms_of_character): Declare Vchar_script_table as a lisp variable
        and initialize it.

        * chartab.c (Fmake_char_table): Doc fixed.  If PURPOSE doesn't
        have property char-table-extra-slots, make no extra slot.

        * dispextern.h (struct face): Member `charset' deleted.
        (FACE_SUITABLE_FOR_CHAR_P): Use ASCII_CHAR_P, not
        SINGLE_BYTE_CHAR_P.
        (FACE_FOR_CHAR): Likewise.
        (choose_face_font, lookup_non_ascii_face, font_name_registry): Add
        prototypes
        (lookup_face, lookup_named_face, lookup_derived_face): Prototype
        fixed.
        (generate_ascii_font_name): Renamed from generate_ascii_font.
        
        * fontset.h (get_font_repertory_func): New prototype.
        (make_fontset_for_ascii_face, fs_load_font): Prototypes fixed.
        (FS_LOAD_FONT): Call fs_load_font with the 3rd arg charset_ascii.

        * fontset.c (Qprepend, Qappend): New variables.
        (FONTSET_CHARSET_ALIST, FONTSET_FACE_ALIST): These macros deleted.
        (FONTSET_NOFONT_FACE, FONTSET_REPERTORY): New macros.
        (FONTSET_REF): Optimize if FONTSET is Vdefault_fontset.
        (FONTSET_REF_AND_RANGE, FONTSET_ADD): New macros.
        (fontset_ref_and_range, fontset_add, reorder_font_vector)
        (load_font_get_repertory): New functions.
        (fontset_set): This function deleted.
        (fontset_face): New arg FACE.  Return face ID, not face.
        Completely re-written to handle new fontset structure.  Caller
        changed.
        (free_face_fontset): Use ASET istead of AREF (X) = Y.
        (face_for_char): Don't call lookup_face.
        (make_fontset_for_ascii_face): New arg FACE.
        (fs_load_font): New arg CHARSET_ID.  Don't check
        Vfont_encoding_alist here.
        (find_font_encoding): New function.
        (list_fontsets): Use STRINGP, not ! NILP.
        (accumulate_script_ranges): New function.
        (Fset_fontset_font, Fnew_fontset, Ffontset_info): Completely
        re-written to handle new fontset structure.
        (Ffontset_font): Return a copy of element.
        (syms_of_fontset): Define symbols Qprepend and Qappend.  Fix
        docstring of font-encoding-alist.

        * lisp.h (CHAR_TABLE_REF): Remove unnecessary check (IDX >= 0).
        (Fset_fotset_font): Fix arguments to 5.

        * msdos.c (XMenuActivate): Adjuted for the change of
        lookup_derived_face.

        * xdisp.c (message_dolog, set_message_1, extend_face_to_end_of_line):
        Use ASCII_CHAR_P, not SINGLE_BYTE_CHAR_P.
        (highlight_trailing_whitespace): Adjusted for the change of
        lookup_named_face.

        * xfaces.c: Include charset.h.
        (load_face_font): Argument C deleted.  Caller changed.
        (generate_ascii_font_name): Renamed from generate_ascii_font.
        (font_name_registry): New function.
        (cache_face): Store ascii faces before non-ascii faces in buckets.
        (lookup_face): Arguments C and BASE_FACE deleted.  Caller changed.
        Lookup only ascii faces.
        (lookup_non_ascii_face): New function.
        (lookup_named_face): Argument C deleted.  Caller changed.
        (lookup_derived_face): Argument C deleted.  Caller changed.
        (try_font_list): New arg PATTERN.  Caller changed.  If PATTERN is
        a string, just call font_list with it.
        (choose_face_font): Arguments FACE and C deleted.  New arg
        FONT_SPEC.  Caller changed.
        (realize_face): Arguments C and BASE_FACE deleted.  Caller
        (realize_x_face): Likewise.
        (realize_non_ascii_face): New function.
        (realize_x_face): Call load_face_font here.
        (realize_tty_face): Argument C deleted.  Caller changed.
        (compute_char_face): If CH is not ascii, call FACE_FOR_CHAR to
        get a face ID.
        (dump_realized_face): Don't print charset of FACE.

        * xfns.c (x_set_font): Always call x_new_fontset and
        store_frame_parameter.
        (Fx_create_frame): Call x_new_fontset, not x_new_font.
        (syms_of_xfns): Set get_font_repertory_func to
        x_get_font_repertory.

        * xterm.h (x_get_font_repertory): Extern it.

        * xterm.c (x_produce_glyphs): Use ASCII_CHAR_P, not
        SINGLE_BYTE_CHAR_P.  Fix the logic of handling non-ASCII char when
        it->multibyte_p is zero.
        (XTread_socket): Use ASCII_CHAR_P, not SINGLE_BYTE_CHAR_P.
        (x_new_fontset): If FONTSETNAME doesn't match any existing
        fontsets, create a new one.
        (x_get_font_repertory): New function.

2002-07-25  Kenichi Handa  <handa@m17n.org>

        * coding.c (Ffind_coding_systems_region_internal): Detect an
        ASCII only string correctly.

        * lread.c (Fload): Don't load with Qload_force_doc_strings t if
        version is 0.

2002-07-24  Kenichi Handa  <handa@m17n.org>

        * lread.c: Include "coding.h".
        (Qget_emacs_mule_file_char, Qload_force_doc_strings,
        load_each_byte, unread_char): New variables.
        (readchar_backlog): This variable deleted.
        (readchar): Return a character unless load_each_byte is nonzero.
        Handle the case that readcharfun is Qget_emacs_mule_file_char or a
        cons.  If unread_char is not -1, simply return it.
        (unreadchar): Handle the case that readcharfun is
        Qget_emacs_mule_file_char or a cons.  Set unread_char if
        necessary.
        (read_multibyte): This function deleted.
        (readbyte_for_lambda, readbyte_from_file, readbyte_from_string)
        (read_emacs_mule_char): New functions.
        (Fload): Even if the file doesn't have the extention ".elc", if
        safe_to_load_p returns a positive version number, assume that the
        file contains bytecompiled code.  If the version is less than 22,
        load the file while decoding multibyte sequences by emacs-mule.
        (readevalloop): Don't use readchar_backlog.
        (Fread): Likewise.  Pay attention to the case that STREAM is a
        cons.
        (Fread_from_string): Pay attention to the case that STREAM is a
        cons.
        (read_escape): The arg BYTEREP deleted.
        (read1): Set load_each_byte to 1 temporarily while handling
        #@NUMBER.  Don't call read_multibyte.
        (read_vector): Call Fread with a cons.  If readcharfun is
        Qget_emacs_mule_file_char, decode the read string by emacs-mule.
        (read_list): If doc_reference is 2, make the cdr part string as
        unibyte.
        (syms_of_lread): Intern and staticpro Qget_emacs_mule_file_char
        and Qload_force_doc_strings.

2002-07-23  Kenichi Handa  <handa@m17n.org>

        * xdisp.c (face_before_or_after_it_pos): Call
        FETCH_MULTIBYTE_CHAR with byte postion, not char position.

2002-07-22  Kenichi Handa  <handa@m17n.org>

        * character.h (TRAILING_CODE_P): New macro.
        (MAYBE_UNIFY_CHAR): Adjusted for the change of Funify_charset.
        (string_char_with_unification): Fix prototype.
        (Vscript_alist): Extern it.

        * character.c (Vscript_alist): New variable.
        (string_char_with_unification): Add `const' to local variables.
        (str_as_unibyte): Likewise.
        (string_escape_byte8): Likewise.
        (syms_of_character): Declare script-alist as a Lisp variable.

        * charset.h (Vcharset_ordered_list): Extern it.
        (charset_ordered_list_tick): Extern it.
        (EMACS_MULE_LEADING_CODE_PRIVATE_11)
        (EMACS_MULE_LEADING_CODE_PRIVATE_12)
        (EMACS_MULE_LEADING_CODE_PRIVATE_21)
        (EMACS_MULE_LEADING_CODE_PRIVATE_22): New macros
        (Funify_charset): Adjusted for the change of Funify_charset.

        * charset.c (charset_ordered_list_tick): New variable.
        (Fdefine_charset_internal): Increment charset_ordered_list_tick.
        (Funify_charset): New optional arg DEUNIFY.  If it is non-nil,
        deunify intead of unify a charset.
        (string_xstring_p): Add `const' to local variables.
        (find_charsets_in_text): Add `const' to arguemnts and local
        variables.
        (encode_char): Adjusted for the change of Funify_charset.  Fix
        detecting of invalid code.
        (Fset_charset_priority): Increment charset_ordered_list_tick.
        (Fmap_charset_chars): Fix handling of default value for FROM_CODE
        and TO_CODE.

        * coding.c (LEADING_CODE_PRIVATE_11, LEADING_CODE_PRIVATE_12)
        (LEADING_CODE_PRIVATE_21, LEADING_CODE_PRIVATE_22): Macros
        deleted.  Callers changed to use
        EMACS_MULE_LEADING_CODE_PRIVATE_11, etc.
        (decode_coding_ccl): Add `const' to local variables.
        (consume_chars): Likewise.
        (Ffind_coding_systems_region_internal): Likewise.
        (Fcheck_coding_systems_region): Likewise.

        * print.c (print_object): Use octal form for printing the
        contents of a bool vector.

2002-07-18  Dave Love  <fx@gnu.org>

        * lread.c (Fload) <!load_dangerous_libraries>: Don't leak fd.
        <version == 20>: Refuse to load.

2002-07-17  Dave Love  <fx@gnu.org>

        * fns.c: Move coding.h.
        (Qcodeset, Qdays, Qmonths): New.
        (concat): Use CHARACTERP instead of INTERGERP.
        (Flocale_codeset): Deleted.
        (Flanginfo): New function.
        (syms_of_fns): Changed accordingly.

        * coding.c (adjust_coding_eol_type): Fix eol_type/eol_seen mixup.

2002-07-16  Dave Love  <fx@gnu.org>

        * casetab.c (init_casetab_once, init_casetab_once): Fix
        CHAR_TABLE_SET call.

        * category.c (Fmodify_category_entry): Fix CATEGORY_MEMBER call.

        * character.c (syms_of_character): Fix CHAR_TABLE_SET call.

        * charset.c (Fmap_charset_chars): Check args.  Convert Lisp types.
        (load_charset_map, Fdeclare_equiv_charset, Fencode_char)
        (Fset_charset_priority, syms_of_charset): Convert Lisp types.

        * charset.h (CHECK_CHARSET_GET_ID): Use XINT on AREF result.

        * coding.c (ENCODE_DESIGNATION, decode_eol)
        (make_conversion_work_buffer, code_conversion_restore)
        (Fdefine_coding_system_internal): Convert Lisp types.
        (code_conversion_restore): Use EQ, not ==.
        (Fencode_coding_string): Fix code_convert_string call.

        * coding.h (code_convert_region): Fix prototype.

        * dispextern.h (redraw_frame, redraw_garbaged_frames): Removed.

        * fontset.c (fontset_ref, fontset_set, fs_load_font) 
        (Ffontset_info): Convert Lisp types.

        * syntax.h (SYNTAX_ENTRY_INT): Don't use make_number.

        * xterm.c (note_mouse_movement): Fix call of window_from_coordinates.

        * xdisp.c (display_mode_element): Fix call of Fset_text_properties.

        * chartab.c: Include "...h", not <...h> in some cases.

        * callproc.c (Fcall_process): Remove unused variables.

2002-07-12  Dave Love  <fx@gnu.org>

        * coding.c (Fset_coding_system_priority): Allow null arg list.

2002-07-03  Dave Love  <fx@gnu.org>

        * minibuf.c (Fminibuffer_complete_word): Remove unused var.
        (Fself_insert_and_exit): Use CHARACTERP.

        * callproc.c (Fcall_process): Remove unused vars.

        * xterm.c (XTread_socket): Add extra dead keysyms.

        * xdisp.c (decode_mode_spec_coding): Use CHARACTERP.

        * dispextern.h: Remove prototypes for redraw_frame,
        redraw_garbaged_frames.

        * cmds.c (Fself_insert_command): Use CHARACTERP.

        * chartab.c (make_sub_char_table): Remove unused var.
        (Fset_char_table_default, Fmap_char_table): Doc fix.

        * keymap.c (access_keymap): Remove generic char code.
        (push_key_description): Use CHARACTERP.

2002-07-01  Dave Love  <fx@gnu.org>

        * charset.c: Doc fixes.
        (Funify_charset): Extra checking.

2002-06-24  Dave Love  <fx@gnu.org>

        * lread.c: Remove some unused variables.
        (safe_to_load_p): If safe, return the magic number version byte.
        (Fload): Maybe use load-with-code-conversion.

2002-06-12  Kenichi Handa  <handa@m17n.org>

        * category.c (Fmodify_category_entry): Don't modify the contents
        of category_set for characters out of the range.  Avoid
        unnecessary modification.

        * character.h (MAYBE_UNIFY_CHAR): Adjusted for the change of
        Vchar_unify_table.  The default value of the table is now nil.

        * character.c (syms_of_character): Setup Vchar_width_table for
        eight-bit-control and raw-byte chars.

        * charset.h (enum define_charset_arg_index): Delete
        charset_arg_parents and add charset_arg_subset and
        charset_arg_superset.
        (enum charset_attr_index): Delete charset_parents and add
        charset_subset and charset_superset.
        (enum charset_method): Delete CHARSET_METHOD_INHERIT and add
        CHARSET_METHOD_SUBSET and CHARSET_METHOD_SUPERSET.
        (CHARSET_ATTR_PARENTS, CHARSET_PARENTS): Macros deleted.
        (CHARSET_ATTR_SUBSET, CHARSET_ATTR_SUPERSET, CHARSET_SUBSET)
        (CHARSET_SUPERSET): New macros.
        (charset_work): Extern it.
        (ENCODE_CHAR): Use charset_work.
        (CHAR_CHARSET_P): Adjusted for the change of encoder format.
        (map_charset_chars): Extern it.

        * charset.c (load_charset_map): Set the default value of encoder
        and deunifier char-tables to nil.
        (map_charset_chars): Argument changed.  Callers changed.  Use
        map_char_table_for_charset instead of map_char_table.
        (Fmap_charset_chars): New optional args from_code and to_code.
        (Fdefine_charset_internal): Adjusted for the change of
        `define-charset' (:parents -> :subset or :superset).
        (charset_work): New variable.
        (encode_char): Adjusted for the change of
        Fdefine_charset_internal.
        (syms_of_charset): Likewise.
        (Ffind_charset_string): Setup the vector `charsets' correctly.

        * chartab.c (sub_char_table_ref_and_range): New arg defalt.  Fix
        the previous change.
        (char_table_ref_and_range): Adjusted for the above change.
        (map_sub_char_table_for_charset): New function.
        (map_char_table_for_charset): New function.

        * keymap.c (describe_vector): Handle a char-table directly here.
        (describe_char_table): Deleted.

        * lisp.h (map_charset_chars): Deleted.

2002-06-11  Dave Love  <fx@gnu.org>

        * fns.c (count_combining): Comment out (unused).
        (Flocale_codeset): New.
        (syms_of_fns): Defsubr it.

        * config.in (HAVE_PTY_H, HAVE_SIZE_T, HAVE_LANGINFO_CODESET): New.
        (size_t): Removed.
        
2002-06-06  Dave Love  <fx@gnu.org>

        * Makefile.in (chartab.o): Depend on charset.h

2002-06-03  Kenichi Handa  <handa@m17n.org>

        * character.c (syms_of_character): Set the default value of
        Vprintable_chars to Qnil.

2002-05-31  Dave Love  <fx@gnu.org>

        * Makefile.in (lisp, shortlisp): Change indian.elc to indian.el.

2002-05-31  Kenichi Handa  <handa@m17n.org>

        * charset.c (load_charset_map): Handle the case that from < to
        correctly.

        * coding.c (encode_coding_emacs_mule): Pay attention to raw-8-bit
        chars.
        (encode_coding_iso_2022): Likewise.
        (encode_coding_sjis): Likewise.
        (encode_coding_big5): Likewise.
        (encode_coding_charset): Likewise.

2002-05-30  Kenichi Handa  <handa@m17n.org>

        * Makefile.in (lisp): Change chinese.elc to chinese.el.  They are
        not bytecompiled now.
        (shortlisp): Likewise.

        * charset.c (charset_jisx0201_roman, charset_jisx0208_1978)
        (charset_jisx0208): New variables.
        (Fdefine_charset_internal): Setup them if appropriate.
        (init_charset_once): Initialize them to -1.

        * charset.h (charset_jisx0201_roman, charset_jisx0208_1978,
        charset_jisx0208): Extern them.

        * coding.c (CODING_ISO_FLAG_USE_ROMAN): New macro
        (CODING_ISO_FLAG_USE_OLDJIS): New macro.
        (CODING_ISO_FLAG_FULL_SUPPORT): Macro definition changed.
        (setup_iso_safe_charsets): Fix arguemtns to Fassq.
        (DECODE_DESIGNATION): Pay attention to CODING_ISO_FLAG_USE_ROMAN
        and CODING_ISO_FLAG_USE_OLDJIS.
        (ENCODE_ISO_CHARACTER_DIMENSION1): Likewise.
        (ENCODE_ISO_CHARACTER_DIMENSION2): Likewise.
        (encode_coding_iso_2022): Change the 1st arg to
        ENCODE_ISO_CHARACTER to a variable.

2002-05-29  Kenichi Handa  <handa@m17n.org>

        * charset.h (enum define_charset_arg_index): New enums
        charset_arg_min_code and charset_arg_max_code.
        (struct charset): New member char_index_offset.

        * charset.c (CODE_POINT_TO_INDEX): Take charset->char_index_offset
        into account.
        (INDEX_TO_CODE_POINT): Likewise.
        (Fdefine_charset_internal): Handle args[charset_arg_min_code] and
        args[charset_arg_max_code].  Setup charset.char_index_offset.
        (syms_of_charset): Fix args to Fdefine_charset_internal.

2002-05-27  Dave Love  <fx@gnu.org>

        * coding.c (decode_coding_utf_8): Reject overlong sequences.

2002-05-26  Dave Love  <fx@gnu.org>

        * coding.c: Doc fixes.
        (Fcoding_system_aliases): Fix return value.
        (Qmac): Remove (duplicated) definition.

2002-05-25  Dave Love  <fx@gnu.org>

        * charset.c (Fcharset_priority_list, Fset_charset_priority): New
        functions.

        * character.c (Fstring): Doc fix.

        * charset.c (Fdefine_charset_alias): Update Vcharset_list.

        * fontset.c (Ffontset_info): Doc fix.  Return charset names, not
        ids.
        (font-encoding-alist): Doc fix.

2002-05-24  Dave Love  <fx@gnu.org>

        * term.c (costs_set): Declare static, non-initialized for pcc.
        (encode_terminal_code): Remove ensued var.

        * keyboard.c (kbd_buffer_store_event): Fix interrupt_signal decl
        for K&R.

        * xterm.c (xlwmenu_window_p): Fix prototype for K&R.

        * coding.c (setup_iso_safe_charsets): Fix arg decl for K&R.
        (suffixes): Moved out of make_subsidiaries for K&R.

        * charset.c (map_charset_chars): Fix c_function declaration for
        K&R.

        * lisp.h (DEFUN) [!PROTOTYPES]: Remove spurious `args'.

2002-05-23  Dave Love  <fx@gnu.org>

        * data.c (Fchar_or_string_p): Doc fix.  Use CHARACTERP.

        * category.c (Fmodify_category_entry): Doc fix.  Remove unused
        vars.

2002-05-23  Yong Lu <lyongu@asia-infonet.com>

        * charset.c (Fdefine_charset_internal): Fix argument to bzero.

        * coding.c (Fdefine_coding_system_internal): Fix previous change.
        (decode_coding_charset): Workaround for the bug of GCC 2.96.

2002-05-23  Kenichi Handa  <handa@m17n.org>

        * Makefile.in (lisp): Change cyrillic.elc to cyrillic.el,
        vietnamese.elc to vietnamese.el.  They are not bytecompiled now.
        (shortlisp): Likewise.

2002-05-22  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding_charset): Adjusted for the change of
        Fdefine_coding_system_internal.
        (Fdefine_coding_system_internal): For a coding system of
        `charset' type, store a list of charset IDs in
        `charset_attr_charset_valids' element of coding attributes.

        * charset.c (Fmake_char): Fix previous change.

2002-05-21  Kenichi Handa  <handa@m17n.org>

        * coding.c (ONE_MORE_BYTE_NO_CHECK): Increment consumed_chars.
        (emacs_mule_char): New arg src.  Delete arg `composition'.  Caller
        changed.  Handle 2-byte and 3-byte charsets correctly.
        (DECODE_EMACS_MULE_COMPOSITION_RULE_20): Renamed from
        DECODE_EMACS_MULE_COMPOSITION_RULE.  Caller changed.
        (DECODE_EMACS_MULE_COMPOSITION_RULE_21): New macro.
        (DECODE_EMACS_MULE_21_COMPOSITION): Call
        DECODE_EMACS_MULE_COMPOSITION_RULE_21.  Produce correct annotation
        sequence.
        (decode_coding_emacs_mule): Handle composition correctly.  Rewind
        `src' and `consumed_chars' correctly before calling
        emacs_mule_char.
        (DECODE_COMPOSITION_START): Correctly handle the case of altchar
        and alt&rule composition.
        (decode_coding_iso_2022): Handle composition correctly.
        (init_coding_once): Setup emacs_mule_bytes for private charsets.
        
        * charset.c (Fdefine_charset_internal): Fix bug for the case of
        re-defining a charset.  If the charset has :emacs-mule-id, setup
        emacs_mule_bytes.
        (Fmake_char): If CODE1 is nil, use the minimum code of the
        charset.

2002-05-20  Kenichi Handa  <handa@m17n.org>

        * coding.c (encode_coding_iso_2022): If coding requires safe
        encoding, produce a character specified by
        CODING_INHIBIT_CHARACTER_SUBSTITUTION.
        (encode_coding_sjis): Likewise.
        (encode_coding_big5): Likewise.
        (encode_coding_charset): Likewise.

2002-05-17  Dave Love  <fx@gnu.org>

        * xterm.c (XSetIMValues): Declare.

        * process.c: Conditionally include sys/wait.h, pty.h.

        * print.c (print_object): Fix print format for 64-bit
        systems.

        * keyboard.c (modify_event_symbol): Fix print format for 64-bit
        systems.

        * buffer.c (emacs_strerror): Declare.
        (MMAP_ALLOCATED_P, mmap_enlarge, syms_of_buffer): Import changes
        from trunk.

        * fontset.c (Fclear_face_cache): Declare.
        (accumulate_font_info): Commented-out (unused).
        (face_for_char, Fset_fontset_font, Ffontset_info): Remove unused
        variables.

        * character.h (string_escape_byte8): Declare.

        * charset.c (load_charset_map, load_charset_map_from_file): Remove
        unused vars.
        (Fdefine_charset_internal, Fsplit_char, syms_of_charset)
        (Fmap_charset_chars): Doc fix.

        * coding.c (Vchar_coding_system_table, Qchar_coding_system):
        Removed.
        (Fset_coding_system_priority, Fset_coding_system_priority) 
        (Fdefine_coding_system_internal): Doc fix.

2002-05-16  Dave Love  <fx@gnu.org>

        * s/osf5-0.h (C_SWITCH_SYSTEM) [!__GNUC__]: Remove -nointrinsics.

2002-05-16  Kenichi Handa  <handa@m17n.org>

        * character.c (string_escape_byte8): Make multibyte string with
        correct size.

        * charset.c (Fmake_char): Delete unnecessary code.

2002-05-14  Kenichi Handa  <handa@m17n.org>

        * xfns.c (x_encode_text): Allocate coding.destination here, and
        call encode_coding_object with dst_object Qnil.

        * buffer.c (Fset_buffer_multibyte): Convert 8-bit bytes to
        multibyte form correctly.

        * fontset.c (fs_load_font): Check fontp->full_name (not fontname)
        against Vfont_encoding_alist.

        * coding.c (Fdecode_sjis_char): Fix typo (0x7F->0xFF).  Fix the
        handling of charset list.
        (encode_coding_iso_2022): Setup coding->safe_charsets in advance.
        (decode_coding_object): Move point to coding->dst_pos before
        calling post-read-conversion function.
        (encode_coding_object): Give correct arguments to
        pre-write-conversion.  Ignore the return value of
        pre-write-conversion function.  Pay attention to the case that
        pre-write-conversion changes the current buffer.  If dst_object is
        Qt, even if coding->src_bytes is zero, allocate at least one byte
        to coding->destination.

        * coding.h (JIS_TO_SJIS): Fix typo (j1->s1, j2->s2).

        * charset.c (Fmake_char): Make it more backward compatible.
        (Fmap_charset_chars): Fix docstring.

2002-05-13  Dave Love  <fx@gnu.org>

        * coding.c: Doc fixes.
        (Fdefine_coding_system_alias): Use names, not symbols, in
        coding-system-alist.

2002-05-13  Kenichi Handa  <handa@m17n.org>

        * fontset.c (free_realized_fontsets): Call Fclear_face_cache instead
        of calling free_realized_face.

2002-05-10  Yong Lu <lyongu@asia-infonet.com>

        * charset.c (load_charset_map): Fix previous change.
        (read_hex): Don't treat SPC as a comment starter.
        (decode_char): If CODE_POINT_TO_INDEX retruns -1, always return
        -1.
        (Fdecode_char): Fix typo.

2002-05-10  Kenichi Handa  <handa@m17n.org>

        * charset.h (struct charset): New member `code_space_mask'.

        * coding.c (coding_set_source): Delete the local variable
        beg_byte.
        (encode_coding_charset): Delete the local variable charset.
        (Fdefine_coding_system_internal): Likewise.
        (Fdefine_coding_system_internal): Setup
        attrs[coding_attr_charset_valids] correctly.

        * charset.c (CODE_POINT_TO_INDEX): Utilize `code_space_mask'
        member to check if CODE is valid or not.
        (Fdefine_charset_internal): Initialize `code_space_mask' member.
        (encode_char): Before calling CODE_POINT_TO_INDEX, check if CODE
        is within the range of charset->min_code and carset->max_code.

2002-05-09  Dave Love  <fx@gnu.org>

        * syntax.h (syntax_temp) [!__GNUC__]: Declare.

        * dispextern.h (generate_ascii_font): Fix return type.

        * xfaces.c (generate_ascii_font): Fix arg declaration.

        * coding.c (coding_inherit_eol_type)
        (Fset_terminal_coding_system_internal)
        (Fset_safe_terminal_coding_system_internal): Fix arg declarations.

2002-05-08  Kenichi Handa  <handa@m17n.org>

        * coding.c (decode_coding_charset, encode_coding_charset): Handle
        multiple charsets correctly.

2002-05-07  Kenichi Handa  <handa@m17n.org>

        * search.c (boyer_moore): Fix handling of mulitbyte character
        translation.

        * xdisp.c (display_mode_element): When the variable `elt' is
        changed, update `this' and `lisp_string'.

2002-05-07  Kenichi Handa  <handa@m17n.org>

        * buffer.c (Fset_buffer_multibyte): Fix 8-bit char handling.

        * callproc.c (Fcall_process): Be sure to give the current buffer
        to decode_coding_c_string.  Update PT and PT_BYTE after the
        insertion.

        * charset.c (struct charset_map_entries): New struct.
        (load_charset_map): Renamed from parse_charset_map.  New args
        entries and n_entries.  Caller changed.
        (load_charset_map_from_file): Renamed from load_charset_map.
        Caller changed.  New arg control_flag.  Call load_charset_map at
        the tail.
        (load_charset_map_from_vector): New function.
        (Fdefine_charset_internal): Setup charset.compact_codes_p.
        (encode_char): If the charset is compact, change a character index
        to a code point.

        * coding.c (coding_alloc_by_making_gap): Check the case that the
        source and destination are the same correctly.
        (decode_coding_raw_text): Set coding->consumed_char and
        coding->consumed to 0.
        (produce_chars): If coding->chars_at_source is nonzero, update
        coding->consumed_char and coding->consumed before calling
        alloc_destination.
        (Fdefine_coding_system_alias): Register ALIAS in
        Vcoding_system_alist.
        (syms_of_coding): Define `no-convesion' coding system at the tail.

        * fileio.c (Finsert_file_contents): Set coding_system instead of
        val.  If the current buffer is multibyte, always call
        decode_coding_gap.

        * xfaces.c (try_font_list): Give higher priority to fontset's
        family than face's family.

2002-04-18  Kenichi Handa  <handa@m17n.org>

        * callproc.c (Fcall_process): Be sure to give the current buffer
        to decode_coding_c_string.

        * xfaces.c (try_font_list): Give a family specified in a fontset
        higher priority than a family specified in a face.

2002-04-09  Kenichi Handa  <handa@m17n.org>

        * fileio.c (Finsert_file_contents): Fix calculation of `inserted'.
        Fix arguments to insert_from_buffer.

        * xdisp.c (display_mode_element): Fix calculation of `bytepos'.

2002-03-11  Kenichi Handa  <handa@m17n.org>

        * coding.c (produce_chars): Set the variable `multibytep' correctly.
        (decode_coding_gap): Set coding->dst_multibyte correctly.

2002-03-07  Kenichi Handa  <handa@m17n.org>

        * coding.c (encode_coding_utf_8): Initialize produced_chars to 0.
        (decode_coding_utf_16): Fix converting high and low bytes to
        code-point.
        (encode_coding_utf_16): Substitute coding->default_char for
        non-Unicode characters.
        (decode_coding): Don't call record_insert here.
        (setup_coding_system): Initialize `surrogate' of
        coding->spec.utf_16 to 0.
        (EMIT_ONE_BYTE): Fix for multibyte case.

        * insdel.c (insert_from_gap): Call record_insert.

2002-03-04  Kenichi Handa  <handa@m17n.org>

        * casefiddle.c (casify_region): Fix multibyte case.

        * character.c (c_string_width): Add return type `int'.
        (char_string_with_unification): Arg ADVANCED deleted.

        * character.h (CHAR_VALID_P): Don't call CHARACTERP.
        (CHAR_STRING): Adjusted for the change of
        char_string_with_unification.
        (CHAR_STRING_ADVANCE): Make it do-while statement.

        * chartab.c (sub_char_table_set_range): Optimized for the case
        DEPTH == 3.  Add workaround code for a GCC optimization bug.

        * charset.c (parse_charset_map): Remove an unused variable.

        * coding.c: Delete unused variables.

        * fileio.c (Finsert_file_contents): Set coding_system to Qnil
        earlier.  If inserted is zero and the coding system doesn't
        require flushing, don't call decode_coding_gap.

        * syntax.h (SET_RAW_SYNTAX_ENTRY): Don't call make_number.

2002-03-01  Kenichi Handa  <handa@m17n.org>

        The following changes are for using Unicode as an internal
        character model, and use UTF-8 format for buffer/string
        representation.

        * .gdbinit (xchartable): Adjusted for the change of char table
        structure.
        (xsubchartable, xcoding, xcharset, xcurbuf): New commands.
        
        * Makefile.in (obj): Add character.o and chartab.o.
        (lisp, shortlisp): Remove utf-8.elc:
        (*.o): For many files, change dependency on charset.h to
        character.h, and add dependency on character.h.
        (character.o, chartab.o): New targets.
        
        * abbrev.c, bytecode.c, casefiddle.c, cmds.c, dispnew.c, doc.c,
        doprnt.c, dosfns.c, frame.c, marker.c, minibuf.c, msdos.c,
        w16select.c, w32bdf.c, w32console.c: Include "character.h" instead
        of "charset.h".
        
        * dired.c, filelock.c: Include "character.h".

        * alloc.c: Include "character.h" instead of "charset.h".
        (Fmake_char_table): Moved to chartab.c.
        (make_sub_char_table): Likewise.
        (syms_of_alloc): Remove defsubr for Smake_char_table.

        * buffer.c: Include "character.h" instead of "charset.h", don't
        include "coding.h".
        (Fset_buffer_multibyte): Adjuted for UTF-8.

        * buffer.h: EXFUN Fbuffer_live_p.

        * callproc.c: Include "character.h" instead of "charset.h".
        (Fcall_process): Big change for the new code-conversion APIs.

        * casetab.c: Include "character.h" instead of "charset.h".
        (set_canon, set_identity, shuffle): Adjusted for the new
        map_char_table spec.
        (init_casetab_once): Call CHAR_TABLE_SET instead of directly
        accessing the char table structure.

        * chartab.c: New file that implements char table.

        * category.c: Include "character.h".
        (copy_category_entry): New function.
        (copy_category_table): Call map_char_table and copy_category_entry.
        (Fmake_category_table): Initialize all top-vel slots.
        (char_category_set): New function.
        (modify_lower_category_set): Deleted.
        (Fmodify_category_entry): Call char_table_ref_and_range.

        * category.h (CATEGORY_SET): Just call char_category_set.

        * ccl.c: Include "character.h".
        (Qccl, Qcclp): New variables.
        (CCL_WRITE_CHAR): Alway treat the arg CH as a character even if
        it's less than 256.
        (CCL_WRITE_MULTIBYTE_CHAR): Deleted.
        (CCL_WRITE_STRING, CCL_READ_CHAR): Adjusted for the change of SRC
        and DST type.
        (ccl_driver): Types of arguments changed.  Code adjusted for that.
        (Fccl_execute, Fccl_execute_on_string): Adjusted for the change of
        ccl_driver.
        (syms_of_ccl): Intern and staticpro Qccl and Qcclp.

        * ccl.h (struct ccl_program): Members eol_type and multibyte
        deleted.  New members src_multibyte, dst_multibyte, consumed, and
        produced.
        (struct ccl_spec): Members decoder and encoder deleted.  New
        memeber ccl.
        (CODING_SPEC_CCL_PROGRAM): New macro.
        (ccl_driver): Prototype updated.
        (Qccl, Qcclp, Fccl_program_p): Extern them.
        (CHECK_CCL_PROGRAM): New macro.

        * character.c, character.h, chartab.c: New files. 

        * charset.c: Mostly re-written.  Character and multibyte sequence
        handling codes are moved to character.c.

        * charset.h: Mostly re-written.  Character and multibyte sequence
        handling codes are moved to character.h.

        * coding.c, coding.h: Mostly re-written.

        * composite.c: Include "character.h" instead of "charset.h".
        (CHAR_WIDTH): Moved to character.h.
        (HASH_KEY, HASH_VALUE): Deleted.

        * composite.h (enum composition_method): Order of enumeration
        symbols changed.

        * data.c: Include "character.h" instead of "charset.h".
        (Faref): Call CHAR_TABLE_REF for a char table.
        (Faset): Call CHAR_TABLE_SET for a char table.

        * dispextern.h (free_realized_face, check_face_attribytes,
        generate_ascii_font): Extern them.
        (free_realized_multibyte_face): Extern deleted.

        * disptab.h (DISP_CHAR_VECTOR): Adjusted for the change of char
        table structure.
        
        * editfns.c: Include "character.h" instead of "charset.h".
        (Fchar_to_string): Always call CHAR_STRING.

        * emacs.c (main): Call init_charset_once, init_charset,
        syms_of_chartab, and syms_of_character.

        * fileio.c: Include "character.h" instead of "charset.h".
        (Finsert_file_contents): Big change for the new code-conversion
        API.
        (choose_write_coding_system): Likewise.
        (Fwrite_region): Likewise.
        (build_annotations_2): Deleted.
        (e_write): Big change for the new code-conversion API.

        * fns.c: Include "character.h" instead of "charset.h".
        (copy_sub_char_table): Moved to chartab.c.
        (Fcopy_sequence): Call copy_char_table for a char table.
        (concat): Delete codes calling count_multibyte.
        (string_char_to_byte): Adjusted for the new multibyte form.
        (string_byte_to_char): Likewise.
        (internal_equal): Adjusted for the change of char table structure.
        (Fchar_table_subtype, Fchar_table_parent, Fset_char_table_parent,
        Fchar_table_extra_slot, Fset_char_table_extra_slot,
        Fchar_table_range, Fset_char_table_range, Fset_char_table_default,
        char_table_translate, optimize_sub_char_table,
        Foptimize_char_table, map_char_table, Fmap_char_table): Moved to
        chartab.c.
        (char_table_ref_and_index): Deleted.
        (HASH_KEY, HASH_VALUE): Moved to lisp.h.
        (Fmd5): Call preferred_coding_system instead of accessing
        Vcoding_category_list.  Adjusted for the new code-conversion API.
        (syms_of_fns): Defsubr for char table related functions moved to
        chartab.c.

        * fontset.c: Mostly re-written.

        * fontset.h (struct font_info): Type of the member encoding_type
        changed.
        (enum FONT_SPEC_INDEX): New enum.
        (fontset_font_pattern, fs_load_font): Prototype updated.
        (FS_LOAD_FONT): Adjusted for the change of fs_load_font.

        * indent.c: Include "character.h" instead of "charset.h".
        (MULTIBYTE_BYTES_WIDTH): Call CHAR_WIDTH instead of
        WIDTH_BY_CHAR_HEAD.

        * insdel.c: Include "character.h" instead of "charset.h".
        (copy_text): Don't refer to Vnonascii_translation_table.
        (insert_from_gap): New function.

        * keyboard.c: Include "character.h" instead of "charset.h".
        (command_loop_1): Never call direct_output_forward_char before
        a non-ASCII character.
        (read_char): If Vkeyboard_translate_table is a char table, always
        translated a character.

        * keymap.c: Include "character.h".
        (store_in_keymap): Handle the case that IDX is a cons.
        (Fdefine_key): Handle the case that KEY is a cons and the car part
        is also a cons (range).
        (push_key_description): Adjusted for the new character code.
        (describe_vector): Call describe_char_table for a char table.
        (describe_char_table): New function.

        * keymap.h (describe_char_table): Extern it.

        * lisp.h (enum pvec_type): New member PVEC_SUB_CHAR_TABLE.
        (XSUB_CHAR_TABLE, XSETSUB_CHAR_TABLE): New macros.
        (CHAR_TABLE_ORDINARY_SLOTS, CHAR_TABLE_SINGLE_BYTE_SLOTS,
        SUB_CHAR_TABLE_ORDINARY_SLOTS, SUB_CHAR_TABLE_STANDARD_SLOTS):
        Deleted.
        (CHAR_TABLE_REF, CHAR_TABLE_SET): Adjusted for the new char table
        structure.
        (CHAR_TABLE_TRANSLATE): Just call char_table_translate.
        (CHARTAB_SIZE_BITS_0, CHARTAB_SIZE_BITS_1, CHARTAB_SIZE_BITS_2,
        CHARTAB_SIZE_BITS_3): New macros.
        (chartab_size): Extern it.
        (struct Lisp_Char_Table): Re-designed.
        (struct Lisp_Sub_Char_Table): New structure.
        (HASH_KEY, HASH_VALUE): Moved from fns.c.
        (CHARACTERBITS): Defined as 22.
        (GLYPH_MASK_FACE, GLYPH_MASK_CHAR): Adjusted for the above change.
        (SUB_CHAR_TABLE_P): Check PVEC_CHAR_TABLE.
        (GC_SUB_CHAR_TABLE_P): New macro.
        (Fencode_coding_string, Fdecode_coding_string): EXFUN Updated.
        (code_convert_string_norecord): Extern deleted.
        (init_character_once, syms_of_character, init_charset,
        syms_of_composite, Qeq, Fmakehash, insert_from_gap): Extern them.

        * lread.c: Include "character.h".
        (read_multibyte): New arg NBYTES.
        (read_escape): The meaning of returned *BYTEREP changed.
        (to_multibyte): Deleted.
        (read1): Adjuted the handling of char table and string.

        * print.c: Include "character.h" instead of "charset.h".
        (print_string): Convert 8-bit raw bytes to octal form by
        string_escape_byte8.
        (print_object): Adjusted for the new multibyte form.  Print 8-bit
        raw bytes always in octal form.  Handle sub char table correctly.

        * process.c: Include "character.h" instead of "charset.h".
        (read_process_output): Adjusted for the new code-conversion API.
        (send_process): Likewise.

        * puresize.h (BASE_PURESIZE): Increased.

        * regex.c: Include "character.h" instead of "charset.h".
        (BYTE8_TO_CHAR, CHAR_BYTE8_P) [not emacs]: New dummy macros.
        (regex_compile): Accept a range whose starting and ending
        character have different leading bytes.
        (analyse_first): Adjusted for the above change.

        * search.c: Include "character.h" instead of "charset.h".
        (search_buffer, boyer_moore): Adjusted for the new multibyte form.
        (Freplace_match): Adjusted for the change of
        multibyte_char_to_unibyte.

        * syntax.c: Include "character.h" instead of "charset.h".
        (syntax_parent_lookup): Deleted.
        (Fmodify_syntax_entry): Accept a cons as CHAR.
        (skip_chars): Adjusted for the new multibyte form.
        (init_syntax_once): Call char_table_set_range instead of directly
        accessing the structure of a char table.

        * syntax.h (SET_RAW_SYNTAX_ENTRY): Call CHAR_TABLE_SET.
        (SYNTAX_ENTRY_FOLLOW_PARENT): Macro deleted.
        (SET_RAW_SYNTAX_ENTRY_RANGE): New macro.
        (SYNTAX_ENTRY_INT): Call CHAR_TABLE_REF.

        * term.c: Include "buffer.h" and "character.h".
        (encode_terminal_code): Adjusted for the new code-conversion API.
        (write_glyphs): Likewise.
        (produce_glyphs): Call CHAR_WIDTH instead of CHARSET_WIDTH.

        * w32term.c (x_new_font): Adjusted for the change of FS_LOAD_FONT.

        * xdisp.c: Include "character.h".
        (get_next_display_element): Adjusted for the new multibyte form.
        (disp_char_vector): Adjusted for the new char table structure.
        (decode_mode_spec_coding): Adjusted for the new structure of
        coding system.
        (decode_mode_spec): Adjusted for the new code-conversion API.

        * xfaces.c: Include "character.h" instead of "charset.h".
        (load_face_font): Adjusted for the change of choose_face_font and
        FS_LOAD_FONT.
        (generate_ascii_font): New function.
        (set_lface_from_font_name): Adjusted for the change of
        FS_LOAD_FONT.
        (set_font_frame_param): Adjusted for the change of
        choose_face_font.
        (free_realized_face): Make it public.
        (free_realized_faces_for_fontset): Renamed from
        free_realized_multibyte_face.  Free also faces realized for ASCII.
        (choose_face_font): Argments changed.  Adjusted for the change of
        fontset_font_pattern and FS_LOAD_FONT.

        * xfns.c: Include "character.h".
        (x_encode_text): Adjusted for the new code-conversion API.

        * xselect.c: Don't include "charset.h".
        (selection_data_to_lisp_data): Adjusted for the new code
        covnersion API.

        * xterm.c: Include "character.h".
        (x_encode_char): New argument CHARSET.  Caller changed.
        (x_get_char_face_and_encoding): Call ENCODE_CHAR instead of
        SPLIT_CHAR.
        (x_get_glyph_face_and_encoding): Likewise.
        (x_produce_glyphs): Don't check Vnonascii_translation_table Call
        CHAR_WIDTH instead of CHARSET_WIDTH.
        (XTread_socket): Adjusted for the new code-conversion API.
        (x_new_font): Adjusted for the change of FS_LOAD_FONT.
        (x_load_font): Adjusted for the change of struct font.

;; Local Variables:
;; coding: iso-2022-7bit
;; End:

    Copyright (C) 2002 Free Software Foundation, Inc.
  Copying and distribution of this file, with or without modification,
  are permitted provided the copyright notice and this notice are preserved.

;;; arch-tag: 1bff38bd-2030-46ae-9d18-f15e6006b665