From f14db65234364c75d1efbad1f1a7d1e32ed4686a Mon Sep 17 00:00:00 2001 From: Tim Waugh Date: May 15 2006 14:10:12 +0000 Subject: Fixed i18n patch. --- diff --git a/coreutils-i18n.patch b/coreutils-i18n.patch index 481f8bb..d841385 100644 --- a/coreutils-i18n.patch +++ b/coreutils-i18n.patch @@ -1,49 +1,23 @@ ---- coreutils-5.93/lib/linebuffer.h.i18n 2005-05-14 08:58:06.000000000 +0100 -+++ coreutils-5.93/lib/linebuffer.h 2005-12-23 08:53:01.000000000 +0000 -@@ -22,6 +22,11 @@ - - # include - -+/* Get mbstate_t. */ -+# if HAVE_WCHAR_H -+# include -+# endif -+ - /* A `struct linebuffer' holds a line of text. */ - - struct linebuffer -@@ -29,6 +34,9 @@ - size_t size; /* Allocated. */ - size_t length; /* Used. */ - char *buffer; -+# if HAVE_WCHAR_H -+ mbstate_t state; -+# endif - }; - - /* Initialize linebuffer LINEBUFFER for use. */ ---- coreutils-5.93/src/cut.c.i18n 2005-08-12 08:16:25.000000000 +0100 -+++ coreutils-5.93/src/cut.c 2005-12-23 08:53:01.000000000 +0000 -@@ -29,6 +29,11 @@ - #include +--- coreutils-5.95/src/expand.c.i18n 2005-08-12 08:16:25.000000000 +0100 ++++ coreutils-5.95/src/expand.c 2006-05-15 15:08:57.000000000 +0100 +@@ -38,11 +38,28 @@ + #include #include #include + -+/* Get mbstate_t, mbrtowc(). */ ++/* Get mbstate_t, mbrtowc(), wcwidth(). */ +#if HAVE_WCHAR_H +# include +#endif ++ #include "system.h" - #include "error.h" -@@ -37,6 +42,18 @@ #include "quote.h" #include "xstrndup.h" +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC -+ installation; work around this configuration error. */ ++ installation; work around this configuration error. */ +#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 -+# undef MB_LEN_MAX +# define MB_LEN_MAX 16 +#endif + @@ -53,570 +27,175 @@ +#endif + /* The official name of this program (e.g., no `g' prefix). */ - #define PROGRAM_NAME "cut" - -@@ -67,6 +84,52 @@ - } \ - while (0) - -+/* Refill the buffer BUF to get a multibyte character. */ -+#define REFILL_BUFFER(BUF, BUFPOS, BUFLEN, STREAM) \ -+ do \ -+ { \ -+ if (BUFLEN < MB_LEN_MAX && !feof (STREAM) && !ferror (STREAM)) \ -+ { \ -+ memmove (BUF, BUFPOS, BUFLEN); \ -+ BUFLEN += fread (BUF + BUFLEN, sizeof(char), BUFSIZ, STREAM); \ -+ BUFPOS = BUF; \ -+ } \ -+ } \ -+ while (0) -+ -+/* Get wide character on BUFPOS. BUFPOS is not included after that. -+ If byte sequence is not valid as a character, CONVFAIL is 1. Otherwise 0. */ -+#define GET_NEXT_WC_FROM_BUFFER(WC, BUFPOS, BUFLEN, MBLENGTH, STATE, CONVFAIL) \ -+ do \ -+ { \ -+ mbstate_t state_bak; \ -+ \ -+ if (BUFLEN < 1) \ -+ { \ -+ WC = WEOF; \ -+ break; \ -+ } \ -+ \ -+ /* Get a wide character. */ \ -+ CONVFAIL = 0; \ -+ state_bak = STATE; \ -+ MBLENGTH = mbrtowc ((wchar_t *)&WC, BUFPOS, BUFLEN, &STATE); \ -+ \ -+ switch (MBLENGTH) \ -+ { \ -+ case (size_t)-1: \ -+ case (size_t)-2: \ -+ CONVFAIL++; \ -+ STATE = state_bak; \ -+ /* Fall througn. */ \ -+ \ -+ case 0: \ -+ MBLENGTH = 1; \ -+ break; \ -+ } \ -+ } \ -+ while (0) -+ - struct range_pair - { - size_t lo; -@@ -85,7 +148,7 @@ - /* The number of bytes allocated for FIELD_1_BUFFER. */ - static size_t field_1_bufsize; - --/* The largest field or byte index used as an endpoint of a closed -+/* The largest byte, character or field index used as an endpoint of a closed - or degenerate range specification; this doesn't include the starting - index of right-open-ended ranges. For example, with either range spec - `2-5,9-', `2-3,5,9-' this variable would be set to 5. */ -@@ -97,10 +160,11 @@ - - /* This is a bit vector. - In byte mode, which bytes to output. -+ In character mode, which characters to output. - In field mode, which DELIM-separated fields to output. -- Both bytes and fields are numbered starting with 1, -+ Bytes, characters and fields are numbered starting with 1, - so the zeroth bit of this array is unused. -- A field or byte K has been selected if -+ A byte, character or field K has been selected if - (K <= MAX_RANGE_ENDPOINT and is_printable_field(K)) - || (EOL_RANGE_START > 0 && K >= EOL_RANGE_START). */ - static unsigned char *printable_field; -@@ -109,9 +173,12 @@ - { - undefined_mode, - -- /* Output characters that are in the given bytes. */ -+ /* Output bytes that are at the given positions. */ - byte_mode, - -+ /* Output characters that are at the given positions. */ -+ character_mode, -+ - /* Output the given delimeter-separated fields. */ - field_mode - }; -@@ -121,6 +188,13 @@ - - static enum operating_mode operating_mode; - -+/* If nonzero, when in byte mode, don't split multibyte characters. */ -+static int byte_mode_character_aware; -+ -+/* If nonzero, the function for single byte locale is work -+ if this program runs on multibyte locale. */ -+static int force_singlebyte_mode; -+ - /* If true do not output lines containing no delimeter characters. - Otherwise, all such lines are printed. This option is valid only - with field mode. */ -@@ -132,6 +206,9 @@ - - /* The delimeter character for field mode. */ - static unsigned char delim; -+#if HAVE_WCHAR_H -+static wchar_t wcdelim; -+#endif - - /* True if the --output-delimiter=STRING option was specified. */ - static bool output_delimiter_specified; -@@ -205,7 +282,7 @@ - -f, --fields=LIST select only these fields; also print any line\n\ - that contains no delimiter character, unless\n\ - the -s option is specified\n\ -- -n (ignored)\n\ -+ -n with -b: don't split multibyte characters\n\ - "), stdout); - fputs (_("\ - --complement complement the set of selected bytes, characters\n\ -@@ -360,7 +437,7 @@ - in_digits = false; - /* Starting a range. */ - if (dash_found) -- FATAL_ERROR (_("invalid byte or field list")); -+ FATAL_ERROR (_("invalid byte, character or field list")); - dash_found = true; - fieldstr++; - -@@ -385,14 +462,16 @@ - if (value == 0) - { - /* `n-'. From `initial' to end of line. */ -- eol_range_start = initial; -+ if (eol_range_start == 0 || -+ (eol_range_start != 0 && eol_range_start > initial)) -+ eol_range_start = initial; - field_found = true; - } - else - { - /* `m-n' or `-n' (1-n). */ - if (value < initial) -- FATAL_ERROR (_("invalid byte or field list")); -+ FATAL_ERROR (_("invalid byte, character or field list")); + #define PROGRAM_NAME "expand" - /* Is there already a range going to end of line? */ - if (eol_range_start != 0) -@@ -465,6 +544,9 @@ - if (operating_mode == byte_mode) - error (0, 0, - _("byte offset %s is too large"), quote (bad_num)); -+ else if (operating_mode == character_mode) -+ error (0, 0, -+ _("character offset %s is too large"), quote (bad_num)); - else - error (0, 0, - _("field number %s is too large"), quote (bad_num)); -@@ -475,7 +557,7 @@ - fieldstr++; +@@ -182,6 +199,7 @@ + stops = num_start + len - 1; + } } ++ else -- FATAL_ERROR (_("invalid byte or field list")); -+ FATAL_ERROR (_("invalid byte, character or field list")); - } - - max_range_endpoint = 0; -@@ -568,6 +650,63 @@ + { + error (0, 0, _("tab size contains invalid character(s): %s"), +@@ -364,6 +382,142 @@ } } +#if HAVE_MBRTOWC -+/* This function is in use for the following case. -+ -+ 1. Read from the stream STREAM, printing to standard output any selected -+ characters. -+ -+ 2. Read from stream STREAM, printing to standard output any selected bytes, -+ without splitting multibyte characters. */ -+ +static void -+cut_characters_or_cut_bytes_no_split (FILE *stream) ++expand_multibyte (void) +{ -+ int idx; /* number of bytes or characters in the line so far. */ ++ FILE *fp; /* Input strem. */ ++ mbstate_t i_state; /* Current shift state of the input stream. */ ++ mbstate_t i_state_bak; /* Back up the I_STATE. */ ++ mbstate_t o_state; /* Current shift state of the output stream. */ + char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ -+ char *bufpos; /* Next read position of BUF. */ -+ size_t buflen; /* The length of the byte sequence in buf. */ -+ wint_t wc; /* A gotten wide character. */ -+ size_t mblength; /* The byte size of a multibyte character which shows -+ as same character as WC. */ -+ mbstate_t state; /* State of the stream. */ -+ int convfail; /* 1, when conversion is failed. Otherwise 0. */ ++ char *bufpos; /* Next read position of BUF. */ ++ size_t buflen = 0; /* The length of the byte sequence in buf. */ ++ wchar_t wc; /* A gotten wide character. */ ++ size_t mblength; /* The byte size of a multibyte character ++ which shows as same character as WC. */ ++ int tab_index = 0; /* Index in `tab_list' of next tabstop. */ ++ int column = 0; /* Column on screen of the next char. */ ++ int next_tab_column; /* Column the next tab stop is on. */ ++ int convert = 1; /* If nonzero, perform translations. */ + -+ idx = 0; -+ buflen = 0; -+ bufpos = buf; -+ memset (&state, '\0', sizeof(mbstate_t)); ++ fp = next_file ((FILE *) NULL); ++ if (fp == NULL) ++ return; + -+ while (1) -+ { -+ REFILL_BUFFER (buf, bufpos, buflen, stream); ++ memset (&o_state, '\0', sizeof(mbstate_t)); ++ memset (&i_state, '\0', sizeof(mbstate_t)); + -+ GET_NEXT_WC_FROM_BUFFER (wc, bufpos, buflen, mblength, state, convfail); -+ -+ if (wc == WEOF) -+ { -+ if (idx > 0) -+ putchar ('\n'); -+ break; -+ } -+ else if (wc == L'\n') -+ { -+ putchar ('\n'); -+ idx = 0; -+ } -+ else ++ for (;;) ++ { ++ /* Refill the buffer BUF. */ ++ if (buflen < MB_LEN_MAX && !feof(fp) && !ferror(fp)) + { -+ idx += (operating_mode == byte_mode) ? mblength : 1; -+ if (print_kth (idx, NULL)) -+ fwrite (bufpos, mblength, sizeof(char), stdout); ++ memmove (buf, bufpos, buflen); ++ buflen += fread (buf + buflen, sizeof(char), BUFSIZ, fp); ++ bufpos = buf; + } + -+ buflen -= mblength; -+ bufpos += mblength; -+ } -+} -+#endif -+ - /* Read from stream STREAM, printing to standard output any selected fields. */ - - static void -@@ -689,13 +828,192 @@ - } - } - -+#if HAVE_MBRTOWC -+static void -+cut_fields_mb (FILE *stream) -+{ -+ int c; -+ unsigned int field_idx; -+ int found_any_selected_field; -+ int buffer_first_field; -+ int empty_input; -+ char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ -+ char *bufpos; /* Next read position of BUF. */ -+ size_t buflen; /* The length of the byte sequence in buf. */ -+ wint_t wc = 0; /* A gotten wide character. */ -+ size_t mblength; /* The byte size of a multibyte character which shows -+ as same character as WC. */ -+ mbstate_t state; /* State of the stream. */ -+ int convfail; /* 1, when conversion is failed. Otherwise 0. */ -+ -+ found_any_selected_field = 0; -+ field_idx = 1; -+ bufpos = buf; -+ buflen = 0; -+ memset (&state, '\0', sizeof(mbstate_t)); -+ -+ c = getc (stream); -+ empty_input = (c == EOF); -+ if (c != EOF) -+ ungetc (c, stream); -+ else -+ wc = WEOF; -+ -+ /* To support the semantics of the -s flag, we may have to buffer -+ all of the first field to determine whether it is `delimited.' -+ But that is unnecessary if all non-delimited lines must be printed -+ and the first field has been selected, or if non-delimited lines -+ must be suppressed and the first field has *not* been selected. -+ That is because a non-delimited line has exactly one field. */ -+ buffer_first_field = (suppress_non_delimited ^ !print_kth (1, NULL)); -+ -+ while (1) -+ { -+ if (field_idx == 1 && buffer_first_field) ++ /* No character is left in BUF. */ ++ if (buflen < 1) + { -+ int len = 0; ++ fp = next_file (fp); + -+ while (1) ++ if (fp == NULL) ++ break; /* No more files. */ ++ else + { -+ REFILL_BUFFER (buf, bufpos, buflen, stream); -+ -+ GET_NEXT_WC_FROM_BUFFER -+ (wc, bufpos, buflen, mblength, state, convfail); -+ -+ if (wc == WEOF) -+ break; ++ memset (&i_state, '\0', sizeof(mbstate_t)); ++ continue; ++ } ++ } + -+ field_1_buffer = xrealloc (field_1_buffer, len + mblength); -+ memcpy (field_1_buffer + len, bufpos, mblength); -+ len += mblength; -+ buflen -= mblength; -+ bufpos += mblength; ++ /* Get a wide character. */ ++ i_state_bak = i_state; ++ mblength = mbrtowc (&wc, bufpos, buflen, &i_state); + -+ if (!convfail && (wc == L'\n' || wc == wcdelim)) -+ break; ++ switch (mblength) ++ { ++ case (size_t)-1: /* illegal byte sequence. */ ++ case (size_t)-2: ++ mblength = 1; ++ i_state = i_state_bak; ++ if (convert) ++ { ++ ++column; ++ if (convert_entire_line == 0) ++ convert = 0; + } ++ putchar (*bufpos); ++ break; + -+ if (wc == WEOF) -+ break; ++ case 0: /* null. */ ++ mblength = 1; ++ if (convert && convert_entire_line == 0) ++ convert = 0; ++ putchar ('\0'); ++ break; + -+ /* If the first field extends to the end of line (it is not -+ delimited) and we are printing all non-delimited lines, -+ print this one. */ -+ if (convfail || (!convfail && wc != wcdelim)) ++ default: ++ if (wc == L'\n') /* LF. */ + { -+ if (suppress_non_delimited) ++ tab_index = 0; ++ column = 0; ++ convert = 1; ++ putchar ('\n'); ++ } ++ else if (wc == L'\t' && convert) /* Tab. */ ++ { ++ if (tab_size == 0) + { -+ /* Empty. */ ++ /* Do not let tab_index == first_free_tab; ++ stop when it is 1 less. */ ++ while (tab_index < first_free_tab - 1 ++ && column >= tab_list[tab_index]) ++ tab_index++; ++ next_tab_column = tab_list[tab_index]; ++ if (tab_index < first_free_tab - 1) ++ tab_index++; ++ if (column >= next_tab_column) ++ next_tab_column = column + 1; + } + else -+ { -+ fwrite (field_1_buffer, sizeof (char), len, stdout); -+ /* Make sure the output line is newline terminated. */ -+ if (convfail || (!convfail && wc != L'\n')) -+ putchar ('\n'); -+ } -+ continue; -+ } -+ -+ if (print_kth (1, NULL)) -+ { -+ /* Print the field, but not the trailing delimiter. */ -+ fwrite (field_1_buffer, sizeof (char), len - 1, stdout); -+ found_any_selected_field = 1; -+ } -+ ++field_idx; -+ } ++ next_tab_column = column + tab_size - column % tab_size; + -+ if (wc != WEOF) -+ { -+ if (print_kth (field_idx, NULL)) -+ { -+ if (found_any_selected_field) ++ while (column < next_tab_column) + { -+ fwrite (output_delimiter_string, sizeof (char), -+ output_delimiter_length, stdout); ++ putchar (' '); ++ ++column; + } -+ found_any_selected_field = 1; + } -+ -+ while (1) ++ else /* Others. */ + { -+ REFILL_BUFFER (buf, bufpos, buflen, stream); -+ -+ GET_NEXT_WC_FROM_BUFFER -+ (wc, bufpos, buflen, mblength, state, convfail); -+ -+ if (wc == WEOF) -+ break; -+ else if (!convfail && (wc == wcdelim || wc == L'\n')) ++ if (convert) + { -+ buflen -= mblength; -+ bufpos += mblength; -+ break; -+ } -+ -+ if (print_kth (field_idx, NULL)) -+ fwrite (bufpos, mblength, sizeof(char), stdout); ++ if (wc == L'\b') ++ { ++ if (column > 0) ++ --column; ++ } ++ else ++ { ++ int width; /* The width of WC. */ + -+ buflen -= mblength; -+ bufpos += mblength; ++ width = wcwidth (wc); ++ column += (width > 0) ? width : 0; ++ if (convert_entire_line == 0) ++ convert = 0; ++ } ++ } ++ fwrite (bufpos, sizeof(char), mblength, stdout); + } + } -+ -+ if ((!convfail || wc == L'\n') && buflen < 1) -+ wc = WEOF; -+ -+ if (!convfail && wc == wcdelim) -+ ++field_idx; -+ else if (wc == WEOF || (!convfail && wc == L'\n')) -+ { -+ if (found_any_selected_field -+ || (!empty_input && !(suppress_non_delimited && field_idx == 1))) -+ putchar ('\n'); -+ if (wc == WEOF) -+ break; -+ field_idx = 1; -+ found_any_selected_field = 0; -+ } ++ buflen -= mblength; ++ bufpos += mblength; + } +} +#endif + - static void - cut_stream (FILE *stream) + int + main (int argc, char **argv) { -- if (operating_mode == byte_mode) -- cut_bytes (stream); +@@ -428,7 +582,12 @@ + + file_list = (optind < argc ? &argv[optind] : stdin_argv); + +- expand (); +#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1 && !force_singlebyte_mode) -+ { -+ switch (operating_mode) -+ { -+ case byte_mode: -+ if (byte_mode_character_aware) -+ cut_characters_or_cut_bytes_no_split (stream); -+ else -+ cut_bytes (stream); -+ break; -+ -+ case character_mode: -+ cut_characters_or_cut_bytes_no_split (stream); -+ break; -+ -+ case field_mode: -+ cut_fields_mb (stream); -+ break; -+ -+ default: -+ abort (); -+ } -+ } - else -- cut_fields (stream); -+#endif -+ { -+ if (operating_mode == field_mode) -+ cut_fields (stream); -+ else -+ cut_bytes (stream); -+ } - } - - /* Process file FILE to standard output. -@@ -745,6 +1063,8 @@ - bool ok; - bool delim_specified = false; - char *spec_list_string IF_LINT(= NULL); -+ char mbdelim[MB_LEN_MAX + 1]; -+ size_t delimlen = 0; - - initialize_main (&argc, &argv); - program_name = argv[0]; -@@ -767,7 +1087,6 @@ - switch (optc) - { - case 'b': -- case 'c': - /* Build the byte list. */ - if (operating_mode != undefined_mode) - FATAL_ERROR (_("only one type of list may be specified")); -@@ -775,6 +1094,14 @@ - spec_list_string = optarg; - break; - -+ case 'c': -+ /* Build the character list. */ -+ if (operating_mode != undefined_mode) -+ FATAL_ERROR (_("only one type of list may be specified")); -+ operating_mode = character_mode; -+ spec_list_string = optarg; -+ break; -+ - case 'f': - /* Build the field list. */ - if (operating_mode != undefined_mode) -@@ -786,10 +1113,35 @@ - case 'd': - /* New delimiter. */ - /* Interpret -d '' to mean `use the NUL byte as the delimiter.' */ -- if (optarg[0] != '\0' && optarg[1] != '\0') -- FATAL_ERROR (_("the delimiter must be a single character")); -- delim = optarg[0]; -- delim_specified = true; -+#if HAVE_MBRTOWC -+ { -+ if(MB_CUR_MAX > 1) -+ { -+ mbstate_t state; -+ -+ memset (&state, '\0', sizeof(mbstate_t)); -+ delimlen = mbrtowc (&wcdelim, optarg, strnlen(optarg, MB_LEN_MAX), &state); -+ -+ if (delimlen == (size_t)-1 || delimlen == (size_t)-2) -+ ++force_singlebyte_mode; -+ else -+ { -+ delimlen = (delimlen < 1) ? 1 : delimlen; -+ if (wcdelim != L'\0' && *(optarg + delimlen) != '\0') -+ FATAL_ERROR (_("the delimiter must be a single character")); -+ memcpy (mbdelim, optarg, delimlen); -+ } -+ } -+ -+ if (MB_CUR_MAX <= 1 || force_singlebyte_mode) -+#endif -+ { -+ if (optarg[0] != '\0' && optarg[1] != '\0') -+ FATAL_ERROR (_("the delimiter must be a single character")); -+ delim = (unsigned char) optarg[0]; -+ } -+ delim_specified = true; -+ } - break; - - case OUTPUT_DELIMITER_OPTION: -@@ -802,6 +1154,7 @@ - break; - - case 'n': -+ byte_mode_character_aware = 1; - break; - - case 's': -@@ -824,7 +1177,7 @@ - if (operating_mode == undefined_mode) - FATAL_ERROR (_("you must specify a list of bytes, characters, or fields")); - -- if (delim != '\0' && operating_mode != field_mode) -+ if (delim_specified && operating_mode != field_mode) - FATAL_ERROR (_("an input delimiter may be specified only\ - when operating on fields")); - -@@ -851,15 +1204,34 @@ - } - - if (!delim_specified) -- delim = '\t'; -+ { -+ delim = '\t'; -+#ifdef HAVE_MBRTOWC -+ wcdelim = L'\t'; -+ mbdelim[0] = '\t'; -+ mbdelim[1] = '\0'; -+ delimlen = 1; -+#endif -+ } - - if (output_delimiter_string == NULL) - { -- static char dummy[2]; -- dummy[0] = delim; -- dummy[1] = '\0'; -- output_delimiter_string = dummy; -- output_delimiter_length = 1; -+#ifdef HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1 && !force_singlebyte_mode) -+ { -+ output_delimiter_string = xstrdup(mbdelim); -+ output_delimiter_length = delimlen; -+ } -+ -+ if (MB_CUR_MAX <= 1 || force_singlebyte_mode) ++ if (MB_CUR_MAX > 1) ++ expand_multibyte (); ++ else +#endif -+ { -+ static char dummy[2]; -+ dummy[0] = delim; -+ dummy[1] = '\0'; -+ output_delimiter_string = dummy; -+ output_delimiter_length = 1; -+ } - } ++ expand (); - if (optind == argc) ---- coreutils-5.93/src/pr.c 2005-12-23 08:53:01.000000000 +0000 -+++ coreutils-5.95/src/pr.c 2006-05-15 13:45:56.000000000 +0100 + if (have_read_stdin && fclose (stdin) != 0) + error (EXIT_FAILURE, errno, "-"); +--- coreutils-5.95/src/pr.c.i18n 2005-09-16 08:50:33.000000000 +0100 ++++ coreutils-5.95/src/pr.c 2006-05-15 15:09:08.000000000 +0100 @@ -313,6 +313,32 @@ #include @@ -846,7 +425,7 @@ use_col_separator = true; if (optarg) separator_string (optarg); -@@ -1188,17 +1279,52 @@ +@@ -1188,10 +1279,45 @@ a number. */ static void @@ -894,14 +473,6 @@ if (*arg) { long int tmp_long; - We've to use 8 as the lower limit, if we use chars_per_default_tab = 8 - to expand a tab which is not an input_tab-char. */ - free (clump_buff); -- clump_buff = xmalloc (MAX (8, chars_per_input_tab)); -+ clump_buff = xmalloc (MAX (80, chars_per_input_tab)); - } - - /* Open the necessary files, @@ -1256,7 +1382,7 @@ else col_sep_string = column_separator; @@ -934,6 +505,15 @@ if (chars_per_column < 1) error (EXIT_FAILURE, 0, _("page width too narrow")); +@@ -1324,7 +1450,7 @@ + We've to use 8 as the lower limit, if we use chars_per_default_tab = 8 + to expand a tab which is not an input_tab-char. */ + free (clump_buff); +- clump_buff = xmalloc (MAX (8, chars_per_input_tab)); ++ clump_buff = xmalloc (MAX (80, chars_per_input_tab)); + } + + /* Open the necessary files, @@ -1432,7 +1558,7 @@ /* Enlarge p->start_position of first column to use the same form of @@ -1347,1479 +927,1711 @@ /* We've just printed some files and need to clean up things before looking for more options and printing the next batch of files. ---- coreutils-5.93/src/uniq.c.i18n 2005-07-05 07:32:54.000000000 +0100 -+++ coreutils-5.93/src/uniq.c 2005-12-23 08:53:01.000000000 +0000 -@@ -23,6 +23,16 @@ - #include +--- coreutils-5.95/src/join.c.i18n 2005-08-12 08:16:25.000000000 +0100 ++++ coreutils-5.95/src/join.c 2006-05-15 15:08:57.000000000 +0100 +@@ -23,16 +23,30 @@ #include + #include -+/* Get mbstate_t, mbrtowc(). */ ++/* Get mbstate_t, mbrtowc(), mbrtowc(), wcwidth(). */ +#if HAVE_WCHAR_H +# include +#endif + -+/* Get isw* functions. */ ++/* Get iswblank(), towupper. */ +#if HAVE_WCTYPE_H +# include +#endif + #include "system.h" - #include "argmatch.h" + #include "error.h" + #include "hard-locale.h" #include "linebuffer.h" -@@ -32,7 +42,19 @@ +-#include "memcasecmp.h" #include "quote.h" + #include "stdio--.h" #include "xmemcoll.h" #include "xstrtol.h" --#include "memcasecmp.h" -+#include "xmemcoll.h" -+ -+/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC -+ installation; work around this configuration error. */ -+#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 -+# define MB_LEN_MAX 16 -+#endif -+ + +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ +#if HAVE_MBRTOWC && defined mbstate_t +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) +#endif + - /* The official name of this program (e.g., no `g' prefix). */ - #define PROGRAM_NAME "uniq" -@@ -109,6 +131,10 @@ - /* Select whether/how to delimit groups of duplicate lines. */ - static enum delimit_method delimit_groups; + #define PROGRAM_NAME "join" -+/* Function pointers. */ -+static char * -+(*find_field) (struct linebuffer *line); +@@ -104,10 +118,12 @@ + /* Last element in `outlist', where a new element can be added. */ + static struct outlist *outlist_end = &outlist_head; + +-/* Tab character separating fields. If negative, fields are separated +- by any nonempty string of blanks, otherwise by exactly one +- tab character whose value (when cast to unsigned char) equals TAB. */ +-static int tab = -1; ++/* Tab character separating fields. If NULL, fields are separated ++ by any nonempty string of blanks. */ ++static char *tab = NULL; + ++/* The number of bytes used for tab. */ ++static size_t tablen = 0; + static struct option const longopts[] = { - {"count", no_argument, NULL, 'c'}, -@@ -189,7 +215,7 @@ - return a pointer to the beginning of the line's field to be compared. */ +@@ -197,6 +213,8 @@ - static char * --find_field (const struct linebuffer *line) -+find_field_uni (struct linebuffer *line) - { - size_t count; - char *lp = line->buffer; -@@ -210,6 +236,83 @@ - return lp + i; - } + /* Fill in the `fields' structure in LINE. */ -+#if HAVE_MBRTOWC ++/* Fill in the `fields' structure in LINE. */ + -+# define MBCHAR_TO_WCHAR(WC, MBLENGTH, LP, POS, SIZE, STATEP, CONVFAIL) \ -+ do \ -+ { \ -+ mbstate_t state_bak; \ -+ \ -+ CONVFAIL = 0; \ -+ state_bak = *STATEP; \ -+ \ -+ MBLENGTH = mbrtowc (&WC, LP + POS, SIZE - POS, STATEP); \ -+ \ -+ switch (MBLENGTH) \ -+ { \ -+ case (size_t)-2: \ -+ case (size_t)-1: \ -+ *STATEP = state_bak; \ -+ CONVFAIL++; \ -+ /* Fall through */ \ -+ case 0: \ -+ MBLENGTH = 1; \ -+ } \ -+ } \ -+ while (0) -+ -+static char * -+find_field_multi (struct linebuffer *line) + static void + xfields (struct line *line) + { +@@ -206,10 +224,11 @@ + if (ptr == lim) + return; + +- if (0 <= tab) ++ if (tab != NULL) + { ++ unsigned char t = tab[0]; + char *sep; +- for (; (sep = memchr (ptr, tab, lim - ptr)) != NULL; ptr = sep + 1) ++ for (; (sep = memchr (ptr, t, lim - ptr)) != NULL; ptr = sep + 1) + extract_field (line, ptr, sep - ptr); + } + else +@@ -236,6 +255,148 @@ + extract_field (line, ptr, lim - ptr); + } + ++#if HAVE_MBRTOWC ++static void ++xfields_multibyte (struct line *line) +{ -+ size_t count; -+ char *lp = line->buffer; -+ size_t size = line->length - 1; -+ size_t pos; -+ size_t mblength; -+ wchar_t wc; -+ mbstate_t *statep; -+ int convfail; ++ char *ptr = line->buf.buffer; ++ char const *lim = ptr + line->buf.length - 1; ++ wchar_t wc = 0; ++ size_t mblength = 1; ++ mbstate_t state, state_bak; + -+ pos = 0; -+ statep = &(line->state); ++ memset (&state, 0, sizeof (mbstate_t)); + -+ /* skip fields. */ -+ for (count = 0; count < skip_fields && pos < size; count++) ++ if (ptr == lim) ++ return; ++ ++ if (tab != NULL) + { -+ while (pos < size) ++ unsigned char t = tab[0]; ++ char *sep = ptr; ++ for (; ptr < lim; ptr = sep + mblength) + { -+ MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); -+ -+ if (convfail || !iswblank (wc)) ++ sep = ptr; ++ while (sep < lim) + { -+ pos += mblength; -+ break; -+ } -+ pos += mblength; -+ } ++ state_bak = state; ++ mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); + -+ while (pos < size) -+ { -+ MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ { ++ mblength = 1; ++ state = state_bak; ++ } ++ mblength = (mblength < 1) ? 1 : mblength; + -+ if (!convfail && iswblank (wc)) ++ if (mblength == tablen && !memcmp (sep, tab, mblength)) ++ break; ++ else ++ { ++ sep += mblength; ++ continue; ++ } ++ } ++ ++ if (sep == lim) + break; + -+ pos += mblength; ++ extract_field (line, ptr, sep - ptr); + } + } -+ -+ /* skip fields. */ -+ for (count = 0; count < skip_chars && pos < size; count++) ++ else + { -+ MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); -+ pos += mblength; -+ } -+ -+ return lp + pos; -+} -+#endif -+ - /* Return false if two strings OLD and NEW match, true if not. - OLD and NEW point not to the beginnings of the lines - but rather to the beginnings of the fields to compare. -@@ -218,6 +321,8 @@ - static bool - different (char *old, char *new, size_t oldlen, size_t newlen) - { -+ char *copy_old, *copy_new; ++ /* Skip leading blanks before the first field. */ ++ while(ptr < lim) ++ { ++ state_bak = state; ++ mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); + - if (check_chars < oldlen) - oldlen = check_chars; - if (check_chars < newlen) -@@ -225,14 +330,92 @@ - - if (ignore_case) - { -- /* FIXME: This should invoke strcoll somehow. */ -- return oldlen != newlen || memcasecmp (old, new, oldlen); -+ size_t i; ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ { ++ mblength = 1; ++ state = state_bak; ++ break; ++ } ++ mblength = (mblength < 1) ? 1 : mblength; + -+ copy_old = alloca (oldlen + 1); -+ copy_new = alloca (oldlen + 1); ++ if (!iswblank(wc)) ++ break; ++ ptr += mblength; ++ } + -+ for (i = 0; i < oldlen; i++) ++ do + { -+ copy_old[i] = toupper (old[i]); -+ copy_new[i] = toupper (new[i]); -+ } - } -- else if (hard_LC_COLLATE) -- return xmemcoll (old, oldlen, new, newlen) != 0; - else -- return oldlen != newlen || memcmp (old, new, oldlen); -+ { -+ copy_old = (char *)old; -+ copy_new = (char *)new; -+ } -+ -+ return xmemcoll (copy_old, oldlen, copy_new, newlen); -+} ++ char *sep; ++ state_bak = state; ++ mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ { ++ mblength = 1; ++ state = state_bak; ++ break; ++ } ++ mblength = (mblength < 1) ? 1 : mblength; + -+#if HAVE_MBRTOWC -+static int -+different_multi (const char *old, const char *new, size_t oldlen, size_t newlen, mbstate_t oldstate, mbstate_t newstate) -+{ -+ size_t i, j, chars; -+ const char *str[2]; -+ char *copy[2]; -+ size_t len[2]; -+ mbstate_t state[2]; -+ size_t mblength; -+ wchar_t wc, uwc; -+ mbstate_t state_bak; ++ sep = ptr + mblength; ++ while (sep != lim) ++ { ++ state_bak = state; ++ mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ { ++ mblength = 1; ++ state = state_bak; ++ break; ++ } ++ mblength = (mblength < 1) ? 1 : mblength; + -+ str[0] = old; -+ str[1] = new; -+ len[0] = oldlen; -+ len[1] = newlen; -+ state[0] = oldstate; -+ state[1] = newstate; ++ if (iswblank (wc)) ++ break; + -+ for (i = 0; i < 2; i++) -+ { -+ copy[i] = alloca (len[i] + 1); ++ sep += mblength; ++ } + -+ for (j = 0, chars = 0; j < len[i] && chars < check_chars; chars++) -+ { -+ state_bak = state[i]; -+ mblength = mbrtowc (&wc, str[i] + j, len[i] - j, &(state[i])); ++ extract_field (line, ptr, sep - ptr); ++ if (sep == lim) ++ return; + -+ switch (mblength) ++ state_bak = state; ++ mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) + { -+ case (size_t)-1: -+ case (size_t)-2: -+ state[i] = state_bak; -+ /* Fall through */ -+ case 0: + mblength = 1; ++ state = state_bak; + break; ++ } ++ mblength = (mblength < 1) ? 1 : mblength; + -+ default: -+ if (ignore_case) ++ ptr = sep + mblength; ++ while (ptr != lim) ++ { ++ state_bak = state; ++ mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) + { -+ uwc = towupper (wc); ++ mblength = 1; ++ state = state_bak; ++ break; ++ } ++ mblength = (mblength < 1) ? 1 : mblength; + -+ if (uwc != wc) -+ { -+ mbstate_t state_wc; ++ if (!iswblank (wc)) ++ break; + -+ memset (&state_wc, '\0', sizeof(mbstate_t)); -+ wcrtomb (copy[i] + j, uwc, &state_wc); -+ } -+ else -+ memcpy (copy[i] + j, str[i] + j, mblength); -+ } -+ else -+ memcpy (copy[i] + j, str[i] + j, mblength); ++ ptr += mblength; + } -+ j += mblength; + } -+ copy[i][j] = '\0'; -+ len[i] = j; ++ while (ptr != lim); + } + -+ return xmemcoll (copy[0], len[0], copy[1], len[1]); - } -+#endif - - /* Output the line in linebuffer LINE to standard output - provided that the switches say it should be output. -@@ -286,15 +469,43 @@ - { - char *prevfield IF_LINT (= NULL); - size_t prevlen IF_LINT (= 0); -+#if HAVE_MBRTOWC -+ mbstate_t prevstate; -+ -+ memset (&prevstate, '\0', sizeof (mbstate_t)); -+#endif - - while (!feof (stdin)) - { - char *thisfield; - size_t thislen; -+#if HAVE_MBRTOWC -+ mbstate_t thisstate; ++ extract_field (line, ptr, lim - ptr); ++} +#endif + - if (readlinebuffer (thisline, stdin) == 0) - break; - thisfield = find_field (thisline); - thislen = thisline->length - 1 - (thisfield - thisline->buffer); -+#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ { -+ thisstate = thisline->state; -+ -+ if (prevline->length == 0 || different_multi -+ (thisfield, prevfield, thislen, prevlen, thisstate, prevstate)) -+ { -+ fwrite (thisline->buffer, sizeof (char), -+ thisline->length, stdout); -+ -+ SWAP_LINES (prevline, thisline); -+ prevfield = thisfield; -+ prevlen = thislen; -+ prevstate = thisstate; -+ } -+ } -+ else -+#endif - if (prevline->length == 0 - || different (thisfield, prevfield, thislen, prevlen)) - { -@@ -313,17 +524,26 @@ - size_t prevlen; - uintmax_t match_count = 0; - bool first_delimiter = true; -+#if HAVE_MBRTOWC -+ mbstate_t prevstate; -+#endif + /* Read a line from FP into LINE and split it into fields. + Return true if successful. */ - if (readlinebuffer (prevline, stdin) == 0) - goto closefiles; - prevfield = find_field (prevline); - prevlen = prevline->length - 1 - (prevfield - prevline->buffer); +@@ -256,6 +417,11 @@ + line->nfields_allocated = 0; + line->nfields = 0; + line->fields = NULL; +#if HAVE_MBRTOWC -+ prevstate = prevline->state; ++ if (MB_CUR_MAX > 1) ++ xfields_multibyte (line); ++ else +#endif + xfields (line); + return true; + } +@@ -310,56 +476,114 @@ + keycmp (struct line const *line1, struct line const *line2) + { + /* Start of field to compare in each file. */ +- char *beg1; +- char *beg2; +- +- size_t len1; +- size_t len2; /* Length of fields to compare. */ ++ char *beg[2]; ++ char *copy[2]; ++ size_t len[2]; /* Length of fields to compare. */ + int diff; ++ int i, j; - while (!feof (stdin)) - { - bool match; - char *thisfield; - size_t thislen; -+#if HAVE_MBRTOWC -+ mbstate_t thisstate; -+#endif - if (readlinebuffer (thisline, stdin) == 0) - { - if (ferror (stdin)) -@@ -332,6 +552,15 @@ - } - thisfield = find_field (thisline); - thislen = thisline->length - 1 - (thisfield - thisline->buffer); -+#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ { -+ thisstate = thisline->state; -+ match = !different_multi (thisfield, prevfield, -+ thislen, prevlen, thisstate, prevstate); -+ } -+ else -+#endif - match = !different (thisfield, prevfield, thislen, prevlen); - match_count += match; + if (join_field_1 < line1->nfields) + { +- beg1 = line1->fields[join_field_1].beg; +- len1 = line1->fields[join_field_1].len; ++ beg[0] = line1->fields[join_field_1].beg; ++ len[0] = line1->fields[join_field_1].len; + } + else + { +- beg1 = NULL; +- len1 = 0; ++ beg[0] = NULL; ++ len[0] = 0; + } -@@ -364,6 +593,9 @@ - SWAP_LINES (prevline, thisline); - prevfield = thisfield; - prevlen = thislen; -+#if HAVE_MBRTOWC -+ prevstate = thisstate; -+#endif - if (!match) - match_count = 0; - } -@@ -408,6 +640,19 @@ + if (join_field_2 < line2->nfields) + { +- beg2 = line2->fields[join_field_2].beg; +- len2 = line2->fields[join_field_2].len; ++ beg[1] = line2->fields[join_field_2].beg; ++ len[1] = line2->fields[join_field_2].len; + } + else + { +- beg2 = NULL; +- len2 = 0; ++ beg[1] = NULL; ++ len[1] = 0; + } - atexit (close_stdout); +- if (len1 == 0) +- return len2 == 0 ? 0 : -1; +- if (len2 == 0) ++ if (len[0] == 0) ++ return len[1] == 0 ? 0 : -1; ++ if (len[1] == 0) + return 1; -+#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ { -+ find_field = find_field_multi; -+ } -+ else -+#endif -+ { -+ find_field = find_field_uni; -+ } + if (ignore_case) + { +- /* FIXME: ignore_case does not work with NLS (in particular, +- with multibyte chars). */ +- diff = memcasecmp (beg1, beg2, MIN (len1, len2)); ++#ifdef HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1) ++ { ++ size_t mblength; ++ wchar_t wc, uwc; ++ mbstate_t state, state_bak; + ++ memset (&state, '\0', sizeof (mbstate_t)); + ++ for (i = 0; i < 2; i++) ++ { ++ copy[i] = alloca (len[i] + 1); + - skip_chars = 0; - skip_fields = 0; - check_chars = SIZE_MAX; ---- coreutils-5.93/src/expand.c.i18n 2005-08-12 08:16:25.000000000 +0100 -+++ coreutils-5.93/src/expand.c 2005-12-23 08:53:01.000000000 +0000 -@@ -38,11 +38,28 @@ - #include - #include - #include ++ for (j = 0; j < MIN (len[0], len[1]);) ++ { ++ state_bak = state; ++ mblength = mbrtowc (&wc, beg[i] + j, len[i] - j, &state); + -+/* Get mbstate_t, mbrtowc(), wcwidth(). */ -+#if HAVE_WCHAR_H -+# include -+#endif ++ switch (mblength) ++ { ++ case (size_t) -1: ++ case (size_t) -2: ++ state = state_bak; ++ /* Fall through */ ++ case 0: ++ mblength = 1; ++ break; + - #include "system.h" - #include "error.h" - #include "quote.h" - #include "xstrndup.h" - -+/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC -+ installation; work around this configuration error. */ -+#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 -+# define MB_LEN_MAX 16 -+#endif ++ default: ++ uwc = towupper (wc); + -+/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ -+#if HAVE_MBRTOWC && defined mbstate_t -+# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) ++ if (uwc != wc) ++ { ++ mbstate_t state_wc; ++ ++ memset (&state_wc, '\0', sizeof (mbstate_t)); ++ wcrtomb (copy[i] + j, uwc, &state_wc); ++ } ++ else ++ memcpy (copy[i] + j, beg[i] + j, mblength); ++ } ++ j += mblength; ++ } ++ copy[i][j] = '\0'; ++ } ++ } ++ else +#endif ++ { ++ for (i = 0; i < 2; i++) ++ { ++ copy[i] = alloca (len[i] + 1); + - /* The official name of this program (e.g., no `g' prefix). */ - #define PROGRAM_NAME "expand" - -@@ -182,6 +199,7 @@ - stops = num_start + len - 1; - } - } ++ for (j = 0; j < MIN (len[0], len[1]); j++) ++ copy[i][j] = toupper (beg[i][j]); + - else - { - error (0, 0, _("tab size contains invalid character(s): %s"), -@@ -364,6 +382,142 @@ ++ copy[i][j] = '\0'; ++ } ++ } + } + else + { +- if (hard_LC_COLLATE) +- return xmemcoll (beg1, len1, beg2, len2); +- diff = memcmp (beg1, beg2, MIN (len1, len2)); ++ copy[0] = (unsigned char *) beg[0]; ++ copy[1] = (unsigned char *) beg[1]; } - } -+#if HAVE_MBRTOWC -+static void -+expand_multibyte (void) -+{ -+ FILE *fp; /* Input strem. */ -+ mbstate_t i_state; /* Current shift state of the input stream. */ -+ mbstate_t i_state_bak; /* Back up the I_STATE. */ -+ mbstate_t o_state; /* Current shift state of the output stream. */ -+ char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ -+ char *bufpos; /* Next read position of BUF. */ -+ size_t buflen = 0; /* The length of the byte sequence in buf. */ -+ wchar_t wc; /* A gotten wide character. */ -+ size_t mblength; /* The byte size of a multibyte character -+ which shows as same character as WC. */ -+ int tab_index = 0; /* Index in `tab_list' of next tabstop. */ -+ int column = 0; /* Column on screen of the next char. */ -+ int next_tab_column; /* Column the next tab stop is on. */ -+ int convert = 1; /* If nonzero, perform translations. */ -+ -+ fp = next_file ((FILE *) NULL); -+ if (fp == NULL) -+ return; -+ -+ memset (&o_state, '\0', sizeof(mbstate_t)); -+ memset (&i_state, '\0', sizeof(mbstate_t)); -+ -+ for (;;) -+ { -+ /* Refill the buffer BUF. */ -+ if (buflen < MB_LEN_MAX && !feof(fp) && !ferror(fp)) -+ { -+ memmove (buf, bufpos, buflen); -+ buflen += fread (buf + buflen, sizeof(char), BUFSIZ, fp); -+ bufpos = buf; -+ } -+ -+ /* No character is left in BUF. */ -+ if (buflen < 1) -+ { -+ fp = next_file (fp); -+ -+ if (fp == NULL) -+ break; /* No more files. */ -+ else -+ { -+ memset (&i_state, '\0', sizeof(mbstate_t)); -+ continue; -+ } -+ } -+ -+ /* Get a wide character. */ -+ i_state_bak = i_state; -+ mblength = mbrtowc (&wc, bufpos, buflen, &i_state); -+ -+ switch (mblength) -+ { -+ case (size_t)-1: /* illegal byte sequence. */ -+ case (size_t)-2: -+ mblength = 1; -+ i_state = i_state_bak; -+ if (convert) -+ { -+ ++column; -+ if (convert_entire_line == 0) -+ convert = 0; -+ } -+ putchar (*bufpos); -+ break; -+ -+ case 0: /* null. */ -+ mblength = 1; -+ if (convert && convert_entire_line == 0) -+ convert = 0; -+ putchar ('\0'); -+ break; ++ if (HAVE_SETLOCALE && hard_LC_COLLATE) ++ return xmemcoll ((char *) copy[0], len[0], (char *) copy[1], len[1]); ++ diff = memcmp (copy[0], copy[1], MIN (len[0], len[1])); + -+ default: -+ if (wc == L'\n') /* LF. */ -+ { -+ tab_index = 0; -+ column = 0; -+ convert = 1; -+ putchar ('\n'); -+ } -+ else if (wc == L'\t' && convert) /* Tab. */ -+ { -+ if (tab_size == 0) -+ { -+ /* Do not let tab_index == first_free_tab; -+ stop when it is 1 less. */ -+ while (tab_index < first_free_tab - 1 -+ && column >= tab_list[tab_index]) -+ tab_index++; -+ next_tab_column = tab_list[tab_index]; -+ if (tab_index < first_free_tab - 1) -+ tab_index++; -+ if (column >= next_tab_column) -+ next_tab_column = column + 1; -+ } -+ else -+ next_tab_column = column + tab_size - column % tab_size; + if (diff) + return diff; +- return len1 < len2 ? -1 : len1 != len2; ++ return len[0] - len[1]; + } + + /* Print field N of LINE if it exists and is nonempty, otherwise +@@ -384,11 +608,18 @@ + + /* Print the join of LINE1 and LINE2. */ + ++#define PUT_TAB_CHAR \ ++ do \ ++ { \ ++ (tab != NULL) ? \ ++ fwrite(tab, sizeof(char), tablen, stdout) : putchar (' '); \ ++ } \ ++ while (0) + -+ while (column < next_tab_column) + static void + prjoin (struct line const *line1, struct line const *line2) + { + const struct outlist *outlist; +- char output_separator = tab < 0 ? ' ' : tab; + + outlist = outlist_head.next; + if (outlist) +@@ -404,12 +635,12 @@ + if (o->file == 0) + { + if (line1 == &uni_blank) +- { + { -+ putchar (' '); -+ ++column; -+ } -+ } -+ else /* Others. */ -+ { -+ if (convert) + line = line2; + field = join_field_2; + } + else +- { + { -+ if (wc == L'\b') -+ { -+ if (column > 0) -+ --column; -+ } -+ else -+ { -+ int width; /* The width of WC. */ -+ -+ width = wcwidth (wc); -+ column += (width > 0) ? width : 0; -+ if (convert_entire_line == 0) -+ convert = 0; -+ } -+ } -+ fwrite (bufpos, sizeof(char), mblength, stdout); -+ } -+ } -+ buflen -= mblength; -+ bufpos += mblength; -+ } -+} -+#endif -+ - int - main (int argc, char **argv) - { -@@ -428,7 +582,12 @@ + line = line1; + field = join_field_1; + } +@@ -423,7 +654,7 @@ + o = o->next; + if (o == NULL) + break; +- putchar (output_separator); ++ PUT_TAB_CHAR; + } + putchar ('\n'); + } +@@ -441,23 +672,23 @@ + prfield (join_field_1, line1); + for (i = 0; i < join_field_1 && i < line1->nfields; ++i) + { +- putchar (output_separator); ++ PUT_TAB_CHAR; + prfield (i, line1); + } + for (i = join_field_1 + 1; i < line1->nfields; ++i) + { +- putchar (output_separator); ++ PUT_TAB_CHAR; + prfield (i, line1); + } - file_list = (optind < argc ? &argv[optind] : stdin_argv); + for (i = 0; i < join_field_2 && i < line2->nfields; ++i) + { +- putchar (output_separator); ++ PUT_TAB_CHAR; + prfield (i, line2); + } + for (i = join_field_2 + 1; i < line2->nfields; ++i) + { +- putchar (output_separator); ++ PUT_TAB_CHAR; + prfield (i, line2); + } + putchar ('\n'); +@@ -869,20 +1100,41 @@ -- expand (); + case 't': + { +- unsigned char newtab = optarg[0]; +- if (! newtab) ++ char *newtab; ++ size_t newtablen; ++ if (! optarg[0]) + error (EXIT_FAILURE, 0, _("empty tab")); +- if (optarg[1]) ++ newtab = xstrdup (optarg); +#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ expand_multibyte (); -+ else ++ if (MB_CUR_MAX > 1) ++ { ++ mbstate_t state; ++ ++ memset (&state, 0, sizeof (mbstate_t)); ++ newtablen = mbrtowc (NULL, newtab, ++ strnlen (newtab, MB_LEN_MAX), ++ &state); ++ if (newtablen == (size_t) 0 ++ || newtablen == (size_t) -1 ++ || newtablen == (size_t) -2) ++ newtablen = 1; ++ } ++ else +#endif -+ expand (); ++ newtablen = 1; ++ ++ if (newtablen == 1 && newtab[1]) ++ { ++ if (STREQ (newtab, "\\0")) ++ newtab[0] = '\0'; ++ } ++ if (tab != NULL && strcmp (tab, newtab)) + { +- if (STREQ (optarg, "\\0")) +- newtab = '\0'; +- else +- error (EXIT_FAILURE, 0, _("multi-character tab %s"), +- quote (optarg)); ++ free (newtab); ++ error (EXIT_FAILURE, 0, _("incompatible tabs")); + } +- if (0 <= tab && tab != newtab) +- error (EXIT_FAILURE, 0, _("incompatible tabs")); + tab = newtab; ++ tablen = newtablen; + } + break; - if (have_read_stdin && fclose (stdin) != 0) - error (EXIT_FAILURE, errno, "-"); ---- coreutils-5.93/src/fold.c.i18n 2005-08-12 08:29:38.000000000 +0100 -+++ coreutils-5.93/src/fold.c 2005-12-23 08:53:01.000000000 +0000 -@@ -23,11 +23,33 @@ +--- coreutils-5.95/src/sort.c.i18n 2005-10-07 19:48:28.000000000 +0100 ++++ coreutils-5.95/src/sort.c 2006-05-15 15:08:57.000000000 +0100 +@@ -23,9 +23,18 @@ + + #include + ++#include #include #include - -+/* Get mbstate_t, mbrtowc(), wcwidth(). */ + #include +#if HAVE_WCHAR_H +# include +#endif -+ -+/* Get iswprint(), iswblank(), wcwidth(). */ ++/* Get isw* functions. */ +#if HAVE_WCTYPE_H +# include +#endif + #include "system.h" #include "error.h" - #include "quote.h" - #include "xstrtol.h" + #include "hard-locale.h" +@@ -95,14 +104,38 @@ + /* Thousands separator; if -1, then there isn't one. */ + static int thousands_sep; -+/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC -+ installation; work around this configuration error. */ -+#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 -+# undef MB_LEN_MAX -+# define MB_LEN_MAX 16 ++static int force_general_numcompare = 0; ++ + /* Nonzero if the corresponding locales are hard. */ + static bool hard_LC_COLLATE; +-#if HAVE_NL_LANGINFO ++#if HAVE_LANGINFO_CODESET + static bool hard_LC_TIME; + #endif + + #define NONZERO(x) ((x) != 0) + ++/* get a multibyte character's byte length. */ ++#define GET_BYTELEN_OF_CHAR(LIM, PTR, MBLENGTH, STATE) \ ++ do \ ++ { \ ++ wchar_t wc; \ ++ mbstate_t state_bak; \ ++ \ ++ state_bak = STATE; \ ++ mblength = mbrtowc (&wc, PTR, LIM - PTR, &STATE); \ ++ \ ++ switch (MBLENGTH) \ ++ { \ ++ case (size_t)-1: \ ++ case (size_t)-2: \ ++ STATE = state_bak; \ ++ /* Fall through. */ \ ++ case 0: \ ++ MBLENGTH = 1; \ ++ } \ ++ } \ ++ while (0) ++ + /* The kind of blanks for '-b' to skip in various options. */ + enum blanktype { bl_start, bl_end, bl_both }; + +@@ -239,13 +272,11 @@ + they were read if all keys compare equal. */ + static bool stable; + +-/* If TAB has this value, blanks separate fields. */ +-enum { TAB_DEFAULT = CHAR_MAX + 1 }; +- +-/* Tab character separating fields. If TAB_DEFAULT, then fields are ++/* Tab character separating fields. If tab_length is 0, then fields are + separated by the empty string between a non-blank character and a blank + character. */ +-static int tab = TAB_DEFAULT; ++static char tab[MB_LEN_MAX + 1]; ++static size_t tab_length = 0; + + /* Flag to remove consecutive duplicate lines from the output. + Only the last of a sequence of equal lines will be output. */ +@@ -392,6 +423,44 @@ + static struct tempnode *volatile temphead; + static struct tempnode *volatile *temptail = &temphead; + ++/* Function pointers. */ ++static void ++(*inittables) (void); ++static char * ++(*begfield) (const struct line*, const struct keyfield *); ++static char * ++(*limfield) (const struct line*, const struct keyfield *); ++static int ++(*getmonth) (char const *, size_t); ++static int ++(*keycompare) (const struct line *, const struct line *); ++static int ++(*numcompare) (const char *, const char *); ++ ++/* Test for white space multibyte character. ++ Set LENGTH the byte length of investigated multibyte character. */ ++#if HAVE_MBRTOWC ++static int ++ismbblank (const char *str, size_t len, size_t *length) ++{ ++ size_t mblength; ++ wchar_t wc; ++ mbstate_t state; ++ ++ memset (&state, '\0', sizeof(mbstate_t)); ++ mblength = mbrtowc (&wc, str, len, &state); ++ ++ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ { ++ *length = 1; ++ return 0; ++ } ++ ++ *length = (mblength < 1) ? 1 : mblength; ++ return iswblank (wc); ++} +#endif + -+/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ -+#if HAVE_MBRTOWC && defined mbstate_t -+# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) + /* Clean up any remaining temporary files. */ + + static void +@@ -545,7 +614,7 @@ + free (node); + } + +-#if HAVE_NL_LANGINFO ++#if HAVE_LANGINFO_CODESET + + static int + struct_month_cmp (const void *m1, const void *m2) +@@ -560,7 +629,7 @@ + /* Initialize the character class tables. */ + + static void +-inittables (void) ++inittables_uni (void) + { + size_t i; + +@@ -572,7 +641,7 @@ + fold_toupper[i] = (ISLOWER (i) ? toupper (i) : i); + } + +-#if HAVE_NL_LANGINFO ++#if HAVE_LANGINFO_CODESET + /* If we're not in the "C" locale, read different names for months. */ + if (hard_LC_TIME) + { +@@ -598,6 +667,64 @@ + #endif + } + ++#if HAVE_MBRTOWC ++static void ++inittables_mb (void) ++{ ++ int i, j, k, l; ++ char *name, *s; ++ size_t s_len, mblength; ++ char mbc[MB_LEN_MAX]; ++ wchar_t wc, pwc; ++ mbstate_t state_mb, state_wc; ++ ++ for (i = 0; i < MONTHS_PER_YEAR; i++) ++ { ++ s = (char *) nl_langinfo (ABMON_1 + i); ++ s_len = strlen (s); ++ monthtab[i].name = name = (char *) xmalloc (s_len + 1); ++ monthtab[i].val = i + 1; ++ ++ memset (&state_mb, '\0', sizeof (mbstate_t)); ++ memset (&state_wc, '\0', sizeof (mbstate_t)); ++ ++ for (j = 0; j < s_len;) ++ { ++ if (!ismbblank (s + j, s_len - j, &mblength)) ++ break; ++ j += mblength; ++ } ++ ++ for (k = 0; j < s_len;) ++ { ++ mblength = mbrtowc (&wc, (s + j), (s_len - j), &state_mb); ++ assert (mblength != (size_t)-1 && mblength != (size_t)-2); ++ if (mblength == 0) ++ break; ++ ++ pwc = towupper (wc); ++ if (pwc == wc) ++ { ++ memcpy (mbc, s + j, mblength); ++ j += mblength; ++ } ++ else ++ { ++ j += mblength; ++ mblength = wcrtomb (mbc, pwc, &state_wc); ++ assert (mblength != (size_t)0 && mblength != (size_t)-1); ++ } ++ ++ for (l = 0; l < mblength; l++) ++ name[k++] = mbc[l]; ++ } ++ name[k] = '\0'; ++ } ++ qsort ((void *) monthtab, MONTHS_PER_YEAR, ++ sizeof (struct month), struct_month_cmp); ++} +#endif + - #define TAB_WIDTH 8 + /* Specify the amount of main memory to use when sorting. */ + static void + specify_sort_size (char const *s) +@@ -808,7 +935,7 @@ + by KEY in LINE. */ - /* The official name of this program (e.g., no `g' prefix). */ -@@ -35,23 +57,44 @@ + static char * +-begfield (const struct line *line, const struct keyfield *key) ++begfield_uni (const struct line *line, const struct keyfield *key) + { + char *ptr = line->text, *lim = ptr + line->length - 1; + size_t sword = key->sword; +@@ -818,10 +945,10 @@ + /* The leading field separator itself is included in a field when -t + is absent. */ - #define AUTHORS "David MacKenzie" +- if (tab != TAB_DEFAULT) ++ if (tab_length) + while (ptr < lim && sword--) + { +- while (ptr < lim && *ptr != tab) ++ while (ptr < lim && *ptr != tab[0]) + ++ptr; + if (ptr < lim) + ++ptr; +@@ -849,11 +976,70 @@ + return ptr; + } -+#define FATAL_ERROR(Message) \ -+ do \ -+ { \ -+ error (0, 0, (Message)); \ -+ usage (2); \ -+ } \ -+ while (0) -+ -+enum operating_mode ++#if HAVE_MBRTOWC ++static char * ++begfield_mb (const struct line *line, const struct keyfield *key) +{ -+ /* Fold texts by columns that are at the given positions. */ -+ column_mode, ++ int i; ++ char *ptr = line->text, *lim = ptr + line->length - 1; ++ size_t sword = key->sword; ++ size_t schar = key->schar; ++ size_t mblength; ++ mbstate_t state; + -+ /* Fold texts by bytes that are at the given positions. */ -+ byte_mode, ++ memset (&state, '\0', sizeof(mbstate_t)); + -+ /* Fold texts by characters that are at the given positions. */ -+ character_mode, -+}; ++ if (tab_length) ++ while (ptr < lim && sword--) ++ { ++ while (ptr < lim && memcmp (ptr, tab, tab_length) != 0) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ if (ptr < lim) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ } ++ else ++ while (ptr < lim && sword--) ++ { ++ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) ++ ptr += mblength; ++ if (ptr < lim) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ while (ptr < lim && !ismbblank (ptr, lim - ptr, &mblength)) ++ ptr += mblength; ++ } + - /* The name this program was run with. */ - char *program_name; - -+/* The argument shows current mode. (Default: column_mode) */ -+static enum operating_mode operating_mode; ++ if (key->skipsblanks) ++ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) ++ ptr += mblength; + - /* If nonzero, try to break on whitespace. */ - static bool break_spaces; - --/* If nonzero, count bytes, not column positions. */ --static bool count_bytes; -- - /* If nonzero, at least one of the files we read was standard input. */ - static bool have_read_stdin; - --static char const shortopts[] = "bsw:0::1::2::3::4::5::6::7::8::9::"; -+static char const shortopts[] = "bcsw:0::1::2::3::4::5::6::7::8::9::"; ++ for (i = 0; i < schar; i++) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ++ if (ptr + mblength > lim) ++ break; ++ else ++ ptr += mblength; ++ } ++ ++ return ptr; ++} ++#endif ++ + /* Return the limit of (a pointer to the first character after) the field + in LINE specified by KEY. */ - static struct option const longopts[] = - { - {"bytes", no_argument, NULL, 'b'}, -+ {"characters", no_argument, NULL, 'c'}, - {"spaces", no_argument, NULL, 's'}, - {"width", required_argument, NULL, 'w'}, - {GETOPT_HELP_OPTION_DECL}, -@@ -81,6 +124,7 @@ - "), stdout); - fputs (_("\ - -b, --bytes count bytes rather than columns\n\ -+ -c, --characters count characters rather than columns\n\ - -s, --spaces break at spaces\n\ - -w, --width=WIDTH use WIDTH columns instead of 80\n\ - "), stdout); -@@ -98,7 +142,7 @@ - static size_t - adjust_column (size_t column, char c) + static char * +-limfield (const struct line *line, const struct keyfield *key) ++limfield_uni (const struct line *line, const struct keyfield *key) { -- if (!count_bytes) -+ if (operating_mode != byte_mode) - { - if (c == '\b') - { -@@ -117,35 +161,14 @@ - return column; - } + char *ptr = line->text, *lim = ptr + line->length - 1; + size_t eword = key->eword, echar = key->echar; +@@ -866,10 +1052,10 @@ + `beginning' is the first character following the delimiting TAB. + Otherwise, leave PTR pointing at the first `blank' character after + the preceding field. */ +- if (tab != TAB_DEFAULT) ++ if (tab_length) + while (ptr < lim && eword--) + { +- while (ptr < lim && *ptr != tab) ++ while (ptr < lim && *ptr != tab[0]) + ++ptr; + if (ptr < lim && (eword | echar)) + ++ptr; +@@ -915,10 +1101,10 @@ + */ --/* Fold file FILENAME, or standard input if FILENAME is "-", -- to stdout, with maximum line length WIDTH. -- Return true if successful. */ -- --static bool --fold_file (char *filename, size_t width) -+static void -+fold_text (FILE *istream, size_t width, int *saved_errno) - { -- FILE *istream; - int c; - size_t column = 0; /* Screen column where next char will go. */ - size_t offset_out = 0; /* Index in `line_out' for next char. */ - static char *line_out = NULL; - static size_t allocated_out = 0; -- int saved_errno; -- -- if (STREQ (filename, "-")) -- { -- istream = stdin; -- have_read_stdin = true; -- } -- else -- istream = fopen (filename, "r"); -- -- if (istream == NULL) -- { -- error (0, errno, "%s", filename); -- return false; -- } -- - while ((c = getc (istream)) != EOF) + /* Make LIM point to the end of (one byte past) the current field. */ +- if (tab != TAB_DEFAULT) ++ if (tab_length) { - if (offset_out + 1 >= allocated_out) -@@ -172,6 +195,15 @@ - bool found_blank = false; - size_t logical_end = offset_out; - -+ /* If LINE_OUT has no wide character, -+ put a new wide character in LINE_OUT -+ if column is bigger than width. */ -+ if (offset_out == 0) -+ { -+ line_out[offset_out++] = c; -+ continue; -+ } -+ - /* Look for the last blank. */ - while (logical_end) - { -@@ -218,11 +250,225 @@ - line_out[offset_out++] = c; + char *newlim; +- newlim = memchr (ptr, tab, lim - ptr); ++ newlim = memchr (ptr, tab[0], lim - ptr); + if (newlim) + lim = newlim; } +@@ -951,6 +1137,107 @@ + return ptr; + } -- saved_errno = errno; -+ *saved_errno = errno; ++#if HAVE_MBRTOWC ++static char * ++limfield_mb (const struct line *line, const struct keyfield *key) ++{ ++ char *ptr = line->text, *lim = ptr + line->length - 1; ++ size_t eword = key->eword, echar = key->echar; ++ int i; ++ size_t mblength; ++ mbstate_t state; + -+ if (offset_out) -+ fwrite (line_out, sizeof (char), (size_t) offset_out, stdout); ++ memset (&state, '\0', sizeof(mbstate_t)); + -+ free(line_out); -+} ++ if (tab_length) ++ while (ptr < lim && eword--) ++ { ++ while (ptr < lim && memcmp (ptr, tab, tab_length) != 0) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ if (ptr < lim && (eword | echar)) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ } ++ else ++ while (ptr < lim && eword--) ++ { ++ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) ++ ptr += mblength; ++ if (ptr < lim) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ while (ptr < lim && !ismbblank (ptr, lim - ptr, &mblength)) ++ ptr += mblength; ++ } + -+#if HAVE_MBRTOWC -+static void -+fold_multibyte_text (FILE *istream, int width, int *saved_errno) -+{ -+ char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ -+ size_t buflen = 0; /* The length of the byte sequence in buf. */ -+ char *bufpos; /* Next read position of BUF. */ -+ wint_t wc; /* A gotten wide character. */ -+ size_t mblength; /* The byte size of a multibyte character which shows -+ as same character as WC. */ -+ mbstate_t state, state_bak; /* State of the stream. */ -+ int convfail; /* 1, when conversion is failed. Otherwise 0. */ + -+ char *line_out = NULL; -+ size_t offset_out = 0; /* Index in `line_out' for next char. */ -+ size_t allocated_out = 0; ++# ifdef POSIX_UNSPECIFIED ++ /* Make LIM point to the end of (one byte past) the current field. */ ++ if (tab_length) ++ { ++ char *newlim, *p; + -+ int increment; -+ size_t column = 0; ++ newlim = NULL; ++ for (p = ptr; p < lim;) ++ { ++ if (memcmp (p, tab, tab_length) == 0) ++ { ++ newlim = p; ++ break; ++ } + -+ size_t last_blank_pos; -+ size_t last_blank_column; -+ int is_blank_seen; -+ int last_blank_increment; -+ int is_bs_following_last_blank; -+ size_t bs_following_last_blank_num; -+ int is_cr_after_last_blank; ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ p += mblength; ++ } ++ } ++ else ++ { ++ char *newlim; ++ newlim = ptr; + -+#define CLEAR_FLAGS \ -+ do \ -+ { \ -+ last_blank_pos = 0; \ -+ last_blank_column = 0; \ -+ is_blank_seen = 0; \ -+ is_bs_following_last_blank = 0; \ -+ bs_following_last_blank_num = 0; \ -+ is_cr_after_last_blank = 0; \ -+ } \ -+ while (0) ++ while (newlim < lim && ismbblank (newlim, lim - newlim, &mblength)) ++ newlim += mblength; ++ if (ptr < lim) ++ { ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ ptr += mblength; ++ } ++ while (newlim < lim && !ismbblank (newlim, lim - newlim, &mblength)) ++ newlim += mblength; ++ lim = newlim; ++ } ++# endif + -+#define START_NEW_LINE \ -+ do \ -+ { \ -+ putchar ('\n'); \ -+ column = 0; \ -+ offset_out = 0; \ -+ CLEAR_FLAGS; \ -+ } \ -+ while (0) ++ /* If we're skipping leading blanks, don't start counting characters ++ * until after skipping past any leading blanks. */ ++ if (key->skipsblanks) ++ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) ++ ptr += mblength; + -+ CLEAR_FLAGS; + memset (&state, '\0', sizeof(mbstate_t)); + -+ for (;; bufpos += mblength, buflen -= mblength) ++ /* Advance PTR by ECHAR (if possible), but no further than LIM. */ ++ for (i = 0; i < echar; i++) + { -+ if (buflen < MB_LEN_MAX && !feof (istream) && !ferror (istream)) -+ { -+ memmove (buf, bufpos, buflen); -+ buflen += fread (buf + buflen, sizeof(char), BUFSIZ, istream); -+ bufpos = buf; -+ } ++ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); + -+ if (buflen < 1) ++ if (ptr + mblength > lim) + break; ++ else ++ ptr += mblength; ++ } + -+ /* Get a wide character. */ -+ convfail = 0; -+ state_bak = state; -+ mblength = mbrtowc ((wchar_t *)&wc, bufpos, buflen, &state); -+ -+ switch (mblength) -+ { -+ case (size_t)-1: -+ case (size_t)-2: -+ convfail++; -+ state = state_bak; -+ /* Fall through. */ -+ -+ case 0: -+ mblength = 1; -+ break; -+ } ++ return ptr; ++} ++#endif + -+rescan: -+ if (operating_mode == byte_mode) /* byte mode */ -+ increment = mblength; -+ else if (operating_mode == character_mode) /* character mode */ -+ increment = 1; -+ else /* column mode */ -+ { -+ if (convfail) -+ increment = 1; -+ else -+ { -+ switch (wc) -+ { -+ case L'\n': -+ fwrite (line_out, sizeof(char), offset_out, stdout); -+ START_NEW_LINE; -+ continue; -+ -+ case L'\b': -+ increment = (column > 0) ? -1 : 0; -+ break; + /* Fill BUF reading from FP, moving buf->left bytes from the end + of buf->buf to the beginning first. If EOF is reached and the + file wasn't terminated by a newline, supply one. Set up BUF's line +@@ -1067,7 +1354,7 @@ + hideously fast. */ + + static int +-numcompare (const char *a, const char *b) ++numcompare_uni (const char *a, const char *b) + { + while (blanks[to_uchar (*a)]) + a++; +@@ -1077,6 +1364,25 @@ + return strnumcmp (a, b, decimal_point, thousands_sep); + } + ++#if HAVE_MBRTOWC ++static int ++numcompare_mb (const char *a, const char *b) ++{ ++ size_t mblength, len; ++ len = strlen (a); /* okay for UTF-8 */ ++ while (*a && ismbblank (a, len > MB_CUR_MAX ? MB_CUR_MAX : len, &mblength)) ++ { ++ a += mblength; ++ len -= mblength; ++ } ++ len = strlen (b); /* okay for UTF-8 */ ++ while (*b && ismbblank (b, len > MB_CUR_MAX ? MB_CUR_MAX : len, &mblength)) ++ b += mblength; + -+ case L'\r': -+ increment = -1 * column; -+ break; ++ return strnumcmp (a, b, decimal_point, thousands_sep); ++} ++#endif /* HAV_EMBRTOWC */ + -+ case L'\t': -+ increment = 8 - column % 8; -+ break; + static int + general_numcompare (const char *sa, const char *sb) + { +@@ -1110,7 +1416,7 @@ + Return 0 if the name in S is not recognized. */ + + static int +-getmonth (char const *month, size_t len) ++getmonth_uni (char const *month, size_t len) + { + size_t lo = 0; + size_t hi = MONTHS_PER_YEAR; +@@ -1152,11 +1458,79 @@ + return 0; + } + ++#if HAVE_MBRTOWC ++static int ++getmonth_mb (const char *s, size_t len) ++{ ++ char *month; ++ register size_t i; ++ register int lo = 0, hi = MONTHS_PER_YEAR, result; ++ char *tmp; ++ size_t wclength, mblength; ++ const char **pp; ++ const wchar_t **wpp; ++ wchar_t *month_wcs; ++ mbstate_t state; + -+ default: -+ increment = wcwidth (wc); -+ increment = (increment < 0) ? 0 : increment; -+ } -+ } -+ } ++ while (len > 0 && ismbblank (s, len, &mblength)) ++ { ++ s += mblength; ++ len -= mblength; ++ } + -+ if (column + increment > width && break_spaces && last_blank_pos) -+ { -+ fwrite (line_out, sizeof(char), last_blank_pos, stdout); -+ putchar ('\n'); ++ if (len == 0) ++ return 0; + -+ offset_out = offset_out - last_blank_pos; -+ column = column - last_blank_column + ((is_cr_after_last_blank) -+ ? last_blank_increment : bs_following_last_blank_num); -+ memmove (line_out, line_out + last_blank_pos, offset_out); -+ CLEAR_FLAGS; -+ goto rescan; -+ } ++ month = (char *) alloca (len + 1); + -+ if (column + increment > width && column != 0) -+ { -+ fwrite (line_out, sizeof(char), offset_out, stdout); -+ START_NEW_LINE; -+ goto rescan; -+ } ++ tmp = (char *) alloca (len + 1); ++ memcpy (tmp, s, len); ++ tmp[len] = '\0'; ++ pp = (const char **)&tmp; ++ month_wcs = (wchar_t *) alloca ((len + 1) * sizeof (wchar_t)); ++ memset (&state, '\0', sizeof(mbstate_t)); + -+ if (allocated_out < offset_out + mblength) ++ wclength = mbsrtowcs (month_wcs, pp, len + 1, &state); ++ assert (wclength != (size_t)-1 && *pp == NULL); ++ ++ for (i = 0; i < wclength; i++) ++ { ++ month_wcs[i] = towupper(month_wcs[i]); ++ if (iswblank (month_wcs[i])) + { -+ allocated_out += 1024; -+ line_out = xrealloc (line_out, allocated_out); ++ month_wcs[i] = L'\0'; ++ break; + } ++ } + -+ memcpy (line_out + offset_out, bufpos, mblength); -+ offset_out += mblength; -+ column += increment; ++ wpp = (const wchar_t **)&month_wcs; + -+ if (is_blank_seen && !convfail && wc == L'\r') -+ is_cr_after_last_blank = 1; ++ mblength = wcsrtombs (month, wpp, len + 1, &state); ++ assert (mblength != (-1) && *wpp == NULL); + -+ if (is_bs_following_last_blank && !convfail && wc == L'\b') -+ ++bs_following_last_blank_num; -+ else -+ is_bs_following_last_blank = 0; ++ do ++ { ++ int ix = (lo + hi) / 2; + -+ if (break_spaces && !convfail && iswblank (wc)) -+ { -+ last_blank_pos = offset_out; -+ last_blank_column = column; -+ is_blank_seen = 1; -+ last_blank_increment = increment; -+ is_bs_following_last_blank = 1; -+ bs_following_last_blank_num = 0; -+ is_cr_after_last_blank = 0; -+ } ++ if (strncmp (month, monthtab[ix].name, strlen (monthtab[ix].name)) < 0) ++ hi = ix; ++ else ++ lo = ix; + } ++ while (hi - lo > 1); + -+ *saved_errno = errno; - - if (offset_out) - fwrite (line_out, sizeof (char), (size_t) offset_out, stdout); - -+ free(line_out); ++ result = (!strncmp (month, monthtab[lo].name, strlen (monthtab[lo].name)) ++ ? monthtab[lo].val : 0); ++ ++ return result; +} +#endif + -+/* Fold file FILENAME, or standard input if FILENAME is "-", -+ to stdout, with maximum line length WIDTH. -+ Return 0 if successful, 1 if an error occurs. */ -+ + /* Compare two lines A and B trying every key in sequence until there + are no more keys or a difference is found. */ + + static int +-keycompare (const struct line *a, const struct line *b) ++keycompare_uni (const struct line *a, const struct line *b) + { + struct keyfield const *key = keylist; + +@@ -1326,6 +1700,177 @@ + return key->reverse ? -diff : diff; + } + ++#if HAVE_MBRTOWC +static int -+fold_file (char *filename, int width) ++keycompare_mb (const struct line *a, const struct line *b) +{ -+ FILE *istream; -+ int saved_errno; ++ struct keyfield *key = keylist; + -+ if (STREQ (filename, "-")) -+ { -+ istream = stdin; -+ have_read_stdin = 1; -+ } -+ else -+ istream = fopen (filename, "r"); ++ /* For the first iteration only, the key positions have been ++ precomputed for us. */ ++ char *texta = a->keybeg; ++ char *textb = b->keybeg; ++ char *lima = a->keylim; ++ char *limb = b->keylim; + -+ if (istream == NULL) ++ size_t mblength_a, mblength_b; ++ wchar_t wc_a, wc_b; ++ mbstate_t state_a, state_b; ++ ++ int diff; ++ ++ memset (&state_a, '\0', sizeof(mbstate_t)); ++ memset (&state_b, '\0', sizeof(mbstate_t)); ++ ++ for (;;) + { -+ error (0, errno, "%s", filename); -+ return 1; ++ unsigned char *translate = (unsigned char *) key->translate; ++ bool const *ignore = key->ignore; ++ ++ /* Find the lengths. */ ++ size_t lena = lima <= texta ? 0 : lima - texta; ++ size_t lenb = limb <= textb ? 0 : limb - textb; ++ ++ /* Actually compare the fields. */ ++ if (key->numeric | key->general_numeric) ++ { ++ char savea = *lima, saveb = *limb; ++ ++ *lima = *limb = '\0'; ++ if (force_general_numcompare) ++ diff = general_numcompare (texta, textb); ++ else ++ diff = ((key->numeric ? numcompare : general_numcompare) ++ (texta, textb)); ++ *lima = savea, *limb = saveb; ++ } ++ else if (key->month) ++ diff = getmonth (texta, lena) - getmonth (textb, lenb); ++ else ++ { ++ if (ignore || translate) ++ { ++ char *copy_a = (char *) alloca (lena + 1 + lenb + 1); ++ char *copy_b = copy_a + lena + 1; ++ size_t new_len_a, new_len_b; ++ size_t i, j; ++ ++ /* Ignore and/or translate chars before comparing. */ ++# define IGNORE_CHARS(NEW_LEN, LEN, TEXT, COPY, WC, MBLENGTH, STATE) \ ++ do \ ++ { \ ++ wchar_t uwc; \ ++ char mbc[MB_LEN_MAX]; \ ++ mbstate_t state_wc; \ ++ \ ++ for (NEW_LEN = i = 0; i < LEN;) \ ++ { \ ++ mbstate_t state_bak; \ ++ \ ++ state_bak = STATE; \ ++ MBLENGTH = mbrtowc (&WC, TEXT + i, LEN - i, &STATE); \ ++ \ ++ if (MBLENGTH == (size_t)-2 || MBLENGTH == (size_t)-1 \ ++ || MBLENGTH == 0) \ ++ { \ ++ if (MBLENGTH == (size_t)-2 || MBLENGTH == (size_t)-1) \ ++ STATE = state_bak; \ ++ if (!ignore) \ ++ COPY[NEW_LEN++] = TEXT[i++]; \ ++ continue; \ ++ } \ ++ \ ++ if (ignore) \ ++ { \ ++ if ((ignore == nonprinting && !iswprint (WC)) \ ++ || (ignore == nondictionary \ ++ && !iswalnum (WC) && !iswblank (WC))) \ ++ { \ ++ i += MBLENGTH; \ ++ continue; \ ++ } \ ++ } \ ++ \ ++ if (translate) \ ++ { \ ++ \ ++ uwc = towupper(WC); \ ++ if (WC == uwc) \ ++ { \ ++ memcpy (mbc, TEXT + i, MBLENGTH); \ ++ i += MBLENGTH; \ ++ } \ ++ else \ ++ { \ ++ i += MBLENGTH; \ ++ WC = uwc; \ ++ memset (&state_wc, '\0', sizeof (mbstate_t)); \ ++ \ ++ MBLENGTH = wcrtomb (mbc, WC, &state_wc); \ ++ assert (MBLENGTH != (size_t)-1 && MBLENGTH != 0); \ ++ } \ ++ \ ++ for (j = 0; j < MBLENGTH; j++) \ ++ COPY[NEW_LEN++] = mbc[j]; \ ++ } \ ++ else \ ++ for (j = 0; j < MBLENGTH; j++) \ ++ COPY[NEW_LEN++] = TEXT[i++]; \ ++ } \ ++ COPY[NEW_LEN] = '\0'; \ ++ } \ ++ while (0) ++ IGNORE_CHARS (new_len_a, lena, texta, copy_a, ++ wc_a, mblength_a, state_a); ++ IGNORE_CHARS (new_len_b, lenb, textb, copy_b, ++ wc_b, mblength_b, state_b); ++ diff = xmemcoll (copy_a, new_len_a, copy_b, new_len_b); ++ } ++ else if (lena == 0) ++ diff = - NONZERO (lenb); ++ else if (lenb == 0) ++ goto greater; ++ else ++ diff = xmemcoll (texta, lena, textb, lenb); ++ } ++ ++ if (diff) ++ goto not_equal; ++ ++ key = key->next; ++ if (! key) ++ break; ++ ++ /* Find the beginning and limit of the next field. */ ++ if (key->eword != -1) ++ lima = limfield (a, key), limb = limfield (b, key); ++ else ++ lima = a->text + a->length - 1, limb = b->text + b->length - 1; ++ ++ if (key->sword != -1) ++ texta = begfield (a, key), textb = begfield (b, key); ++ else ++ { ++ texta = a->text, textb = b->text; ++ if (key->skipsblanks) ++ { ++ while (texta < lima && ismbblank (texta, lima - texta, &mblength_a)) ++ texta += mblength_a; ++ while (textb < limb && ismbblank (textb, limb - textb, &mblength_b)) ++ textb += mblength_b; ++ } ++ } + } + -+ /* Define how ISTREAM is being folded. */ ++ return 0; ++ ++greater: ++ diff = 1; ++not_equal: ++ return key->reverse ? -diff : diff; ++} ++#endif ++ + /* Compare two lines A and B, returning negative, zero, or positive + depending on whether A compares less than, equal to, or greater than B. */ + +@@ -2127,7 +2672,7 @@ + atexit (close_stdout); + + hard_LC_COLLATE = hard_locale (LC_COLLATE); +-#if HAVE_NL_LANGINFO ++#if HAVE_LANGINFO_CODESET + hard_LC_TIME = hard_locale (LC_TIME); + #endif + +@@ -2148,6 +2693,27 @@ + thousands_sep = -1; + } + +#if HAVE_MBRTOWC + if (MB_CUR_MAX > 1) -+ fold_multibyte_text (istream, width, &saved_errno); ++ { ++ inittables = inittables_mb; ++ begfield = begfield_mb; ++ limfield = limfield_mb; ++ getmonth = getmonth_mb; ++ keycompare = keycompare_mb; ++ numcompare = numcompare_mb; ++ } + else +#endif -+ fold_text (istream, width, &saved_errno); ++ { ++ inittables = inittables_uni; ++ begfield = begfield_uni; ++ limfield = limfield_uni; ++ getmonth = getmonth_uni; ++ keycompare = keycompare_uni; ++ numcompare = numcompare_uni; ++ } + - if (ferror (istream)) - { - error (0, saved_errno, "%s", filename); -@@ -255,7 +501,8 @@ - - atexit (close_stdout); + have_read_stdin = false; + inittables (); -- break_spaces = count_bytes = have_read_stdin = false; -+ operating_mode = column_mode; -+ break_spaces = have_read_stdin = false; +@@ -2349,13 +2915,35 @@ - while ((optc = getopt_long (argc, argv, shortopts, longopts, NULL)) != -1) - { -@@ -264,7 +511,15 @@ - switch (optc) - { - case 'b': /* Count bytes rather than columns. */ -- count_bytes = true; -+ if (operating_mode != column_mode) -+ FATAL_ERROR (_("only one way of folding may be specified")); -+ operating_mode = byte_mode; -+ break; + case 't': + { +- char newtab = optarg[0]; +- if (! newtab) ++ char newtab[MB_LEN_MAX + 1]; ++ size_t newtab_length = 1; ++ strncpy (newtab, optarg, MB_LEN_MAX); ++ if (! newtab[0]) + error (SORT_FAILURE, 0, _("empty tab")); +- if (optarg[1]) ++#if HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1) ++ { ++ wchar_t wc; ++ mbstate_t state; ++ size_t i; + -+ case 'c': -+ if (operating_mode != column_mode) -+ FATAL_ERROR (_("only one way of folding may be specified")); -+ operating_mode = character_mode; ++ memset (&state, '\0', sizeof (mbstate_t)); ++ newtab_length = mbrtowc (&wc, newtab, strnlen (newtab, ++ MB_LEN_MAX), ++ &state); ++ switch (newtab_length) ++ { ++ case (size_t) -1: ++ case (size_t) -2: ++ case 0: ++ newtab_length = 1; ++ } ++ } ++#endif ++ if (newtab_length == 1 && optarg[1]) + { + if (STREQ (optarg, "\\0")) +- newtab = '\0'; ++ newtab[0] = '\0'; + else + { + /* Provoke with `sort -txx'. Complain about +@@ -2366,9 +2954,12 @@ + quote (optarg)); + } + } +- if (tab != TAB_DEFAULT && tab != newtab) ++ if (tab_length ++ && (tab_length != newtab_length ++ || memcmp (tab, newtab, tab_length) != 0)) + error (SORT_FAILURE, 0, _("incompatible tabs")); +- tab = newtab; ++ memcpy (tab, newtab, newtab_length); ++ tab_length = newtab_length; + } break; - case 's': /* Break at word boundaries. */ ---- coreutils-5.93/src/join.c.i18n 2005-08-12 08:16:25.000000000 +0100 -+++ coreutils-5.93/src/join.c 2005-12-23 08:53:01.000000000 +0000 -@@ -23,16 +23,30 @@ - #include +--- coreutils-5.95/src/uniq.c.i18n 2005-07-05 07:32:54.000000000 +0100 ++++ coreutils-5.95/src/uniq.c 2006-05-15 15:08:57.000000000 +0100 +@@ -23,6 +23,16 @@ #include + #include -+/* Get mbstate_t, mbrtowc(), mbrtowc(), wcwidth(). */ ++/* Get mbstate_t, mbrtowc(). */ +#if HAVE_WCHAR_H +# include +#endif + -+/* Get iswblank(), towupper. */ ++/* Get isw* functions. */ +#if HAVE_WCTYPE_H +# include +#endif + #include "system.h" - #include "error.h" - #include "hard-locale.h" + #include "argmatch.h" #include "linebuffer.h" --#include "memcasecmp.h" +@@ -32,7 +42,19 @@ #include "quote.h" - #include "stdio--.h" #include "xmemcoll.h" #include "xstrtol.h" - -+/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ -+#if HAVE_MBRTOWC && defined mbstate_t -+# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) -+#endif -+ - /* The official name of this program (e.g., no `g' prefix). */ - #define PROGRAM_NAME "join" - -@@ -104,10 +118,12 @@ - /* Last element in `outlist', where a new element can be added. */ - static struct outlist *outlist_end = &outlist_head; - --/* Tab character separating fields. If negative, fields are separated -- by any nonempty string of blanks, otherwise by exactly one -- tab character whose value (when cast to unsigned char) equals TAB. */ --static int tab = -1; -+/* Tab character separating fields. If NULL, fields are separated -+ by any nonempty string of blanks. */ -+static char *tab = NULL; -+ -+/* The number of bytes used for tab. */ -+static size_t tablen = 0; - - static struct option const longopts[] = - { -@@ -197,6 +213,8 @@ - - /* Fill in the `fields' structure in LINE. */ - -+/* Fill in the `fields' structure in LINE. */ -+ - static void - xfields (struct line *line) - { -@@ -206,10 +224,11 @@ - if (ptr == lim) - return; - -- if (0 <= tab) -+ if (tab != NULL) - { -+ unsigned char t = tab[0]; - char *sep; -- for (; (sep = memchr (ptr, tab, lim - ptr)) != NULL; ptr = sep + 1) -+ for (; (sep = memchr (ptr, t, lim - ptr)) != NULL; ptr = sep + 1) - extract_field (line, ptr, sep - ptr); - } - else -@@ -236,6 +255,148 @@ - extract_field (line, ptr, lim - ptr); - } - -+#if HAVE_MBRTOWC -+static void -+xfields_multibyte (struct line *line) -+{ -+ char *ptr = line->buf.buffer; -+ char const *lim = ptr + line->buf.length - 1; -+ wchar_t wc = 0; -+ size_t mblength = 1; -+ mbstate_t state, state_bak; -+ -+ memset (&state, 0, sizeof (mbstate_t)); -+ -+ if (ptr == lim) -+ return; -+ -+ if (tab != NULL) -+ { -+ unsigned char t = tab[0]; -+ char *sep = ptr; -+ for (; ptr < lim; ptr = sep + mblength) -+ { -+ sep = ptr; -+ while (sep < lim) -+ { -+ state_bak = state; -+ mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); -+ -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) -+ { -+ mblength = 1; -+ state = state_bak; -+ } -+ mblength = (mblength < 1) ? 1 : mblength; -+ -+ if (mblength == tablen && !memcmp (sep, tab, mblength)) -+ break; -+ else -+ { -+ sep += mblength; -+ continue; -+ } -+ } -+ -+ if (sep == lim) -+ break; -+ -+ extract_field (line, ptr, sep - ptr); -+ } -+ } -+ else -+ { -+ /* Skip leading blanks before the first field. */ -+ while(ptr < lim) -+ { -+ state_bak = state; -+ mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); +-#include "memcasecmp.h" ++#include "xmemcoll.h" + -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) -+ { -+ mblength = 1; -+ state = state_bak; -+ break; -+ } -+ mblength = (mblength < 1) ? 1 : mblength; ++/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC ++ installation; work around this configuration error. */ ++#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 ++# define MB_LEN_MAX 16 ++#endif + -+ if (!iswblank(wc)) -+ break; -+ ptr += mblength; -+ } ++/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ ++#if HAVE_MBRTOWC && defined mbstate_t ++# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) ++#endif + -+ do -+ { -+ char *sep; -+ state_bak = state; -+ mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) -+ { -+ mblength = 1; -+ state = state_bak; -+ break; -+ } -+ mblength = (mblength < 1) ? 1 : mblength; + + /* The official name of this program (e.g., no `g' prefix). */ + #define PROGRAM_NAME "uniq" +@@ -109,6 +131,10 @@ + /* Select whether/how to delimit groups of duplicate lines. */ + static enum delimit_method delimit_groups; + ++/* Function pointers. */ ++static char * ++(*find_field) (struct linebuffer *line); + -+ sep = ptr + mblength; -+ while (sep != lim) -+ { -+ state_bak = state; -+ mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) -+ { -+ mblength = 1; -+ state = state_bak; -+ break; -+ } -+ mblength = (mblength < 1) ? 1 : mblength; + static struct option const longopts[] = + { + {"count", no_argument, NULL, 'c'}, +@@ -189,7 +215,7 @@ + return a pointer to the beginning of the line's field to be compared. */ + + static char * +-find_field (const struct linebuffer *line) ++find_field_uni (struct linebuffer *line) + { + size_t count; + char *lp = line->buffer; +@@ -210,6 +236,83 @@ + return lp + i; + } + ++#if HAVE_MBRTOWC + -+ if (iswblank (wc)) -+ break; ++# define MBCHAR_TO_WCHAR(WC, MBLENGTH, LP, POS, SIZE, STATEP, CONVFAIL) \ ++ do \ ++ { \ ++ mbstate_t state_bak; \ ++ \ ++ CONVFAIL = 0; \ ++ state_bak = *STATEP; \ ++ \ ++ MBLENGTH = mbrtowc (&WC, LP + POS, SIZE - POS, STATEP); \ ++ \ ++ switch (MBLENGTH) \ ++ { \ ++ case (size_t)-2: \ ++ case (size_t)-1: \ ++ *STATEP = state_bak; \ ++ CONVFAIL++; \ ++ /* Fall through */ \ ++ case 0: \ ++ MBLENGTH = 1; \ ++ } \ ++ } \ ++ while (0) + -+ sep += mblength; -+ } ++static char * ++find_field_multi (struct linebuffer *line) ++{ ++ size_t count; ++ char *lp = line->buffer; ++ size_t size = line->length - 1; ++ size_t pos; ++ size_t mblength; ++ wchar_t wc; ++ mbstate_t *statep; ++ int convfail; + -+ extract_field (line, ptr, sep - ptr); -+ if (sep == lim) -+ return; ++ pos = 0; ++ statep = &(line->state); + -+ state_bak = state; -+ mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ /* skip fields. */ ++ for (count = 0; count < skip_fields && pos < size; count++) ++ { ++ while (pos < size) ++ { ++ MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); ++ ++ if (convfail || !iswblank (wc)) + { -+ mblength = 1; -+ state = state_bak; ++ pos += mblength; + break; + } -+ mblength = (mblength < 1) ? 1 : mblength; ++ pos += mblength; ++ } + -+ ptr = sep + mblength; -+ while (ptr != lim) -+ { -+ state_bak = state; -+ mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) -+ { -+ mblength = 1; -+ state = state_bak; -+ break; -+ } -+ mblength = (mblength < 1) ? 1 : mblength; ++ while (pos < size) ++ { ++ MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); + -+ if (!iswblank (wc)) -+ break; ++ if (!convfail && iswblank (wc)) ++ break; + -+ ptr += mblength; -+ } ++ pos += mblength; + } -+ while (ptr != lim); + } + -+ extract_field (line, ptr, lim - ptr); ++ /* skip fields. */ ++ for (count = 0; count < skip_chars && pos < size; count++) ++ { ++ MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); ++ pos += mblength; ++ } ++ ++ return lp + pos; +} +#endif + - /* Read a line from FP into LINE and split it into fields. - Return true if successful. */ - -@@ -256,6 +417,11 @@ - line->nfields_allocated = 0; - line->nfields = 0; - line->fields = NULL; -+#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ xfields_multibyte (line); -+ else -+#endif - xfields (line); - return true; - } -@@ -310,56 +476,114 @@ - keycmp (struct line const *line1, struct line const *line2) + /* Return false if two strings OLD and NEW match, true if not. + OLD and NEW point not to the beginnings of the lines + but rather to the beginnings of the fields to compare. +@@ -218,6 +321,8 @@ + static bool + different (char *old, char *new, size_t oldlen, size_t newlen) { - /* Start of field to compare in each file. */ -- char *beg1; -- char *beg2; -- -- size_t len1; -- size_t len2; /* Length of fields to compare. */ -+ char *beg[2]; -+ char *copy[2]; -+ size_t len[2]; /* Length of fields to compare. */ - int diff; -+ int i, j; - - if (join_field_1 < line1->nfields) - { -- beg1 = line1->fields[join_field_1].beg; -- len1 = line1->fields[join_field_1].len; -+ beg[0] = line1->fields[join_field_1].beg; -+ len[0] = line1->fields[join_field_1].len; - } - else - { -- beg1 = NULL; -- len1 = 0; -+ beg[0] = NULL; -+ len[0] = 0; - } - - if (join_field_2 < line2->nfields) - { -- beg2 = line2->fields[join_field_2].beg; -- len2 = line2->fields[join_field_2].len; -+ beg[1] = line2->fields[join_field_2].beg; -+ len[1] = line2->fields[join_field_2].len; - } - else - { -- beg2 = NULL; -- len2 = 0; -+ beg[1] = NULL; -+ len[1] = 0; - } - -- if (len1 == 0) -- return len2 == 0 ? 0 : -1; -- if (len2 == 0) -+ if (len[0] == 0) -+ return len[1] == 0 ? 0 : -1; -+ if (len[1] == 0) - return 1; ++ char *copy_old, *copy_new; ++ + if (check_chars < oldlen) + oldlen = check_chars; + if (check_chars < newlen) +@@ -225,14 +330,92 @@ if (ignore_case) { -- /* FIXME: ignore_case does not work with NLS (in particular, -- with multibyte chars). */ -- diff = memcasecmp (beg1, beg2, MIN (len1, len2)); -+#ifdef HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ { -+ size_t mblength; -+ wchar_t wc, uwc; -+ mbstate_t state, state_bak; +- /* FIXME: This should invoke strcoll somehow. */ +- return oldlen != newlen || memcasecmp (old, new, oldlen); ++ size_t i; ++ ++ copy_old = alloca (oldlen + 1); ++ copy_new = alloca (oldlen + 1); ++ ++ for (i = 0; i < oldlen; i++) ++ { ++ copy_old[i] = toupper (old[i]); ++ copy_new[i] = toupper (new[i]); ++ } + } +- else if (hard_LC_COLLATE) +- return xmemcoll (old, oldlen, new, newlen) != 0; + else +- return oldlen != newlen || memcmp (old, new, oldlen); ++ { ++ copy_old = (char *)old; ++ copy_new = (char *)new; ++ } + -+ memset (&state, '\0', sizeof (mbstate_t)); ++ return xmemcoll (copy_old, oldlen, copy_new, newlen); ++} + -+ for (i = 0; i < 2; i++) -+ { -+ copy[i] = alloca (len[i] + 1); ++#if HAVE_MBRTOWC ++static int ++different_multi (const char *old, const char *new, size_t oldlen, size_t newlen, mbstate_t oldstate, mbstate_t newstate) ++{ ++ size_t i, j, chars; ++ const char *str[2]; ++ char *copy[2]; ++ size_t len[2]; ++ mbstate_t state[2]; ++ size_t mblength; ++ wchar_t wc, uwc; ++ mbstate_t state_bak; + -+ for (j = 0; j < MIN (len[0], len[1]);) -+ { -+ state_bak = state; -+ mblength = mbrtowc (&wc, beg[i] + j, len[i] - j, &state); ++ str[0] = old; ++ str[1] = new; ++ len[0] = oldlen; ++ len[1] = newlen; ++ state[0] = oldstate; ++ state[1] = newstate; + -+ switch (mblength) -+ { -+ case (size_t) -1: -+ case (size_t) -2: -+ state = state_bak; -+ /* Fall through */ -+ case 0: -+ mblength = 1; -+ break; ++ for (i = 0; i < 2; i++) ++ { ++ copy[i] = alloca (len[i] + 1); + -+ default: -+ uwc = towupper (wc); ++ for (j = 0, chars = 0; j < len[i] && chars < check_chars; chars++) ++ { ++ state_bak = state[i]; ++ mblength = mbrtowc (&wc, str[i] + j, len[i] - j, &(state[i])); + -+ if (uwc != wc) -+ { -+ mbstate_t state_wc; ++ switch (mblength) ++ { ++ case (size_t)-1: ++ case (size_t)-2: ++ state[i] = state_bak; ++ /* Fall through */ ++ case 0: ++ mblength = 1; ++ break; + -+ memset (&state_wc, '\0', sizeof (mbstate_t)); -+ wcrtomb (copy[i] + j, uwc, &state_wc); -+ } -+ else -+ memcpy (copy[i] + j, beg[i] + j, mblength); -+ } -+ j += mblength; -+ } -+ copy[i][j] = '\0'; -+ } -+ } -+ else -+#endif -+ { -+ for (i = 0; i < 2; i++) -+ { -+ copy[i] = alloca (len[i] + 1); ++ default: ++ if (ignore_case) ++ { ++ uwc = towupper (wc); + -+ for (j = 0; j < MIN (len[0], len[1]); j++) -+ copy[i][j] = toupper (beg[i][j]); ++ if (uwc != wc) ++ { ++ mbstate_t state_wc; + -+ copy[i][j] = '\0'; -+ } -+ } - } - else - { -- if (hard_LC_COLLATE) -- return xmemcoll (beg1, len1, beg2, len2); -- diff = memcmp (beg1, beg2, MIN (len1, len2)); -+ copy[0] = (unsigned char *) beg[0]; -+ copy[1] = (unsigned char *) beg[1]; - } - -+ if (HAVE_SETLOCALE && hard_LC_COLLATE) -+ return xmemcoll ((char *) copy[0], len[0], (char *) copy[1], len[1]); -+ diff = memcmp (copy[0], copy[1], MIN (len[0], len[1])); ++ memset (&state_wc, '\0', sizeof(mbstate_t)); ++ wcrtomb (copy[i] + j, uwc, &state_wc); ++ } ++ else ++ memcpy (copy[i] + j, str[i] + j, mblength); ++ } ++ else ++ memcpy (copy[i] + j, str[i] + j, mblength); ++ } ++ j += mblength; ++ } ++ copy[i][j] = '\0'; ++ len[i] = j; ++ } + - if (diff) - return diff; -- return len1 < len2 ? -1 : len1 != len2; -+ return len[0] - len[1]; ++ return xmemcoll (copy[0], len[0], copy[1], len[1]); } ++#endif - /* Print field N of LINE if it exists and is nonempty, otherwise -@@ -384,11 +608,18 @@ - - /* Print the join of LINE1 and LINE2. */ - -+#define PUT_TAB_CHAR \ -+ do \ -+ { \ -+ (tab != NULL) ? \ -+ fwrite(tab, sizeof(char), tablen, stdout) : putchar (' '); \ -+ } \ -+ while (0) + /* Output the line in linebuffer LINE to standard output + provided that the switches say it should be output. +@@ -286,15 +469,43 @@ + { + char *prevfield IF_LINT (= NULL); + size_t prevlen IF_LINT (= 0); ++#if HAVE_MBRTOWC ++ mbstate_t prevstate; + - static void - prjoin (struct line const *line1, struct line const *line2) - { - const struct outlist *outlist; -- char output_separator = tab < 0 ? ' ' : tab; ++ memset (&prevstate, '\0', sizeof (mbstate_t)); ++#endif - outlist = outlist_head.next; - if (outlist) -@@ -404,12 +635,12 @@ - if (o->file == 0) - { - if (line1 == &uni_blank) -- { -+ { - line = line2; - field = join_field_2; - } - else -- { -+ { - line = line1; - field = join_field_1; - } -@@ -423,7 +654,7 @@ - o = o->next; - if (o == NULL) - break; -- putchar (output_separator); -+ PUT_TAB_CHAR; - } - putchar ('\n'); - } -@@ -441,23 +672,23 @@ - prfield (join_field_1, line1); - for (i = 0; i < join_field_1 && i < line1->nfields; ++i) - { -- putchar (output_separator); -+ PUT_TAB_CHAR; - prfield (i, line1); - } - for (i = join_field_1 + 1; i < line1->nfields; ++i) + while (!feof (stdin)) { -- putchar (output_separator); -+ PUT_TAB_CHAR; - prfield (i, line1); - } + char *thisfield; + size_t thislen; ++#if HAVE_MBRTOWC ++ mbstate_t thisstate; ++#endif ++ + if (readlinebuffer (thisline, stdin) == 0) + break; + thisfield = find_field (thisline); + thislen = thisline->length - 1 - (thisfield - thisline->buffer); ++#if HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1) ++ { ++ thisstate = thisline->state; ++ ++ if (prevline->length == 0 || different_multi ++ (thisfield, prevfield, thislen, prevlen, thisstate, prevstate)) ++ { ++ fwrite (thisline->buffer, sizeof (char), ++ thisline->length, stdout); ++ ++ SWAP_LINES (prevline, thisline); ++ prevfield = thisfield; ++ prevlen = thislen; ++ prevstate = thisstate; ++ } ++ } ++ else ++#endif + if (prevline->length == 0 + || different (thisfield, prevfield, thislen, prevlen)) + { +@@ -313,17 +524,26 @@ + size_t prevlen; + uintmax_t match_count = 0; + bool first_delimiter = true; ++#if HAVE_MBRTOWC ++ mbstate_t prevstate; ++#endif - for (i = 0; i < join_field_2 && i < line2->nfields; ++i) - { -- putchar (output_separator); -+ PUT_TAB_CHAR; - prfield (i, line2); - } - for (i = join_field_2 + 1; i < line2->nfields; ++i) - { -- putchar (output_separator); -+ PUT_TAB_CHAR; - prfield (i, line2); - } - putchar ('\n'); -@@ -869,20 +1100,41 @@ + if (readlinebuffer (prevline, stdin) == 0) + goto closefiles; + prevfield = find_field (prevline); + prevlen = prevline->length - 1 - (prevfield - prevline->buffer); ++#if HAVE_MBRTOWC ++ prevstate = prevline->state; ++#endif + + while (!feof (stdin)) + { + bool match; + char *thisfield; + size_t thislen; ++#if HAVE_MBRTOWC ++ mbstate_t thisstate; ++#endif + if (readlinebuffer (thisline, stdin) == 0) + { + if (ferror (stdin)) +@@ -332,6 +552,15 @@ + } + thisfield = find_field (thisline); + thislen = thisline->length - 1 - (thisfield - thisline->buffer); ++#if HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1) ++ { ++ thisstate = thisline->state; ++ match = !different_multi (thisfield, prevfield, ++ thislen, prevlen, thisstate, prevstate); ++ } ++ else ++#endif + match = !different (thisfield, prevfield, thislen, prevlen); + match_count += match; - case 't': - { -- unsigned char newtab = optarg[0]; -- if (! newtab) -+ char *newtab; -+ size_t newtablen; -+ if (! optarg[0]) - error (EXIT_FAILURE, 0, _("empty tab")); -- if (optarg[1]) -+ newtab = xstrdup (optarg); +@@ -364,6 +593,9 @@ + SWAP_LINES (prevline, thisline); + prevfield = thisfield; + prevlen = thislen; +#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ { -+ mbstate_t state; -+ -+ memset (&state, 0, sizeof (mbstate_t)); -+ newtablen = mbrtowc (NULL, newtab, -+ strnlen (newtab, MB_LEN_MAX), -+ &state); -+ if (newtablen == (size_t) 0 -+ || newtablen == (size_t) -1 -+ || newtablen == (size_t) -2) -+ newtablen = 1; -+ } -+ else ++ prevstate = thisstate; +#endif -+ newtablen = 1; -+ -+ if (newtablen == 1 && newtab[1]) -+ { -+ if (STREQ (newtab, "\\0")) -+ newtab[0] = '\0'; -+ } -+ if (tab != NULL && strcmp (tab, newtab)) - { -- if (STREQ (optarg, "\\0")) -- newtab = '\0'; -- else -- error (EXIT_FAILURE, 0, _("multi-character tab %s"), -- quote (optarg)); -+ free (newtab); -+ error (EXIT_FAILURE, 0, _("incompatible tabs")); - } -- if (0 <= tab && tab != newtab) -- error (EXIT_FAILURE, 0, _("incompatible tabs")); - tab = newtab; -+ tablen = newtablen; - } - break; + if (!match) + match_count = 0; + } +@@ -408,6 +640,19 @@ + + atexit (close_stdout); ---- coreutils-5.93/src/unexpand.c.i18n 2005-08-12 08:16:25.000000000 +0100 -+++ coreutils-5.93/src/unexpand.c 2005-12-23 08:53:01.000000000 +0000 ++#if HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1) ++ { ++ find_field = find_field_multi; ++ } ++ else ++#endif ++ { ++ find_field = find_field_uni; ++ } ++ ++ ++ + skip_chars = 0; + skip_fields = 0; + check_chars = SIZE_MAX; +--- coreutils-5.95/src/unexpand.c.i18n 2005-08-12 08:16:25.000000000 +0100 ++++ coreutils-5.95/src/unexpand.c 2006-05-15 15:08:57.000000000 +0100 @@ -39,11 +39,28 @@ #include #include @@ -3072,846 +2884,1054 @@ if (have_read_stdin && fclose (stdin) != 0) error (EXIT_FAILURE, errno, "-"); ---- coreutils-5.93/src/sort.c.i18n 2005-10-07 19:48:28.000000000 +0100 -+++ coreutils-5.93/src/sort.c 2005-12-23 10:38:44.000000000 +0000 -@@ -23,9 +23,18 @@ - - #include - -+#include +--- coreutils-5.95/src/fold.c.i18n 2005-08-12 08:29:38.000000000 +0100 ++++ coreutils-5.95/src/fold.c 2006-05-15 15:08:57.000000000 +0100 +@@ -23,11 +23,33 @@ #include #include - #include + ++/* Get mbstate_t, mbrtowc(), wcwidth(). */ +#if HAVE_WCHAR_H +# include +#endif -+/* Get isw* functions. */ ++ ++/* Get iswprint(), iswblank(), wcwidth(). */ +#if HAVE_WCTYPE_H +# include +#endif + #include "system.h" #include "error.h" - #include "hard-locale.h" -@@ -95,14 +104,38 @@ - /* Thousands separator; if -1, then there isn't one. */ - static int thousands_sep; + #include "quote.h" + #include "xstrtol.h" -+static int force_general_numcompare = 0; ++/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC ++ installation; work around this configuration error. */ ++#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 ++# undef MB_LEN_MAX ++# define MB_LEN_MAX 16 ++#endif + - /* Nonzero if the corresponding locales are hard. */ - static bool hard_LC_COLLATE; --#if HAVE_NL_LANGINFO -+#if HAVE_LANGINFO_CODESET - static bool hard_LC_TIME; - #endif ++/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ ++#if HAVE_MBRTOWC && defined mbstate_t ++# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) ++#endif ++ + #define TAB_WIDTH 8 - #define NONZERO(x) ((x) != 0) + /* The official name of this program (e.g., no `g' prefix). */ +@@ -35,23 +57,44 @@ -+/* get a multibyte character's byte length. */ -+#define GET_BYTELEN_OF_CHAR(LIM, PTR, MBLENGTH, STATE) \ -+ do \ -+ { \ -+ wchar_t wc; \ -+ mbstate_t state_bak; \ -+ \ -+ state_bak = STATE; \ -+ mblength = mbrtowc (&wc, PTR, LIM - PTR, &STATE); \ -+ \ -+ switch (MBLENGTH) \ -+ { \ -+ case (size_t)-1: \ -+ case (size_t)-2: \ -+ STATE = state_bak; \ -+ /* Fall through. */ \ -+ case 0: \ -+ MBLENGTH = 1; \ -+ } \ -+ } \ + #define AUTHORS "David MacKenzie" + ++#define FATAL_ERROR(Message) \ ++ do \ ++ { \ ++ error (0, 0, (Message)); \ ++ usage (2); \ ++ } \ + while (0) + - /* The kind of blanks for '-b' to skip in various options. */ - enum blanktype { bl_start, bl_end, bl_both }; ++enum operating_mode ++{ ++ /* Fold texts by columns that are at the given positions. */ ++ column_mode, ++ ++ /* Fold texts by bytes that are at the given positions. */ ++ byte_mode, ++ ++ /* Fold texts by characters that are at the given positions. */ ++ character_mode, ++}; ++ + /* The name this program was run with. */ + char *program_name; -@@ -239,13 +272,11 @@ - they were read if all keys compare equal. */ - static bool stable; ++/* The argument shows current mode. (Default: column_mode) */ ++static enum operating_mode operating_mode; ++ + /* If nonzero, try to break on whitespace. */ + static bool break_spaces; --/* If TAB has this value, blanks separate fields. */ --enum { TAB_DEFAULT = CHAR_MAX + 1 }; +-/* If nonzero, count bytes, not column positions. */ +-static bool count_bytes; +- + /* If nonzero, at least one of the files we read was standard input. */ + static bool have_read_stdin; + +-static char const shortopts[] = "bsw:0::1::2::3::4::5::6::7::8::9::"; ++static char const shortopts[] = "bcsw:0::1::2::3::4::5::6::7::8::9::"; + + static struct option const longopts[] = + { + {"bytes", no_argument, NULL, 'b'}, ++ {"characters", no_argument, NULL, 'c'}, + {"spaces", no_argument, NULL, 's'}, + {"width", required_argument, NULL, 'w'}, + {GETOPT_HELP_OPTION_DECL}, +@@ -81,6 +124,7 @@ + "), stdout); + fputs (_("\ + -b, --bytes count bytes rather than columns\n\ ++ -c, --characters count characters rather than columns\n\ + -s, --spaces break at spaces\n\ + -w, --width=WIDTH use WIDTH columns instead of 80\n\ + "), stdout); +@@ -98,7 +142,7 @@ + static size_t + adjust_column (size_t column, char c) + { +- if (!count_bytes) ++ if (operating_mode != byte_mode) + { + if (c == '\b') + { +@@ -117,35 +161,14 @@ + return column; + } + +-/* Fold file FILENAME, or standard input if FILENAME is "-", +- to stdout, with maximum line length WIDTH. +- Return true if successful. */ +- +-static bool +-fold_file (char *filename, size_t width) ++static void ++fold_text (FILE *istream, size_t width, int *saved_errno) + { +- FILE *istream; + int c; + size_t column = 0; /* Screen column where next char will go. */ + size_t offset_out = 0; /* Index in `line_out' for next char. */ + static char *line_out = NULL; + static size_t allocated_out = 0; +- int saved_errno; +- +- if (STREQ (filename, "-")) +- { +- istream = stdin; +- have_read_stdin = true; +- } +- else +- istream = fopen (filename, "r"); +- +- if (istream == NULL) +- { +- error (0, errno, "%s", filename); +- return false; +- } - --/* Tab character separating fields. If TAB_DEFAULT, then fields are -+/* Tab character separating fields. If tab_length is 0, then fields are - separated by the empty string between a non-blank character and a blank - character. */ --static int tab = TAB_DEFAULT; -+static char tab[MB_LEN_MAX + 1]; -+static size_t tab_length = 0; + while ((c = getc (istream)) != EOF) + { + if (offset_out + 1 >= allocated_out) +@@ -172,6 +195,15 @@ + bool found_blank = false; + size_t logical_end = offset_out; - /* Flag to remove consecutive duplicate lines from the output. - Only the last of a sequence of equal lines will be output. */ -@@ -392,6 +423,44 @@ - static struct tempnode *volatile temphead; - static struct tempnode *volatile *temptail = &temphead; ++ /* If LINE_OUT has no wide character, ++ put a new wide character in LINE_OUT ++ if column is bigger than width. */ ++ if (offset_out == 0) ++ { ++ line_out[offset_out++] = c; ++ continue; ++ } ++ + /* Look for the last blank. */ + while (logical_end) + { +@@ -218,11 +250,225 @@ + line_out[offset_out++] = c; + } -+/* Function pointers. */ -+static void -+(*inittables) (void); -+static char * -+(*begfield) (const struct line*, const struct keyfield *); -+static char * -+(*limfield) (const struct line*, const struct keyfield *); -+static int -+(*getmonth) (char const *, size_t); -+static int -+(*keycompare) (const struct line *, const struct line *); -+static int -+(*numcompare) (const char *, const char *); +- saved_errno = errno; ++ *saved_errno = errno; ++ ++ if (offset_out) ++ fwrite (line_out, sizeof (char), (size_t) offset_out, stdout); ++ ++ free(line_out); ++} + -+/* Test for white space multibyte character. -+ Set LENGTH the byte length of investigated multibyte character. */ +#if HAVE_MBRTOWC -+static int -+ismbblank (const char *str, size_t len, size_t *length) ++static void ++fold_multibyte_text (FILE *istream, int width, int *saved_errno) +{ -+ size_t mblength; -+ wchar_t wc; -+ mbstate_t state; ++ char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ ++ size_t buflen = 0; /* The length of the byte sequence in buf. */ ++ char *bufpos; /* Next read position of BUF. */ ++ wint_t wc; /* A gotten wide character. */ ++ size_t mblength; /* The byte size of a multibyte character which shows ++ as same character as WC. */ ++ mbstate_t state, state_bak; /* State of the stream. */ ++ int convfail; /* 1, when conversion is failed. Otherwise 0. */ ++ ++ char *line_out = NULL; ++ size_t offset_out = 0; /* Index in `line_out' for next char. */ ++ size_t allocated_out = 0; ++ ++ int increment; ++ size_t column = 0; ++ ++ size_t last_blank_pos; ++ size_t last_blank_column; ++ int is_blank_seen; ++ int last_blank_increment; ++ int is_bs_following_last_blank; ++ size_t bs_following_last_blank_num; ++ int is_cr_after_last_blank; ++ ++#define CLEAR_FLAGS \ ++ do \ ++ { \ ++ last_blank_pos = 0; \ ++ last_blank_column = 0; \ ++ is_blank_seen = 0; \ ++ is_bs_following_last_blank = 0; \ ++ bs_following_last_blank_num = 0; \ ++ is_cr_after_last_blank = 0; \ ++ } \ ++ while (0) ++ ++#define START_NEW_LINE \ ++ do \ ++ { \ ++ putchar ('\n'); \ ++ column = 0; \ ++ offset_out = 0; \ ++ CLEAR_FLAGS; \ ++ } \ ++ while (0) + ++ CLEAR_FLAGS; + memset (&state, '\0', sizeof(mbstate_t)); -+ mblength = mbrtowc (&wc, str, len, &state); + -+ if (mblength == (size_t)-1 || mblength == (size_t)-2) ++ for (;; bufpos += mblength, buflen -= mblength) + { -+ *length = 1; -+ return 0; -+ } ++ if (buflen < MB_LEN_MAX && !feof (istream) && !ferror (istream)) ++ { ++ memmove (buf, bufpos, buflen); ++ buflen += fread (buf + buflen, sizeof(char), BUFSIZ, istream); ++ bufpos = buf; ++ } + -+ *length = (mblength < 1) ? 1 : mblength; -+ return iswblank (wc); -+} -+#endif ++ if (buflen < 1) ++ break; + - /* Clean up any remaining temporary files. */ - - static void -@@ -545,7 +614,7 @@ - free (node); - } - --#if HAVE_NL_LANGINFO -+#if HAVE_LANGINFO_CODESET - - static int - struct_month_cmp (const void *m1, const void *m2) -@@ -560,7 +629,7 @@ - /* Initialize the character class tables. */ - - static void --inittables (void) -+inittables_uni (void) - { - size_t i; - -@@ -572,7 +641,7 @@ - fold_toupper[i] = (ISLOWER (i) ? toupper (i) : i); - } - --#if HAVE_NL_LANGINFO -+#if HAVE_LANGINFO_CODESET - /* If we're not in the "C" locale, read different names for months. */ - if (hard_LC_TIME) - { -@@ -598,6 +667,64 @@ - #endif - } - -+#if HAVE_MBRTOWC -+static void -+inittables_mb (void) -+{ -+ int i, j, k, l; -+ char *name, *s; -+ size_t s_len, mblength; -+ char mbc[MB_LEN_MAX]; -+ wchar_t wc, pwc; -+ mbstate_t state_mb, state_wc; ++ /* Get a wide character. */ ++ convfail = 0; ++ state_bak = state; ++ mblength = mbrtowc ((wchar_t *)&wc, bufpos, buflen, &state); + -+ for (i = 0; i < MONTHS_PER_YEAR; i++) -+ { -+ s = (char *) nl_langinfo (ABMON_1 + i); -+ s_len = strlen (s); -+ monthtab[i].name = name = (char *) xmalloc (s_len + 1); -+ monthtab[i].val = i + 1; ++ switch (mblength) ++ { ++ case (size_t)-1: ++ case (size_t)-2: ++ convfail++; ++ state = state_bak; ++ /* Fall through. */ + -+ memset (&state_mb, '\0', sizeof (mbstate_t)); -+ memset (&state_wc, '\0', sizeof (mbstate_t)); ++ case 0: ++ mblength = 1; ++ break; ++ } + -+ for (j = 0; j < s_len;) ++rescan: ++ if (operating_mode == byte_mode) /* byte mode */ ++ increment = mblength; ++ else if (operating_mode == character_mode) /* character mode */ ++ increment = 1; ++ else /* column mode */ + { -+ if (!ismbblank (s + j, s_len - j, &mblength)) -+ break; -+ j += mblength; ++ if (convfail) ++ increment = 1; ++ else ++ { ++ switch (wc) ++ { ++ case L'\n': ++ fwrite (line_out, sizeof(char), offset_out, stdout); ++ START_NEW_LINE; ++ continue; ++ ++ case L'\b': ++ increment = (column > 0) ? -1 : 0; ++ break; ++ ++ case L'\r': ++ increment = -1 * column; ++ break; ++ ++ case L'\t': ++ increment = 8 - column % 8; ++ break; ++ ++ default: ++ increment = wcwidth (wc); ++ increment = (increment < 0) ? 0 : increment; ++ } ++ } ++ } ++ ++ if (column + increment > width && break_spaces && last_blank_pos) ++ { ++ fwrite (line_out, sizeof(char), last_blank_pos, stdout); ++ putchar ('\n'); ++ ++ offset_out = offset_out - last_blank_pos; ++ column = column - last_blank_column + ((is_cr_after_last_blank) ++ ? last_blank_increment : bs_following_last_blank_num); ++ memmove (line_out, line_out + last_blank_pos, offset_out); ++ CLEAR_FLAGS; ++ goto rescan; ++ } ++ ++ if (column + increment > width && column != 0) ++ { ++ fwrite (line_out, sizeof(char), offset_out, stdout); ++ START_NEW_LINE; ++ goto rescan; + } + -+ for (k = 0; j < s_len;) ++ if (allocated_out < offset_out + mblength) + { -+ mblength = mbrtowc (&wc, (s + j), (s_len - j), &state_mb); -+ assert (mblength != (size_t)-1 && mblength != (size_t)-2); -+ if (mblength == 0) -+ break; ++ allocated_out += 1024; ++ line_out = xrealloc (line_out, allocated_out); ++ } + -+ pwc = towupper (wc); -+ if (pwc == wc) -+ { -+ memcpy (mbc, s + j, mblength); -+ j += mblength; -+ } -+ else -+ { -+ j += mblength; -+ mblength = wcrtomb (mbc, pwc, &state_wc); -+ assert (mblength != (size_t)0 && mblength != (size_t)-1); -+ } ++ memcpy (line_out + offset_out, bufpos, mblength); ++ offset_out += mblength; ++ column += increment; + -+ for (l = 0; l < mblength; l++) -+ name[k++] = mbc[l]; ++ if (is_blank_seen && !convfail && wc == L'\r') ++ is_cr_after_last_blank = 1; ++ ++ if (is_bs_following_last_blank && !convfail && wc == L'\b') ++ ++bs_following_last_blank_num; ++ else ++ is_bs_following_last_blank = 0; ++ ++ if (break_spaces && !convfail && iswblank (wc)) ++ { ++ last_blank_pos = offset_out; ++ last_blank_column = column; ++ is_blank_seen = 1; ++ last_blank_increment = increment; ++ is_bs_following_last_blank = 1; ++ bs_following_last_blank_num = 0; ++ is_cr_after_last_blank = 0; + } -+ name[k] = '\0'; + } -+ qsort ((void *) monthtab, MONTHS_PER_YEAR, -+ sizeof (struct month), struct_month_cmp); ++ ++ *saved_errno = errno; + + if (offset_out) + fwrite (line_out, sizeof (char), (size_t) offset_out, stdout); + ++ free(line_out); +} +#endif + - /* Specify the amount of main memory to use when sorting. */ - static void - specify_sort_size (char const *s) -@@ -808,7 +935,7 @@ - by KEY in LINE. */ ++/* Fold file FILENAME, or standard input if FILENAME is "-", ++ to stdout, with maximum line length WIDTH. ++ Return 0 if successful, 1 if an error occurs. */ ++ ++static int ++fold_file (char *filename, int width) ++{ ++ FILE *istream; ++ int saved_errno; ++ ++ if (STREQ (filename, "-")) ++ { ++ istream = stdin; ++ have_read_stdin = 1; ++ } ++ else ++ istream = fopen (filename, "r"); ++ ++ if (istream == NULL) ++ { ++ error (0, errno, "%s", filename); ++ return 1; ++ } ++ ++ /* Define how ISTREAM is being folded. */ ++#if HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1) ++ fold_multibyte_text (istream, width, &saved_errno); ++ else ++#endif ++ fold_text (istream, width, &saved_errno); ++ + if (ferror (istream)) + { + error (0, saved_errno, "%s", filename); +@@ -255,7 +501,8 @@ - static char * --begfield (const struct line *line, const struct keyfield *key) -+begfield_uni (const struct line *line, const struct keyfield *key) - { - char *ptr = line->text, *lim = ptr + line->length - 1; - size_t sword = key->sword; -@@ -818,10 +945,10 @@ - /* The leading field separator itself is included in a field when -t - is absent. */ + atexit (close_stdout); -- if (tab != TAB_DEFAULT) -+ if (tab_length) - while (ptr < lim && sword--) - { -- while (ptr < lim && *ptr != tab) -+ while (ptr < lim && *ptr != tab[0]) - ++ptr; - if (ptr < lim) - ++ptr; -@@ -849,11 +976,70 @@ - return ptr; - } +- break_spaces = count_bytes = have_read_stdin = false; ++ operating_mode = column_mode; ++ break_spaces = have_read_stdin = false; -+#if HAVE_MBRTOWC -+static char * -+begfield_mb (const struct line *line, const struct keyfield *key) -+{ -+ int i; -+ char *ptr = line->text, *lim = ptr + line->length - 1; -+ size_t sword = key->sword; -+ size_t schar = key->schar; -+ size_t mblength; -+ mbstate_t state; + while ((optc = getopt_long (argc, argv, shortopts, longopts, NULL)) != -1) + { +@@ -264,7 +511,15 @@ + switch (optc) + { + case 'b': /* Count bytes rather than columns. */ +- count_bytes = true; ++ if (operating_mode != column_mode) ++ FATAL_ERROR (_("only one way of folding may be specified")); ++ operating_mode = byte_mode; ++ break; + -+ memset (&state, '\0', sizeof(mbstate_t)); ++ case 'c': ++ if (operating_mode != column_mode) ++ FATAL_ERROR (_("only one way of folding may be specified")); ++ operating_mode = character_mode; + break; + + case 's': /* Break at word boundaries. */ +--- coreutils-5.95/src/cut.c.i18n 2005-08-12 08:16:25.000000000 +0100 ++++ coreutils-5.95/src/cut.c 2006-05-15 15:08:57.000000000 +0100 +@@ -29,6 +29,11 @@ + #include + #include + #include + -+ if (tab_length) -+ while (ptr < lim && sword--) -+ { -+ while (ptr < lim && memcmp (ptr, tab, tab_length) != 0) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ if (ptr < lim) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ } -+ else -+ while (ptr < lim && sword--) -+ { -+ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) -+ ptr += mblength; -+ if (ptr < lim) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ while (ptr < lim && !ismbblank (ptr, lim - ptr, &mblength)) -+ ptr += mblength; -+ } ++/* Get mbstate_t, mbrtowc(). */ ++#if HAVE_WCHAR_H ++# include ++#endif + #include "system.h" + + #include "error.h" +@@ -37,6 +42,18 @@ + #include "quote.h" + #include "xstrndup.h" + ++/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC ++ installation; work around this configuration error. */ ++#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 ++# undef MB_LEN_MAX ++# define MB_LEN_MAX 16 ++#endif + -+ if (key->skipsblanks) -+ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) -+ ptr += mblength; ++/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ ++#if HAVE_MBRTOWC && defined mbstate_t ++# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) ++#endif ++ + /* The official name of this program (e.g., no `g' prefix). */ + #define PROGRAM_NAME "cut" + +@@ -67,6 +84,52 @@ + } \ + while (0) + ++/* Refill the buffer BUF to get a multibyte character. */ ++#define REFILL_BUFFER(BUF, BUFPOS, BUFLEN, STREAM) \ ++ do \ ++ { \ ++ if (BUFLEN < MB_LEN_MAX && !feof (STREAM) && !ferror (STREAM)) \ ++ { \ ++ memmove (BUF, BUFPOS, BUFLEN); \ ++ BUFLEN += fread (BUF + BUFLEN, sizeof(char), BUFSIZ, STREAM); \ ++ BUFPOS = BUF; \ ++ } \ ++ } \ ++ while (0) ++ ++/* Get wide character on BUFPOS. BUFPOS is not included after that. ++ If byte sequence is not valid as a character, CONVFAIL is 1. Otherwise 0. */ ++#define GET_NEXT_WC_FROM_BUFFER(WC, BUFPOS, BUFLEN, MBLENGTH, STATE, CONVFAIL) \ ++ do \ ++ { \ ++ mbstate_t state_bak; \ ++ \ ++ if (BUFLEN < 1) \ ++ { \ ++ WC = WEOF; \ ++ break; \ ++ } \ ++ \ ++ /* Get a wide character. */ \ ++ CONVFAIL = 0; \ ++ state_bak = STATE; \ ++ MBLENGTH = mbrtowc ((wchar_t *)&WC, BUFPOS, BUFLEN, &STATE); \ ++ \ ++ switch (MBLENGTH) \ ++ { \ ++ case (size_t)-1: \ ++ case (size_t)-2: \ ++ CONVFAIL++; \ ++ STATE = state_bak; \ ++ /* Fall througn. */ \ ++ \ ++ case 0: \ ++ MBLENGTH = 1; \ ++ break; \ ++ } \ ++ } \ ++ while (0) + -+ for (i = 0; i < schar; i++) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); + struct range_pair + { + size_t lo; +@@ -85,7 +148,7 @@ + /* The number of bytes allocated for FIELD_1_BUFFER. */ + static size_t field_1_bufsize; + +-/* The largest field or byte index used as an endpoint of a closed ++/* The largest byte, character or field index used as an endpoint of a closed + or degenerate range specification; this doesn't include the starting + index of right-open-ended ranges. For example, with either range spec + `2-5,9-', `2-3,5,9-' this variable would be set to 5. */ +@@ -97,10 +160,11 @@ + + /* This is a bit vector. + In byte mode, which bytes to output. ++ In character mode, which characters to output. + In field mode, which DELIM-separated fields to output. +- Both bytes and fields are numbered starting with 1, ++ Bytes, characters and fields are numbered starting with 1, + so the zeroth bit of this array is unused. +- A field or byte K has been selected if ++ A byte, character or field K has been selected if + (K <= MAX_RANGE_ENDPOINT and is_printable_field(K)) + || (EOL_RANGE_START > 0 && K >= EOL_RANGE_START). */ + static unsigned char *printable_field; +@@ -109,9 +173,12 @@ + { + undefined_mode, + +- /* Output characters that are in the given bytes. */ ++ /* Output bytes that are at the given positions. */ + byte_mode, + ++ /* Output characters that are at the given positions. */ ++ character_mode, + -+ if (ptr + mblength > lim) -+ break; -+ else -+ ptr += mblength; -+ } + /* Output the given delimeter-separated fields. */ + field_mode + }; +@@ -121,6 +188,13 @@ + + static enum operating_mode operating_mode; + ++/* If nonzero, when in byte mode, don't split multibyte characters. */ ++static int byte_mode_character_aware; + -+ return ptr; -+} -+#endif ++/* If nonzero, the function for single byte locale is work ++ if this program runs on multibyte locale. */ ++static int force_singlebyte_mode; + - /* Return the limit of (a pointer to the first character after) the field - in LINE specified by KEY. */ + /* If true do not output lines containing no delimeter characters. + Otherwise, all such lines are printed. This option is valid only + with field mode. */ +@@ -132,6 +206,9 @@ - static char * --limfield (const struct line *line, const struct keyfield *key) -+limfield_uni (const struct line *line, const struct keyfield *key) - { - char *ptr = line->text, *lim = ptr + line->length - 1; - size_t eword = key->eword, echar = key->echar; -@@ -866,10 +1052,10 @@ - `beginning' is the first character following the delimiting TAB. - Otherwise, leave PTR pointing at the first `blank' character after - the preceding field. */ -- if (tab != TAB_DEFAULT) -+ if (tab_length) - while (ptr < lim && eword--) - { -- while (ptr < lim && *ptr != tab) -+ while (ptr < lim && *ptr != tab[0]) - ++ptr; - if (ptr < lim && (eword | echar)) - ++ptr; -@@ -915,10 +1101,10 @@ - */ + /* The delimeter character for field mode. */ + static unsigned char delim; ++#if HAVE_WCHAR_H ++static wchar_t wcdelim; ++#endif - /* Make LIM point to the end of (one byte past) the current field. */ -- if (tab != TAB_DEFAULT) -+ if (tab_length) - { - char *newlim; -- newlim = memchr (ptr, tab, lim - ptr); -+ newlim = memchr (ptr, tab[0], lim - ptr); - if (newlim) - lim = newlim; + /* True if the --output-delimiter=STRING option was specified. */ + static bool output_delimiter_specified; +@@ -205,7 +282,7 @@ + -f, --fields=LIST select only these fields; also print any line\n\ + that contains no delimiter character, unless\n\ + the -s option is specified\n\ +- -n (ignored)\n\ ++ -n with -b: don't split multibyte characters\n\ + "), stdout); + fputs (_("\ + --complement complement the set of selected bytes, characters\n\ +@@ -360,7 +437,7 @@ + in_digits = false; + /* Starting a range. */ + if (dash_found) +- FATAL_ERROR (_("invalid byte or field list")); ++ FATAL_ERROR (_("invalid byte, character or field list")); + dash_found = true; + fieldstr++; + +@@ -385,14 +462,16 @@ + if (value == 0) + { + /* `n-'. From `initial' to end of line. */ +- eol_range_start = initial; ++ if (eol_range_start == 0 || ++ (eol_range_start != 0 && eol_range_start > initial)) ++ eol_range_start = initial; + field_found = true; + } + else + { + /* `m-n' or `-n' (1-n). */ + if (value < initial) +- FATAL_ERROR (_("invalid byte or field list")); ++ FATAL_ERROR (_("invalid byte, character or field list")); + + /* Is there already a range going to end of line? */ + if (eol_range_start != 0) +@@ -465,6 +544,9 @@ + if (operating_mode == byte_mode) + error (0, 0, + _("byte offset %s is too large"), quote (bad_num)); ++ else if (operating_mode == character_mode) ++ error (0, 0, ++ _("character offset %s is too large"), quote (bad_num)); + else + error (0, 0, + _("field number %s is too large"), quote (bad_num)); +@@ -475,7 +557,7 @@ + fieldstr++; + } + else +- FATAL_ERROR (_("invalid byte or field list")); ++ FATAL_ERROR (_("invalid byte, character or field list")); + } + + max_range_endpoint = 0; +@@ -568,6 +650,63 @@ } -@@ -951,6 +1137,107 @@ - return ptr; } +#if HAVE_MBRTOWC -+static char * -+limfield_mb (const struct line *line, const struct keyfield *key) -+{ -+ char *ptr = line->text, *lim = ptr + line->length - 1; -+ size_t eword = key->eword, echar = key->echar; -+ int i; -+ size_t mblength; -+ mbstate_t state; -+ -+ memset (&state, '\0', sizeof(mbstate_t)); -+ -+ if (tab_length) -+ while (ptr < lim && eword--) -+ { -+ while (ptr < lim && memcmp (ptr, tab, tab_length) != 0) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ if (ptr < lim && (eword | echar)) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ } -+ else -+ while (ptr < lim && eword--) -+ { -+ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) -+ ptr += mblength; -+ if (ptr < lim) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ while (ptr < lim && !ismbblank (ptr, lim - ptr, &mblength)) -+ ptr += mblength; -+ } -+ -+ -+# ifdef POSIX_UNSPECIFIED -+ /* Make LIM point to the end of (one byte past) the current field. */ -+ if (tab_length) -+ { -+ char *newlim, *p; -+ -+ newlim = NULL; -+ for (p = ptr; p < lim;) -+ { -+ if (memcmp (p, tab, tab_length) == 0) -+ { -+ newlim = p; -+ break; -+ } -+ -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ p += mblength; -+ } -+ } -+ else -+ { -+ char *newlim; -+ newlim = ptr; ++/* This function is in use for the following case. + -+ while (newlim < lim && ismbblank (newlim, lim - newlim, &mblength)) -+ newlim += mblength; -+ if (ptr < lim) -+ { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); -+ ptr += mblength; -+ } -+ while (newlim < lim && !ismbblank (newlim, lim - newlim, &mblength)) -+ newlim += mblength; -+ lim = newlim; -+ } -+# endif ++ 1. Read from the stream STREAM, printing to standard output any selected ++ characters. + -+ /* If we're skipping leading blanks, don't start counting characters -+ * until after skipping past any leading blanks. */ -+ if (key->skipsblanks) -+ while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) -+ ptr += mblength; ++ 2. Read from stream STREAM, printing to standard output any selected bytes, ++ without splitting multibyte characters. */ ++ ++static void ++cut_characters_or_cut_bytes_no_split (FILE *stream) ++{ ++ int idx; /* number of bytes or characters in the line so far. */ ++ char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ ++ char *bufpos; /* Next read position of BUF. */ ++ size_t buflen; /* The length of the byte sequence in buf. */ ++ wint_t wc; /* A gotten wide character. */ ++ size_t mblength; /* The byte size of a multibyte character which shows ++ as same character as WC. */ ++ mbstate_t state; /* State of the stream. */ ++ int convfail; /* 1, when conversion is failed. Otherwise 0. */ + ++ idx = 0; ++ buflen = 0; ++ bufpos = buf; + memset (&state, '\0', sizeof(mbstate_t)); + -+ /* Advance PTR by ECHAR (if possible), but no further than LIM. */ -+ for (i = 0; i < echar; i++) ++ while (1) + { -+ GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); ++ REFILL_BUFFER (buf, bufpos, buflen, stream); + -+ if (ptr + mblength > lim) -+ break; -+ else -+ ptr += mblength; -+ } ++ GET_NEXT_WC_FROM_BUFFER (wc, bufpos, buflen, mblength, state, convfail); + -+ return ptr; -+} -+#endif ++ if (wc == WEOF) ++ { ++ if (idx > 0) ++ putchar ('\n'); ++ break; ++ } ++ else if (wc == L'\n') ++ { ++ putchar ('\n'); ++ idx = 0; ++ } ++ else ++ { ++ idx += (operating_mode == byte_mode) ? mblength : 1; ++ if (print_kth (idx, NULL)) ++ fwrite (bufpos, mblength, sizeof(char), stdout); ++ } + - /* Fill BUF reading from FP, moving buf->left bytes from the end - of buf->buf to the beginning first. If EOF is reached and the - file wasn't terminated by a newline, supply one. Set up BUF's line -@@ -1067,7 +1354,7 @@ - hideously fast. */ - - static int --numcompare (const char *a, const char *b) -+numcompare_uni (const char *a, const char *b) - { - while (blanks[to_uchar (*a)]) - a++; -@@ -1077,6 +1364,25 @@ - return strnumcmp (a, b, decimal_point, thousands_sep); - } - -+#if HAVE_MBRTOWC -+static int -+numcompare_mb (const char *a, const char *b) -+{ -+ size_t mblength, len; -+ len = strlen (a); /* okay for UTF-8 */ -+ while (*a && ismbblank (a, len > MB_CUR_MAX ? MB_CUR_MAX : len, &mblength)) -+ { -+ a += mblength; -+ len -= mblength; ++ buflen -= mblength; ++ bufpos += mblength; + } -+ len = strlen (b); /* okay for UTF-8 */ -+ while (*b && ismbblank (b, len > MB_CUR_MAX ? MB_CUR_MAX : len, &mblength)) -+ b += mblength; -+ -+ return strnumcmp (a, b, decimal_point, thousands_sep); +} -+#endif /* HAV_EMBRTOWC */ -+ - static int - general_numcompare (const char *sa, const char *sb) - { -@@ -1110,7 +1416,7 @@ - Return 0 if the name in S is not recognized. */ ++#endif ++ + /* Read from stream STREAM, printing to standard output any selected fields. */ - static int --getmonth (char const *month, size_t len) -+getmonth_uni (char const *month, size_t len) - { - size_t lo = 0; - size_t hi = MONTHS_PER_YEAR; -@@ -1152,11 +1458,79 @@ - return 0; + static void +@@ -689,13 +828,192 @@ + } } +#if HAVE_MBRTOWC -+static int -+getmonth_mb (const char *s, size_t len) ++static void ++cut_fields_mb (FILE *stream) +{ -+ char *month; -+ register size_t i; -+ register int lo = 0, hi = MONTHS_PER_YEAR, result; -+ char *tmp; -+ size_t wclength, mblength; -+ const char **pp; -+ const wchar_t **wpp; -+ wchar_t *month_wcs; -+ mbstate_t state; -+ -+ while (len > 0 && ismbblank (s, len, &mblength)) -+ { -+ s += mblength; -+ len -= mblength; -+ } -+ -+ if (len == 0) -+ return 0; -+ -+ month = (char *) alloca (len + 1); ++ int c; ++ unsigned int field_idx; ++ int found_any_selected_field; ++ int buffer_first_field; ++ int empty_input; ++ char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ ++ char *bufpos; /* Next read position of BUF. */ ++ size_t buflen; /* The length of the byte sequence in buf. */ ++ wint_t wc = 0; /* A gotten wide character. */ ++ size_t mblength; /* The byte size of a multibyte character which shows ++ as same character as WC. */ ++ mbstate_t state; /* State of the stream. */ ++ int convfail; /* 1, when conversion is failed. Otherwise 0. */ + -+ tmp = (char *) alloca (len + 1); -+ memcpy (tmp, s, len); -+ tmp[len] = '\0'; -+ pp = (const char **)&tmp; -+ month_wcs = (wchar_t *) alloca ((len + 1) * sizeof (wchar_t)); ++ found_any_selected_field = 0; ++ field_idx = 1; ++ bufpos = buf; ++ buflen = 0; + memset (&state, '\0', sizeof(mbstate_t)); + -+ wclength = mbsrtowcs (month_wcs, pp, len + 1, &state); -+ assert (wclength != (size_t)-1 && *pp == NULL); ++ c = getc (stream); ++ empty_input = (c == EOF); ++ if (c != EOF) ++ ungetc (c, stream); ++ else ++ wc = WEOF; + -+ for (i = 0; i < wclength; i++) ++ /* To support the semantics of the -s flag, we may have to buffer ++ all of the first field to determine whether it is `delimited.' ++ But that is unnecessary if all non-delimited lines must be printed ++ and the first field has been selected, or if non-delimited lines ++ must be suppressed and the first field has *not* been selected. ++ That is because a non-delimited line has exactly one field. */ ++ buffer_first_field = (suppress_non_delimited ^ !print_kth (1, NULL)); ++ ++ while (1) + { -+ month_wcs[i] = towupper(month_wcs[i]); -+ if (iswblank (month_wcs[i])) ++ if (field_idx == 1 && buffer_first_field) + { -+ month_wcs[i] = L'\0'; -+ break; -+ } -+ } ++ int len = 0; + -+ wpp = (const wchar_t **)&month_wcs; ++ while (1) ++ { ++ REFILL_BUFFER (buf, bufpos, buflen, stream); + -+ mblength = wcsrtombs (month, wpp, len + 1, &state); -+ assert (mblength != (-1) && *wpp == NULL); ++ GET_NEXT_WC_FROM_BUFFER ++ (wc, bufpos, buflen, mblength, state, convfail); + -+ do -+ { -+ int ix = (lo + hi) / 2; ++ if (wc == WEOF) ++ break; + -+ if (strncmp (month, monthtab[ix].name, strlen (monthtab[ix].name)) < 0) -+ hi = ix; -+ else -+ lo = ix; -+ } -+ while (hi - lo > 1); ++ field_1_buffer = xrealloc (field_1_buffer, len + mblength); ++ memcpy (field_1_buffer + len, bufpos, mblength); ++ len += mblength; ++ buflen -= mblength; ++ bufpos += mblength; + -+ result = (!strncmp (month, monthtab[lo].name, strlen (monthtab[lo].name)) -+ ? monthtab[lo].val : 0); ++ if (!convfail && (wc == L'\n' || wc == wcdelim)) ++ break; ++ } + -+ return result; -+} -+#endif ++ if (wc == WEOF) ++ break; + - /* Compare two lines A and B trying every key in sequence until there - are no more keys or a difference is found. */ - - static int --keycompare (const struct line *a, const struct line *b) -+keycompare_uni (const struct line *a, const struct line *b) - { - struct keyfield const *key = keylist; - -@@ -1326,6 +1700,177 @@ - return key->reverse ? -diff : diff; - } - -+#if HAVE_MBRTOWC -+static int -+keycompare_mb (const struct line *a, const struct line *b) -+{ -+ struct keyfield *key = keylist; ++ /* If the first field extends to the end of line (it is not ++ delimited) and we are printing all non-delimited lines, ++ print this one. */ ++ if (convfail || (!convfail && wc != wcdelim)) ++ { ++ if (suppress_non_delimited) ++ { ++ /* Empty. */ ++ } ++ else ++ { ++ fwrite (field_1_buffer, sizeof (char), len, stdout); ++ /* Make sure the output line is newline terminated. */ ++ if (convfail || (!convfail && wc != L'\n')) ++ putchar ('\n'); ++ } ++ continue; ++ } + -+ /* For the first iteration only, the key positions have been -+ precomputed for us. */ -+ char *texta = a->keybeg; -+ char *textb = b->keybeg; -+ char *lima = a->keylim; -+ char *limb = b->keylim; ++ if (print_kth (1, NULL)) ++ { ++ /* Print the field, but not the trailing delimiter. */ ++ fwrite (field_1_buffer, sizeof (char), len - 1, stdout); ++ found_any_selected_field = 1; ++ } ++ ++field_idx; ++ } ++ ++ if (wc != WEOF) ++ { ++ if (print_kth (field_idx, NULL)) ++ { ++ if (found_any_selected_field) ++ { ++ fwrite (output_delimiter_string, sizeof (char), ++ output_delimiter_length, stdout); ++ } ++ found_any_selected_field = 1; ++ } + -+ size_t mblength_a, mblength_b; -+ wchar_t wc_a, wc_b; -+ mbstate_t state_a, state_b; ++ while (1) ++ { ++ REFILL_BUFFER (buf, bufpos, buflen, stream); + -+ int diff; ++ GET_NEXT_WC_FROM_BUFFER ++ (wc, bufpos, buflen, mblength, state, convfail); + -+ memset (&state_a, '\0', sizeof(mbstate_t)); -+ memset (&state_b, '\0', sizeof(mbstate_t)); ++ if (wc == WEOF) ++ break; ++ else if (!convfail && (wc == wcdelim || wc == L'\n')) ++ { ++ buflen -= mblength; ++ bufpos += mblength; ++ break; ++ } + -+ for (;;) -+ { -+ unsigned char *translate = (unsigned char *) key->translate; -+ bool const *ignore = key->ignore; ++ if (print_kth (field_idx, NULL)) ++ fwrite (bufpos, mblength, sizeof(char), stdout); + -+ /* Find the lengths. */ -+ size_t lena = lima <= texta ? 0 : lima - texta; -+ size_t lenb = limb <= textb ? 0 : limb - textb; ++ buflen -= mblength; ++ bufpos += mblength; ++ } ++ } + -+ /* Actually compare the fields. */ -+ if (key->numeric | key->general_numeric) -+ { -+ char savea = *lima, saveb = *limb; ++ if ((!convfail || wc == L'\n') && buflen < 1) ++ wc = WEOF; + -+ *lima = *limb = '\0'; -+ if (force_general_numcompare) -+ diff = general_numcompare (texta, textb); -+ else -+ diff = ((key->numeric ? numcompare : general_numcompare) -+ (texta, textb)); -+ *lima = savea, *limb = saveb; -+ } -+ else if (key->month) -+ diff = getmonth (texta, lena) - getmonth (textb, lenb); -+ else ++ if (!convfail && wc == wcdelim) ++ ++field_idx; ++ else if (wc == WEOF || (!convfail && wc == L'\n')) + { -+ if (ignore || translate) -+ { -+ char *copy_a = (char *) alloca (lena + 1 + lenb + 1); -+ char *copy_b = copy_a + lena + 1; -+ size_t new_len_a, new_len_b; -+ size_t i, j; ++ if (found_any_selected_field ++ || (!empty_input && !(suppress_non_delimited && field_idx == 1))) ++ putchar ('\n'); ++ if (wc == WEOF) ++ break; ++ field_idx = 1; ++ found_any_selected_field = 0; ++ } ++ } ++} ++#endif + -+ /* Ignore and/or translate chars before comparing. */ -+# define IGNORE_CHARS(NEW_LEN, LEN, TEXT, COPY, WC, MBLENGTH, STATE) \ -+ do \ -+ { \ -+ wchar_t uwc; \ -+ char mbc[MB_LEN_MAX]; \ -+ mbstate_t state_wc; \ -+ \ -+ for (NEW_LEN = i = 0; i < LEN;) \ -+ { \ -+ mbstate_t state_bak; \ -+ \ -+ state_bak = STATE; \ -+ MBLENGTH = mbrtowc (&WC, TEXT + i, LEN - i, &STATE); \ -+ \ -+ if (MBLENGTH == (size_t)-2 || MBLENGTH == (size_t)-1 \ -+ || MBLENGTH == 0) \ -+ { \ -+ if (MBLENGTH == (size_t)-2 || MBLENGTH == (size_t)-1) \ -+ STATE = state_bak; \ -+ if (!ignore) \ -+ COPY[NEW_LEN++] = TEXT[i++]; \ -+ continue; \ -+ } \ -+ \ -+ if (ignore) \ -+ { \ -+ if ((ignore == nonprinting && !iswprint (WC)) \ -+ || (ignore == nondictionary \ -+ && !iswalnum (WC) && !iswblank (WC))) \ -+ { \ -+ i += MBLENGTH; \ -+ continue; \ -+ } \ -+ } \ -+ \ -+ if (translate) \ -+ { \ -+ \ -+ uwc = towupper(WC); \ -+ if (WC == uwc) \ -+ { \ -+ memcpy (mbc, TEXT + i, MBLENGTH); \ -+ i += MBLENGTH; \ -+ } \ -+ else \ -+ { \ -+ i += MBLENGTH; \ -+ WC = uwc; \ -+ memset (&state_wc, '\0', sizeof (mbstate_t)); \ -+ \ -+ MBLENGTH = wcrtomb (mbc, WC, &state_wc); \ -+ assert (MBLENGTH != (size_t)-1 && MBLENGTH != 0); \ -+ } \ -+ \ -+ for (j = 0; j < MBLENGTH; j++) \ -+ COPY[NEW_LEN++] = mbc[j]; \ -+ } \ -+ else \ -+ for (j = 0; j < MBLENGTH; j++) \ -+ COPY[NEW_LEN++] = TEXT[i++]; \ -+ } \ -+ COPY[NEW_LEN] = '\0'; \ -+ } \ -+ while (0) -+ IGNORE_CHARS (new_len_a, lena, texta, copy_a, -+ wc_a, mblength_a, state_a); -+ IGNORE_CHARS (new_len_b, lenb, textb, copy_b, -+ wc_b, mblength_b, state_b); -+ diff = xmemcoll (copy_a, new_len_a, copy_b, new_len_b); -+ } -+ else if (lena == 0) -+ diff = - NONZERO (lenb); -+ else if (lenb == 0) -+ goto greater; + static void + cut_stream (FILE *stream) + { +- if (operating_mode == byte_mode) +- cut_bytes (stream); ++#if HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1 && !force_singlebyte_mode) ++ { ++ switch (operating_mode) ++ { ++ case byte_mode: ++ if (byte_mode_character_aware) ++ cut_characters_or_cut_bytes_no_split (stream); + else -+ diff = xmemcoll (texta, lena, textb, lenb); -+ } ++ cut_bytes (stream); ++ break; + -+ if (diff) -+ goto not_equal; ++ case character_mode: ++ cut_characters_or_cut_bytes_no_split (stream); ++ break; + -+ key = key->next; -+ if (! key) -+ break; ++ case field_mode: ++ cut_fields_mb (stream); ++ break; + -+ /* Find the beginning and limit of the next field. */ -+ if (key->eword != -1) -+ lima = limfield (a, key), limb = limfield (b, key); ++ default: ++ abort (); ++ } ++ } + else +- cut_fields (stream); ++#endif ++ { ++ if (operating_mode == field_mode) ++ cut_fields (stream); + else -+ lima = a->text + a->length - 1, limb = b->text + b->length - 1; ++ cut_bytes (stream); ++ } + } + + /* Process file FILE to standard output. +@@ -745,6 +1063,8 @@ + bool ok; + bool delim_specified = false; + char *spec_list_string IF_LINT(= NULL); ++ char mbdelim[MB_LEN_MAX + 1]; ++ size_t delimlen = 0; + + initialize_main (&argc, &argv); + program_name = argv[0]; +@@ -767,7 +1087,6 @@ + switch (optc) + { + case 'b': +- case 'c': + /* Build the byte list. */ + if (operating_mode != undefined_mode) + FATAL_ERROR (_("only one type of list may be specified")); +@@ -775,6 +1094,14 @@ + spec_list_string = optarg; + break; + ++ case 'c': ++ /* Build the character list. */ ++ if (operating_mode != undefined_mode) ++ FATAL_ERROR (_("only one type of list may be specified")); ++ operating_mode = character_mode; ++ spec_list_string = optarg; ++ break; + -+ if (key->sword != -1) -+ texta = begfield (a, key), textb = begfield (b, key); -+ else -+ { -+ texta = a->text, textb = b->text; -+ if (key->skipsblanks) + case 'f': + /* Build the field list. */ + if (operating_mode != undefined_mode) +@@ -786,10 +1113,35 @@ + case 'd': + /* New delimiter. */ + /* Interpret -d '' to mean `use the NUL byte as the delimiter.' */ +- if (optarg[0] != '\0' && optarg[1] != '\0') +- FATAL_ERROR (_("the delimiter must be a single character")); +- delim = optarg[0]; +- delim_specified = true; ++#if HAVE_MBRTOWC + { -+ while (texta < lima && ismbblank (texta, lima - texta, &mblength_a)) -+ texta += mblength_a; -+ while (textb < limb && ismbblank (textb, limb - textb, &mblength_b)) -+ textb += mblength_b; -+ } -+ } -+ } ++ if(MB_CUR_MAX > 1) ++ { ++ mbstate_t state; + -+ return 0; ++ memset (&state, '\0', sizeof(mbstate_t)); ++ delimlen = mbrtowc (&wcdelim, optarg, strnlen(optarg, MB_LEN_MAX), &state); ++ ++ if (delimlen == (size_t)-1 || delimlen == (size_t)-2) ++ ++force_singlebyte_mode; ++ else ++ { ++ delimlen = (delimlen < 1) ? 1 : delimlen; ++ if (wcdelim != L'\0' && *(optarg + delimlen) != '\0') ++ FATAL_ERROR (_("the delimiter must be a single character")); ++ memcpy (mbdelim, optarg, delimlen); ++ } ++ } + -+greater: -+ diff = 1; -+not_equal: -+ return key->reverse ? -diff : diff; -+} ++ if (MB_CUR_MAX <= 1 || force_singlebyte_mode) +#endif -+ - /* Compare two lines A and B, returning negative, zero, or positive - depending on whether A compares less than, equal to, or greater than B. */ ++ { ++ if (optarg[0] != '\0' && optarg[1] != '\0') ++ FATAL_ERROR (_("the delimiter must be a single character")); ++ delim = (unsigned char) optarg[0]; ++ } ++ delim_specified = true; ++ } + break; -@@ -2127,7 +2672,7 @@ - atexit (close_stdout); + case OUTPUT_DELIMITER_OPTION: +@@ -802,6 +1154,7 @@ + break; - hard_LC_COLLATE = hard_locale (LC_COLLATE); --#if HAVE_NL_LANGINFO -+#if HAVE_LANGINFO_CODESET - hard_LC_TIME = hard_locale (LC_TIME); - #endif + case 'n': ++ byte_mode_character_aware = 1; + break; -@@ -2148,6 +2693,27 @@ - thousands_sep = -1; - } + case 's': +@@ -824,7 +1177,7 @@ + if (operating_mode == undefined_mode) + FATAL_ERROR (_("you must specify a list of bytes, characters, or fields")); -+#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) +- if (delim != '\0' && operating_mode != field_mode) ++ if (delim_specified && operating_mode != field_mode) + FATAL_ERROR (_("an input delimiter may be specified only\ + when operating on fields")); + +@@ -851,15 +1204,34 @@ + } + + if (!delim_specified) +- delim = '\t'; + { -+ inittables = inittables_mb; -+ begfield = begfield_mb; -+ limfield = limfield_mb; -+ getmonth = getmonth_mb; -+ keycompare = keycompare_mb; -+ numcompare = numcompare_mb; -+ } -+ else ++ delim = '\t'; ++#ifdef HAVE_MBRTOWC ++ wcdelim = L'\t'; ++ mbdelim[0] = '\t'; ++ mbdelim[1] = '\0'; ++ delimlen = 1; +#endif -+ { -+ inittables = inittables_uni; -+ begfield = begfield_uni; -+ limfield = limfield_uni; -+ getmonth = getmonth_uni; -+ keycompare = keycompare_uni; -+ numcompare = numcompare_uni; + } + + if (output_delimiter_string == NULL) + { +- static char dummy[2]; +- dummy[0] = delim; +- dummy[1] = '\0'; +- output_delimiter_string = dummy; +- output_delimiter_length = 1; ++#ifdef HAVE_MBRTOWC ++ if (MB_CUR_MAX > 1 && !force_singlebyte_mode) ++ { ++ output_delimiter_string = xstrdup(mbdelim); ++ output_delimiter_length = delimlen; ++ } + - have_read_stdin = false; - inittables (); ++ if (MB_CUR_MAX <= 1 || force_singlebyte_mode) ++#endif ++ { ++ static char dummy[2]; ++ dummy[0] = delim; ++ dummy[1] = '\0'; ++ output_delimiter_string = dummy; ++ output_delimiter_length = 1; ++ } + } -@@ -2349,13 +2915,35 @@ + if (optind == argc) +--- coreutils-5.95/tests/sort/Makefile.am.i18n 2005-10-24 22:02:25.000000000 +0100 ++++ coreutils-5.95/tests/sort/Makefile.am 2006-05-15 15:08:57.000000000 +0100 +@@ -43,14 +43,16 @@ + nul-nls.E use-nl.O use-nl.E o2.O o2.E nul-tab.O nul-tab.E + ##test-files-end - case 't': - { -- char newtab = optarg[0]; -- if (! newtab) -+ char newtab[MB_LEN_MAX + 1]; -+ size_t newtab_length = 1; -+ strncpy (newtab, optarg, MB_LEN_MAX); -+ if (! newtab[0]) - error (SORT_FAILURE, 0, _("empty tab")); -- if (optarg[1]) -+#if HAVE_MBRTOWC -+ if (MB_CUR_MAX > 1) -+ { -+ wchar_t wc; -+ mbstate_t state; -+ size_t i; +-EXTRA_DIST = Test.pm $x-tests $(explicit) $(maint_gen) +-noinst_SCRIPTS = $x-tests ++run_gen += mb1.O mb2.O + -+ memset (&state, '\0', sizeof (mbstate_t)); -+ newtab_length = mbrtowc (&wc, newtab, strnlen (newtab, -+ MB_LEN_MAX), -+ &state); -+ switch (newtab_length) -+ { -+ case (size_t) -1: -+ case (size_t) -2: -+ case 0: -+ newtab_length = 1; -+ } -+ } -+#endif -+ if (newtab_length == 1 && optarg[1]) - { - if (STREQ (optarg, "\\0")) -- newtab = '\0'; -+ newtab[0] = '\0'; - else - { - /* Provoke with `sort -txx'. Complain about -@@ -2366,9 +2954,12 @@ - quote (optarg)); - } - } -- if (tab != TAB_DEFAULT && tab != newtab) -+ if (tab_length -+ && (tab_length != newtab_length -+ || memcmp (tab, newtab, tab_length) != 0)) - error (SORT_FAILURE, 0, _("incompatible tabs")); -- tab = newtab; -+ memcpy (tab, newtab, newtab_length); -+ tab_length = newtab_length; - } - break; ++EXTRA_DIST = Test.pm $x-tests $(explicit) $(maint_gen) mb1.I mb1.X mb2.I mb2.X ++noinst_SCRIPTS = $x-tests # $x-mb-tests + TESTS_ENVIRONMENT = \ + PATH="`pwd`/../../src$(PATH_SEPARATOR)$$PATH" + + editpl = sed -e 's,@''PERL''@,$(PERL),g' -e 's,@''srcdir''@,$(srcdir),g' + +-TESTS = $x-tests ++TESTS = $x-tests $x-mb-tests ---- /dev/null 2005-10-10 09:36:06.437701000 +0100 -+++ coreutils-5.93/tests/sort/sort-mb-tests 2005-12-23 08:53:01.000000000 +0000 + mk_script = $(srcdir)/../mk-script + $(srcdir)/$x-tests: $(mk_script) Test.pm Makefile.am +--- /dev/null 2006-05-15 09:11:19.652773000 +0100 ++++ coreutils-5.95/tests/sort/mb1.I 2006-05-15 15:08:57.000000000 +0100 +@@ -0,0 +1,4 @@ ++Apple＠10 ++Banana＠5 ++Citrus＠20 ++Cherry＠30 +--- /dev/null 2006-05-15 09:11:19.652773000 +0100 ++++ coreutils-5.95/tests/sort/mb1.X 2006-05-15 15:08:57.000000000 +0100 +@@ -0,0 +1,4 @@ ++Banana＠5 ++Apple＠10 ++Citrus＠20 ++Cherry＠30 +--- /dev/null 2006-05-15 09:11:19.652773000 +0100 ++++ coreutils-5.95/tests/sort/mb2.X 2006-05-15 15:08:57.000000000 +0100 +@@ -0,0 +1,4 @@ ++Citrus＠ＡＡ20＠＠5 ++Cherry＠ＡＡ30＠＠10 ++Apple＠ＡＡ10＠＠20 ++Banana＠ＡＡ5＠＠30 +--- /dev/null 2006-05-15 09:11:19.652773000 +0100 ++++ coreutils-5.95/tests/sort/sort-mb-tests 2006-05-15 15:08:57.000000000 +0100 @@ -0,0 +1,58 @@ +#! /bin/sh +case $# in @@ -3971,58 +3991,8 @@ +fi +test $errors = 0 || errors=1 +exit $errors ---- /dev/null 2005-10-10 09:36:06.437701000 +0100 -+++ coreutils-5.93/tests/sort/mb1.I 2005-12-23 08:53:01.000000000 +0000 -@@ -0,0 +1,4 @@ -+Apple＠10 -+Banana＠5 -+Citrus＠20 -+Cherry＠30 ---- /dev/null 2005-10-10 09:36:06.437701000 +0100 -+++ coreutils-5.93/tests/sort/mb2.I 2005-12-23 08:53:01.000000000 +0000 -@@ -0,0 +1,4 @@ -+Apple＠ＡＡ10＠＠20 -+Banana＠ＡＡ5＠＠30 -+Citrus＠ＡＡ20＠＠5 -+Cherry＠ＡＡ30＠＠10 ---- /dev/null 2005-10-10 09:36:06.437701000 +0100 -+++ coreutils-5.93/tests/sort/mb1.X 2005-12-23 08:53:01.000000000 +0000 -@@ -0,0 +1,4 @@ -+Banana＠5 -+Apple＠10 -+Citrus＠20 -+Cherry＠30 ---- /dev/null 2005-10-10 09:36:06.437701000 +0100 -+++ coreutils-5.93/tests/sort/mb2.X 2005-12-23 08:53:01.000000000 +0000 -@@ -0,0 +1,4 @@ -+Citrus＠ＡＡ20＠＠5 -+Cherry＠ＡＡ30＠＠10 -+Apple＠ＡＡ10＠＠20 -+Banana＠ＡＡ5＠＠30 ---- coreutils-5.93/tests/sort/Makefile.am.i18n 2005-10-24 22:02:25.000000000 +0100 -+++ coreutils-5.93/tests/sort/Makefile.am 2005-12-23 08:53:01.000000000 +0000 -@@ -43,14 +43,16 @@ - nul-nls.E use-nl.O use-nl.E o2.O o2.E nul-tab.O nul-tab.E - ##test-files-end - --EXTRA_DIST = Test.pm $x-tests $(explicit) $(maint_gen) --noinst_SCRIPTS = $x-tests -+run_gen += mb1.O mb2.O -+ -+EXTRA_DIST = Test.pm $x-tests $(explicit) $(maint_gen) mb1.I mb1.X mb2.I mb2.X -+noinst_SCRIPTS = $x-tests # $x-mb-tests - TESTS_ENVIRONMENT = \ - PATH="`pwd`/../../src$(PATH_SEPARATOR)$$PATH" - - editpl = sed -e 's,@''PERL''@,$(PERL),g' -e 's,@''srcdir''@,$(srcdir),g' - --TESTS = $x-tests -+TESTS = $x-tests $x-mb-tests - - mk_script = $(srcdir)/../mk-script - $(srcdir)/$x-tests: $(mk_script) Test.pm Makefile.am ---- coreutils-5.93/tests/sort/Makefile.in.i18n 2005-11-05 22:12:25.000000000 +0000 -+++ coreutils-5.93/tests/sort/Makefile.in 2005-12-23 09:00:37.000000000 +0000 +--- coreutils-5.95/tests/sort/Makefile.in.i18n 2006-05-04 13:52:51.000000000 +0100 ++++ coreutils-5.95/tests/sort/Makefile.in 2006-05-15 15:08:57.000000000 +0100 @@ -107,25 +107,25 @@ $(top_srcdir)/m4/onceonly_2_57.m4 $(top_srcdir)/m4/openat.m4 \ $(top_srcdir)/m4/pathmax.m4 $(top_srcdir)/m4/perl.m4 \ @@ -4093,7 +4063,7 @@ LN_S = @LN_S@ LTLIBICONV = @LTLIBICONV@ LTLIBINTL = @LTLIBINTL@ -@@ -268,30 +270,23 @@ +@@ -266,30 +268,23 @@ build_os = @build_os@ build_vendor = @build_vendor@ datadir = @datadir@ @@ -4124,7 +4094,7 @@ sbindir = @sbindir@ sharedstatedir = @sharedstatedir@ sysconfdir = @sysconfdir@ -@@ -318,33 +313,37 @@ +@@ -316,33 +311,37 @@ neg-nls.I neg-nls.X nul-nls.I nul-nls.X use-nl.I use-nl.X o2.I o2.X nul-tab.I \ nul-tab.X @@ -4185,3 +4155,34 @@ mk_script = $(srcdir)/../mk-script MAINTAINERCLEANFILES = $x-tests $(maint_gen) CLEANFILES = $(run_gen) +--- /dev/null 2006-05-15 09:11:19.652773000 +0100 ++++ coreutils-5.95/tests/sort/mb2.I 2006-05-15 15:08:57.000000000 +0100 +@@ -0,0 +1,4 @@ ++Apple＠ＡＡ10＠＠20 ++Banana＠ＡＡ5＠＠30 ++Citrus＠ＡＡ20＠＠5 ++Cherry＠ＡＡ30＠＠10 +--- coreutils-5.95/lib/linebuffer.h.i18n 2005-05-14 08:58:06.000000000 +0100 ++++ coreutils-5.95/lib/linebuffer.h 2006-05-15 15:08:57.000000000 +0100 +@@ -22,6 +22,11 @@ + + # include + ++/* Get mbstate_t. */ ++# if HAVE_WCHAR_H ++# include ++# endif ++ + /* A `struct linebuffer' holds a line of text. */ + + struct linebuffer +@@ -29,6 +34,9 @@ + size_t size; /* Allocated. */ + size_t length; /* Used. */ + char *buffer; ++# if HAVE_WCHAR_H ++ mbstate_t state; ++# endif + }; + + /* Initialize linebuffer LINEBUFFER for use. */