Update.

	* locale/programs/ld-ctype.c (ctype_finish): Take all characters from
	the input charset into account when generating the hash table.
	(allocate_arrays): Correct setting default width.  Not all empty slots
	in the table are filled, only those not covert explicitly by the
	locale description and in the charset.

	* stdio-common/vfscanf.c: Make sure to always return WEOF and EOF for
	wide character version.
	For %C handling, test correct pointer variable for NULL.

	* wcsmbs/wctob.c: Handle WEOF special.

	* wcsmbs/wcwidth.h: 0xff in width array means invalid character.

	* wctype/wctype.h: Protect gcc-isms with __extension__.  Avoid
	always-true test to avoid warning.

33 files changed, 435 insertions, 159 deletions

diff --git a/ChangeLog b/ChangeLog
index e53b21df1f..fa9eaeafe8 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,8 +1,25 @@
 2000-06-27  Ulrich Drepper  <drepper@redhat.com>
 
+	* locale/programs/ld-ctype.c (ctype_finish): Take all characters from
+	the input charset into account when generating the hash table.
+	(allocate_arrays): Correct setting default width.  Not all empty slots
+	in the table are filled, only those not covert explicitly by the
+	locale description and in the charset.
+
 	* wctype/towctrans.c (__towctrans): Be graceful and accept error
 	return values from the wctrans function.
 
+	* stdio-common/vfscanf.c: Make sure to always return WEOF and EOF for
+	wide character version.
+	For %C handling, test correct pointer variable for NULL.
+
+	* wcsmbs/wctob.c: Handle WEOF special.
+
+	* wcsmbs/wcwidth.h: 0xff in width array means invalid character.
+
+	* wctype/wctype.h: Protect gcc-isms with __extension__.  Avoid
+	always-true test to avoid warning.
+
 2000-06-27  Greg McGary  <greg@mcgary.org>
 
 	* elf/dl-open.c (_dl_sysdep_start): Wrap weak_extern decl in BP_SYM ().
diff --git a/locale/programs/ld-ctype.c b/locale/programs/ld-ctype.c
index ce097e741b..417660160f 100644
--- a/locale/programs/ld-ctype.c
+++ b/locale/programs/ld-ctype.c
@@ -345,6 +345,10 @@ ctype_finish (struct localedef_t *locale, struct charmap_t *charmap)
   struct charseq *space_seq;
   struct locale_ctype_t *ctype = locale->categories[LC_CTYPE].ctype;
   int warned;
+  const void *key;
+  size_t len;
+  void *vdata;
+  void *curs;
 
   /* Now resolve copying and also handle completely missing definitions.  */
   if (ctype == NULL)
@@ -637,6 +641,21 @@ character '%s' in class `%s' must not be in class `%s'"),
 	  }
       }
 
+  /* Now set all the other characters of the character set to the
+     default width.  */
+  curs = NULL;
+  while (iterate_table (&charmap->char_table, &curs, &key, &len, &vdata) == 0)
+    {
+      struct charseq *data = (struct charseq *) vdata;
+
+      if (data->ucs4 == UNINITIALIZED_CHAR_VALUE)
+	data->ucs4 = repertoire_find_value (ctype->repertoire,
+					    data->name, len);
+
+      if (data->ucs4 != ILLEGAL_CHAR_VALUE)
+	(void) find_idx (ctype, NULL, NULL, NULL, data->ucs4);
+    }
+
   /* There must be a multiple of 10 digits.  */
   if (ctype->mbdigits_act % 10 != 0)
     {
@@ -3158,6 +3177,10 @@ allocate_arrays (struct locale_ctype_t *ctype, struct charmap_t *charmap,
 {
   size_t idx;
   size_t width_table_size;
+  const void *key;
+  size_t len;
+  void *vdata;
+  void *curs;
 
   /* First we have to decide how we organize the arrays.  It is easy
      for a one-byte character set.  But multi-byte character set
@@ -3345,8 +3368,8 @@ Computing table size for character classes might take a while..."),
   width_table_size = (ctype->plane_size * ctype->plane_cnt + 3) & ~3ul;
   ctype->width = (unsigned char *) xmalloc (width_table_size);
 
-  /* Initialize with default width value.  */
-  memset (ctype->width, charmap->width_default, width_table_size);
+  /* Initialize with -1.  */
+  memset (ctype->width, '\xff', width_table_size);
   if (charmap->width_rules != NULL)
     {
       size_t cnt;
@@ -3389,8 +3412,10 @@ Computing table size for character classes might take a while..."),
 		  size_t depth = 0;
 
 		  while (ctype->names[nr + depth * ctype->plane_size] != wch)
-		    ++depth;
-		  assert (depth < ctype->plane_cnt);
+		    {
+		      ++depth;
+		      assert (depth < ctype->plane_cnt);
+		    }
 
 		  ctype->width[nr + depth * ctype->plane_size]
 		    = charmap->width_rules[cnt].width;
@@ -3421,6 +3446,37 @@ Computing table size for character classes might take a while..."),
 	}
     }
 
+  /* Now set all the other characters of the character set to the
+     default width.  */
+  curs = NULL;
+  while (iterate_table (&charmap->char_table, &curs, &key, &len, &vdata) == 0)
+    {
+      struct charseq *data = (struct charseq *) vdata;
+      size_t nr;
+      size_t depth;
+
+      if (data->ucs4 == UNINITIALIZED_CHAR_VALUE)
+	data->ucs4 = repertoire_find_value (ctype->repertoire,
+					    data->name, len);
+
+      if (data->ucs4 != ILLEGAL_CHAR_VALUE)
+	{
+	  nr = data->ucs4 % ctype->plane_size;
+	  depth = 0;
+
+	  while (ctype->names[nr + depth * ctype->plane_size] != data->ucs4)
+	    {
+	      ++depth;
+	      assert (depth < ctype->plane_cnt);
+	    }
+
+	  if (ctype->width[nr + depth * ctype->plane_size]
+	      == (unsigned char) '\xff')
+	    ctype->width[nr + depth * ctype->plane_size] =
+	      charmap->width_default;
+	}
+    }
+
   /* Set MB_CUR_MAX.  */
   ctype->mb_cur_max = charmap->mb_cur_max;
 
diff --git a/localedata/ChangeLog b/localedata/ChangeLog
index 0b7dac9a8a..f733268498 100644
--- a/localedata/ChangeLog
+++ b/localedata/ChangeLog
@@ -1,5 +1,47 @@
 2000-06-27  Ulrich Drepper  <drepper@redhat.com>
 
+	* tests-mbwc/dat_iswcntrl.c: U0000 is not in class cntrl.
+	* tests-mbwc/dat_iswctype.c: U0000 is not in class cntrl.
+	U4E06 is not in EUC-JP.
+	* tests-mbwc/dat_swscanf.c: Correct several bugs in the tests.
+	* tests-mbwc/dat_towctrans.c: Likewise.
+	* tests-mbwc/dat_wcscoll.c: Likewise.
+	* tests-mbwc/dat_wcswidth.c: Likewise.
+	* tests-mbwc/dat_wctob.c: Likewise.
+
+	* tests-mbwc/tst_towctrans.c: Remove hack which avoided crash in an
+	versions.
+
+	* tests-mbwc/dat_iswprint.c: Disable one test until it is decided
+	what is correct.
+	* tests-mbwc/dat_wcsxfrm.c: Likewise.
+	* tests-mbwc/dat_wcwidth.c: Likewise.
+
+	* tests-mbwc/tst_funcs.h: Pretty print.
+
+	* tests-mbwc/tst_scscanf.c: Use correct format to avoid warning.
+	* tests-mbwc/tst_wcschr.c: Likewise.
+	* tests-mbwc/tst_wcscpy.c: Likewise.
+	* tests-mbwc/tst_wcscat.c: Likewise.
+	* tests-mbwc/tst_wcsncpy.c: Likewise.
+	* tests-mbwc/tst_wcspbrk.c: Likewise.
+	* tests-mbwc/tst_wcsstr.c: Likewise.
+	* tests-mbwc/tst_wctrans.c: Likewise.
+	* tests-mbwc/tst_wctype.c: Likewise.
+
+	* tests-mbwc/tst_wcscoll.c: Print better error messages.
+
+	* Makefile (tests): Define as $(locale_test_suite) but only
+	if not cross-compiling and shared libs are built.
+	(locale_test_suite): New variable.  Name all new tests from the
+	locale test suite.
+	Add rule to run new tests only when all data is available.
+	* tst-ctype.sh: Add hack to generate en_US.ANSI_X3.4-1968 locale.
+
+	* tst-ctype-de_DE.in: U00A0 is not in class graph.
+
+	* charmaps/EUC-JP: Remove U005C and U007E entries from non-ASCII range.
+
 	* locales/i18n: Backspace isn't blank, tab is.
 
 	* tst-ctype.c (main): Add tests for control characters and space.
diff --git a/localedata/Makefile b/localedata/Makefile
index 0330a973f3..7c88b045a8 100644
--- a/localedata/Makefile
+++ b/localedata/Makefile
@@ -35,15 +35,6 @@ locales := $(filter-out $(addprefix locales/, CVS RCS SCCS %~), \
 repertoiremaps := $(filter-out $(addprefix repertoiremaps/, CVS RCS SCCS %~), \
 					   $(wildcard repertoiremaps/*))
 
-# Disable the tests for now - first the locales have to be generated
-#tests := tst_iswalnum tst_iswprint tst_towctrans tst_wcsncmp tst_wctrans \
-#	  tst_iswalpha tst_iswpunct tst_wcschr tst_wcspbrk tst_wctype	 \
-#	  tst_iswcntrl tst_iswspace tst_wcscoll tst_wcsspn tst_iswdigit	 \
-#	  tst_iswupper tst_wcscpy tst_wcsstr tst_iswgraph tst_iswxdigit	 \
-#	  tst_wcscspn tst_wcswidth tst_iswlower tst_swscanf tst_wcslen	 \
-#	  tst_wctob tst_iswctype tst_towlower tst_wcscat tst_towupper    \
-#	  tst_wcscmp tst_wcsncat tst_wcsncpy tst_wcsxfrm tst_wcwidth
-
 
 subdir-dirs = tests-mbwc
 vpath %.c tests-mbwc
@@ -69,8 +60,9 @@ fmon-tests = n01y12 n02n40 n10y31 n11y41 n12y11 n20n32 n30y20 n41n00 \
 	     y01y10 y02n22 y22n42 y30y21 y32n31 y40y00 y42n21
 
 generated := $(test-input) $(test-output)
-generated-dirs := $(basename $(test-input)) en_US $(ld-test-names) tt_TT\
-		  de_DE.437 $(addprefix tstfmon_,$(fmon-tests))
+generated-dirs := $(basename $(test-input)) $(ld-test-names) tt_TT	\
+		  de_DE.437 $(addprefix tstfmon_,$(fmon-tests))		\
+		  en_US.ANSI_X3.4-1968 ja_JP.EUC-JP
 
 distribute := CHECKSUMS README SUPPORTED ChangeLog			\
 	      $(charmaps) $(locales) $(repertoiremaps)			\
@@ -82,6 +74,22 @@ distribute := CHECKSUMS README SUPPORTED ChangeLog			\
 # Get $(inst_i18ndir) defined.
 include ../Makeconfig
 
+ifeq (no,$(cross-compiling))
+ifeq (yes,$(build-shared))
+# Disable the tests for now - first the locales have to be generated
+locale_test_suite := tst_iswalnum tst_iswprint tst_towctrans tst_wcsncmp \
+		     tst_wctrans tst_iswalpha tst_iswpunct tst_wcschr	 \
+		     tst_wcspbrk tst_wctype tst_iswcntrl tst_iswspace	 \
+		     tst_wcscoll tst_wcsspn tst_iswdigit tst_iswupper	 \
+		     tst_wcscpy tst_wcsstr tst_iswgraph tst_iswxdigit	 \
+		     tst_wcscspn tst_wcswidth tst_iswlower tst_swscanf	 \
+		     tst_wcslen tst_wctob tst_iswctype tst_towlower	 \
+		     tst_wcscat tst_towupper tst_wcscmp tst_wcsncat	 \
+		     tst_wcsncpy tst_wcsxfrm tst_wcwidth
+tests = $(locale_test_suite)
+endif
+endif
+
 # Files to install.
 install-others := $(addprefix $(inst_i18ndir)/, $(charmaps) $(locales) \
 						$(repertoiremaps))
@@ -101,6 +109,7 @@ CFLAGS-tst-mbswcs4.c = -Wno-format
 CFLAGS-tst-mbswcs5.c = -Wno-format
 CFLAGS-tst-trans.c = -Wno-format
 
+
 ifeq (no,$(cross-compiling))
 ifeq (yes,$(build-shared))
 .PHONY: do-collate-test do-tst-fmon do-tst-locale do-tst-rpmatch do-tst-trans \
@@ -123,6 +132,7 @@ do-tst-mbswcs: tst-mbswcs.sh $(objpfx)tst-mbswcs1 $(objpfx)tst-mbswcs2 \
 	$(SHELL) -e $< $(common-objpfx)
 do-tst-ctype: tst-ctype.sh $(objpfx)tst-ctype do-collate-test
 	$(SHELL) -e $< $(common-objpfx)
+$(addsuffix .out,$(addprefix $(objpfx),$(locale_test_suite))): %: do-tst-ctype
 endif
 endif
 
diff --git a/localedata/charmaps/EUC-JP b/localedata/charmaps/EUC-JP
index f5ff8fa8e9..047d6b8c71 100644
--- a/localedata/charmaps/EUC-JP
+++ b/localedata/charmaps/EUC-JP
@@ -242,7 +242,6 @@ CHARMAP
 <U2015>     /xa1/xbd     HORIZONTAL BAR
 <U2010>     /xa1/xbe     HYPHEN
 <UFF0F>     /xa1/xbf     FULLWIDTH SOLIDUS
-<U005C>     /xa1/xc0     REVERSE SOLIDUS
 <U301C>     /xa1/xc1     WAVE DASH
 <U2016>     /xa1/xc2     DOUBLE VERTICAL LINE
 <UFF5C>     /xa1/xc3     FULLWIDTH VERTICAL LINE
@@ -7102,7 +7101,6 @@ CHARMAP
 <U00AF>     /x8f/xa2/xb4 MACRON
 <U02DB>     /x8f/xa2/xb5 OGONEK
 <U02DA>     /x8f/xa2/xb6 RING ABOVE
-<U007E>     /x8f/xa2/xb7 TILDE
 <U0384>     /x8f/xa2/xb8 GREEK TONOS
 <U0385>     /x8f/xa2/xb9 GREEK DIALYTIKA TONOS
 <U00A1>     /x8f/xa2/xc2 INVERTED EXCLAMATION MARK
diff --git a/localedata/tests-mbwc/dat_iswcntrl.c b/localedata/tests-mbwc/dat_iswcntrl.c
index 89d5021916..4eac814ec0 100644
--- a/localedata/tests-mbwc/dat_iswcntrl.c
+++ b/localedata/tests-mbwc/dat_iswcntrl.c
@@ -46,7 +46,11 @@ TST_ISW_LOC (CNTRL, cntrl) = {
     {	TST_ISW_REC (enUS, cntrl)
 	{
 	  {  { WEOF   }, { 0,0,1,0 }  },
+#ifdef SHOJI_IS_RIGHT
 	  {  { 0x0000 }, { 0,0,0,0 }  },
+#else
+	  {  { 0x0000 }, { 0,0,1,0 }  },
+#endif
 	  {  { 0x001F }, { 0,0,0,0 }  },
 	  {  { 0x0020 }, { 0,0,1,0 }  },
 	  {  { 0x0021 }, { 0,0,1,0 }  },
diff --git a/localedata/tests-mbwc/dat_iswctype.c b/localedata/tests-mbwc/dat_iswctype.c
index ebdaedce57..7e481ff72f 100644
--- a/localedata/tests-mbwc/dat_iswctype.c
+++ b/localedata/tests-mbwc/dat_iswctype.c
@@ -210,7 +210,12 @@ TST_ISWCTYPE tst_iswctype_loc [] = {
       {	 { 0x0009, "blank"  }, { 0,0,0,0 }  },
       {	 { 0x000B, "blank"  }, { 0,0,1,0 }  },
       {	 { 0x0020, "blank"  }, { 0,0,0,0 }  },
+#ifdef SHOJI_IS_RIGHT
       {	 { 0x0000, "cntrl"  }, { 0,0,0,0 }  },
+#else
+      /* XXX U0000 has no properties at all.  */
+      {	 { 0x0000, "cntrl"  }, { 0,0,1,0 }  },
+#endif
       {	 { 0x001F, "cntrl"  }, { 0,0,0,0 }  },
       {	 { 0x0020, "cntrl"  }, { 0,0,1,0 }  },
       {	 { 0x0021, "cntrl"  }, { 0,0,1,0 }  },
@@ -531,8 +536,13 @@ TST_ISWCTYPE tst_iswctype_loc [] = {
       {	 { 0xFF66, "jkata"  }, { 0,0,0,0 }  },	   /* HALF KATA WO	 */
       {	 { 0xFF6F, "jkata"  }, { 0,0,0,0 }  },	   /* HALF KATA tu	 */
       {	 { 0x4E05, "jkanji" }, { 0,0,0,0 }  },	   /* CJK UNI.IDEO.	 */
+#ifdef SHOJI_IS_RIGHT
       /* <NO_WAIVER>: */
       {	 { 0x4E06, "jkanji" }, { 0,0,1,1 }  },	   /* CJK UNI.IDEO.NON-J */
+#else
+      /* XXX This character does not exist in EUC-JP.  */
+      {	 { 0x4E06, "jkanji" }, { 0,0,1,0 }  },	   /* CJK UNI.IDEO.NON-J */
+#endif
       {	 { 0x4E07, "jkanji" }, { 0,0,0,0 }  },	   /* CJK UNI.IDEO.	 */
       { is_last: 1 }
     }
diff --git a/localedata/tests-mbwc/dat_iswprint.c b/localedata/tests-mbwc/dat_iswprint.c
index 47fa4167da..70abb6635e 100644
--- a/localedata/tests-mbwc/dat_iswprint.c
+++ b/localedata/tests-mbwc/dat_iswprint.c
@@ -65,7 +65,9 @@ TST_ISW_LOC (PRINT, print) = {
 	{  { 0x007E }, { 0,0,0,0 }  },
 	{  { 0x007F }, { 0,0,1,0 }  },
 	{  { 0x0080 }, { 0,0,1,0 }  }, /* 20 */
+#ifdef NO_WAIVER
 	{  { 0x3042 }, { 0,0,1,0 }  }, /* <WAIVER> */
+#endif
 	{ is_last: 1 }			/* Last element.  */
       }
   },
@@ -120,4 +122,3 @@ TST_ISW_LOC (PRINT, print) = {
   },
   {   TST_ISW_REC (end, print) }
 };
-
diff --git a/localedata/tests-mbwc/dat_swscanf.c b/localedata/tests-mbwc/dat_swscanf.c
index 24274401ba..cce5b336e6 100644
--- a/localedata/tests-mbwc/dat_swscanf.c
+++ b/localedata/tests-mbwc/dat_swscanf.c
@@ -33,7 +33,6 @@ TST_SWSCANF tst_swscanf_loc [] =
 	},
       },
       /*------------------------ 02 -----------------------*/
-      /* <NO_WAIVER> x 2 */
       { { {
 	0x00E4, 0x00C4, 0x0000		       /* "äÄ" */
       },
@@ -60,8 +59,13 @@ TST_SWSCANF tst_swscanf_loc [] =
       },
 	  L"1%d:2%d:3%d:4%d:5%d:6%d:7%d:8%d:9%d", 0
       },
-	{ 1,EINVAL,1,EOF,
+#ifdef SHOJI_IS_RIGHT
+	{ 1,EINVAL,1,WEOF,
+	  0,0,0,0,"", { 0x0000 },
+#else
+	{ 0,0,1,0,
 	  0,0,0,0,"", { 0x0000 },
+#endif
 	},
       },
       /*---------------------------------------------------*/
@@ -136,7 +140,6 @@ TST_SWSCANF tst_swscanf_loc [] =
 	},
       },
       /*------------------------ 03 -----------------------*/
-      /* <NO_WAIVER> */
       { { {
 	0x0031,			0x003A,
 	0x0030,			0x003A,
@@ -144,12 +147,16 @@ TST_SWSCANF tst_swscanf_loc [] =
 	0x0061,			0x003A,
 	0x0063, 0x0064, 0x0000, 0x0000,
       },
-	  L"%2$d:%1$u:%f:%c:%s", 0
+	  L"%2$d:%1$u:%3$f:%4$c:%5$s", 0
       },
 	{ 1,0,1,5,
 	  0, 1, 3.9, 'a', "cd", { 0x0000 }
 	},
       },
+#ifdef SHOJI_IS_RIGHT
+      /* XXX This test does not make sense.  The format string is
+	 L"\x1\x2\x25\x53" and it is supposed to match the words
+	 0x30A2, 0x30A4, 0x0001.  */
       /*------------------------ 04 -----------------------*/
       /* <NO_WAIVER> x 2 */
       { { {
@@ -161,6 +168,7 @@ TST_SWSCANF tst_swscanf_loc [] =
 	  0,0,0,0,"", { 0x0000 }
 	},
       },
+#endif
       /*---------------------------------------------------*/
       { is_last: 1}	/* Last element.  */