X-Git-Url: http://erislabs.net/gitweb/?a=blobdiff_plain;f=lib%2Fgen-uni-tables.c;h=0a1731751608707dc7c9d18e698907c40b974fb9;hb=665a8a0fca396d67dfed85580a48178df8b4a60e;hp=b949a5bf2f98250714b1bb9141317425ab1aa5a2;hpb=7ef2788bd0cfc453fab33250a85881863640dd7c;p=gnulib.git

diff --git a/lib/gen-uni-tables.c b/lib/gen-uni-tables.c
index b949a5bf2..0a1731751 100644
--- a/lib/gen-uni-tables.c
+++ b/lib/gen-uni-tables.c
@@ -1,5 +1,6 @@
 /* Generate Unicode conforming character classification tables and
-   Line Break Properties tables from a UnicodeData file.
+   line break properties tables and word break property tables and
+   case mapping tables from a UnicodeData file.
    Copyright (C) 2000-2002, 2004, 2007-2009 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2000-2002.
 
@@ -737,7 +738,7 @@ output_predicate (const char *filename, bool (*predicate) (unsigned int), const
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd * sizeof (int) / sizeof (short) + %5zd",
+	fprintf (stream, " %5zu * sizeof (int) / sizeof (short) + %5zu",
 		 1 + t.level1_size, (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
         fprintf (stream, ",");
@@ -757,7 +758,7 @@ output_predicate (const char *filename, bool (*predicate) (unsigned int), const
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd + %5zd * sizeof (short) / sizeof (int) + %5zd",
+	fprintf (stream, " %5zu + %5zu * sizeof (short) / sizeof (int) + %5zu",
 		 1 + t.level1_size, t.level2_size << t.q, (offset - level3_offset) / sizeof (uint32_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -1055,7 +1056,7 @@ output_category (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -1075,7 +1076,7 @@ output_category (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -1206,7 +1207,7 @@ output_combclass (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -1226,7 +1227,7 @@ output_combclass (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -1545,7 +1546,7 @@ output_bidi_category (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -1565,7 +1566,7 @@ output_bidi_category (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -1752,7 +1753,7 @@ output_decimal_digit (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -1772,7 +1773,7 @@ output_decimal_digit (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -1939,7 +1940,7 @@ output_digit (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -1959,7 +1960,7 @@ output_digit (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -2195,7 +2196,7 @@ output_numeric (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -2215,7 +2216,7 @@ output_numeric (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -2433,7 +2434,7 @@ output_mirror (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -2453,7 +2454,7 @@ output_mirror (const char *filename, const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (int32_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -3753,7 +3754,7 @@ output_scripts (const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -3773,7 +3774,7 @@ output_scripts (const char *version)
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -4429,7 +4430,7 @@ output_ident_category (const char *filename, int (*predicate) (unsigned int), co
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level2_offset) / sizeof (uint32_t));
       if (i+1 < t.level1_size)
 	fprintf (stream, ",");
@@ -4449,7 +4450,7 @@ output_ident_category (const char *filename, int (*predicate) (unsigned int), co
       if (offset == 0)
 	fprintf (stream, " %5d", -1);
       else
-	fprintf (stream, " %5zd",
+	fprintf (stream, " %5zu",
 		 (offset - level3_offset) / sizeof (uint8_t));
       if (i+1 < t.level2_size << t.q)
 	fprintf (stream, ",");
@@ -6109,9 +6110,13 @@ output_lbp (FILE *stream1, FILE *stream2)
       if (i > 0 && (i % 8) == 0)
 	fprintf (stream2, "\n   ");
       offset = ((uint32_t *) (t.result + level1_offset))[i];
-      fprintf (stream2, " %5zd%s",
-	       offset == 0 ? -1 : (offset - level2_offset) / sizeof (uint32_t),
-	       (i+1 < t.level1_size ? "," : ""));
+      if (offset == 0)
+	fprintf (stream2, " %5d", -1);
+      else
+	fprintf (stream2, " %5zu",
+		 (offset - level2_offset) / sizeof (uint32_t));
+      if (i+1 < t.level1_size)
+	fprintf (stream2, ",");
     }
   if (t.level1_size > 8)
     fprintf (stream2, "\n ");
@@ -6125,9 +6130,13 @@ output_lbp (FILE *stream1, FILE *stream2)
       if (i > 0 && (i % 8) == 0)
 	fprintf (stream2, "\n   ");
       offset = ((uint32_t *) (t.result + level2_offset))[i];
-      fprintf (stream2, " %5zd%s",
-	       offset == 0 ? -1 : (offset - level3_offset) / sizeof (uint8_t),
-	       (i+1 < t.level2_size << t.q ? "," : ""));
+      if (offset == 0)
+	fprintf (stream2, " %5d", -1);
+      else
+	fprintf (stream2, " %5zu",
+		 (offset - level3_offset) / sizeof (unsigned char));
+      if (i+1 < t.level2_size << t.q)
+	fprintf (stream2, ",");
     }
   if (t.level2_size << t.q > 8)
     fprintf (stream2, "\n ");
@@ -6614,9 +6623,13 @@ output_wbp (FILE *stream)
       if (i > 0 && (i % 8) == 0)
 	fprintf (stream, "\n   ");
       offset = ((uint32_t *) (t.result + level1_offset))[i];
-      fprintf (stream, " %5zd%s",
-	       offset == 0 ? -1 : (offset - level2_offset) / sizeof (uint32_t),
-	       (i+1 < t.level1_size ? "," : ""));
+      if (offset == 0)
+	fprintf (stream, " %5d", -1);
+      else
+	fprintf (stream, " %5zu",
+		 (offset - level2_offset) / sizeof (uint32_t));
+      if (i+1 < t.level1_size)
+	fprintf (stream, ",");
     }
   if (t.level1_size > 8)
     fprintf (stream, "\n ");
@@ -6630,9 +6643,13 @@ output_wbp (FILE *stream)
       if (i > 0 && (i % 8) == 0)
 	fprintf (stream, "\n   ");
       offset = ((uint32_t *) (t.result + level2_offset))[i];
-      fprintf (stream, " %5zd%s",
-	       offset == 0 ? -1 : (offset - level3_offset) / sizeof (uint8_t),
-	       (i+1 < t.level2_size << t.q ? "," : ""));
+      if (offset == 0)
+	fprintf (stream, " %5d", -1);
+      else
+	fprintf (stream, " %5zu",
+		 (offset - level3_offset) / sizeof (unsigned char));
+      if (i+1 < t.level2_size << t.q)
+	fprintf (stream, ",");
     }
   if (t.level2_size << t.q > 8)
     fprintf (stream, "\n ");
@@ -6723,6 +6740,208 @@ output_wbrk_tables (const char *filename, const char *version)
 
 /* ========================================================================= */
 
+/* Output the test for a simple character mapping table to the given file.  */
+
+static void
+output_simple_mapping_test (const char *filename,
+			    const char *function_name,
+			    unsigned int (*func) (unsigned int),
+			    const char *version)
+{
+  FILE *stream;
+  bool need_comma;
+  unsigned int ch;
+
+  stream = fopen (filename, "w");
+  if (stream == NULL)
+    {
+      fprintf (stderr, "cannot open '%s' for writing\n", filename);
+      exit (1);
+    }
+
+  fprintf (stream, "/* DO NOT EDIT! GENERATED AUTOMATICALLY! */\n");
+  fprintf (stream, "/* Test the Unicode character mapping functions.\n");
+  fprintf (stream, "   Copyright (C) 2009 Free Software Foundation, Inc.\n");
+  fprintf (stream, "\n");
+  fprintf (stream, "   This program is free software: you can redistribute it and/or modify\n");
+  fprintf (stream, "   it under the terms of the GNU General Public License as published by\n");
+  fprintf (stream, "   the Free Software Foundation; either version 3 of the License, or\n");
+  fprintf (stream, "   (at your option) any later version.\n");
+  fprintf (stream, "\n");
+  fprintf (stream, "   This program is distributed in the hope that it will be useful,\n");
+  fprintf (stream, "   but WITHOUT ANY WARRANTY; without even the implied warranty of\n");
+  fprintf (stream, "   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n");
+  fprintf (stream, "   GNU General Public License for more details.\n");
+  fprintf (stream, "\n");
+  fprintf (stream, "   You should have received a copy of the GNU General Public License\n");
+  fprintf (stream, "   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */\n");
+  fprintf (stream, "\n");
+  fprintf (stream, "/* Generated automatically by gen-case.c for Unicode %s.  */\n",
+	   version);
+  fprintf (stream, "\n");
+  fprintf (stream, "#include \"test-mapping-part1.h\"\n");
+  fprintf (stream, "\n");
+
+  need_comma = false;
+  for (ch = 0; ch < 0x110000; ch++)
+    {
+      unsigned int value = func (ch);
+
+      if (value != ch)
+	{
+	  if (need_comma)
+	    fprintf (stream, ",\n");
+	  fprintf (stream, "    { 0x%04X, 0x%04X }", ch, value);
+	  need_comma = true;
+	}
+    }
+  if (need_comma)
+    fprintf (stream, "\n");
+
+  fprintf (stream, "\n");
+  fprintf (stream, "#define MAP(c) %s (c)\n", function_name);
+  fprintf (stream, "#include \"test-mapping-part2.h\"\n");
+
+  if (ferror (stream) || fclose (stream))
+    {
+      fprintf (stderr, "error writing to '%s'\n", filename);
+      exit (1);
+    }
+}
+
+/* Construction of sparse 3-level tables.  */
+#define TABLE mapping_table
+#define ELEMENT int32_t
+#define DEFAULT 0
+#define xmalloc malloc
+#define xrealloc realloc
+#include "3level.h"
+
+/* Output a simple character mapping table to the given file.  */
+
+static void
+output_simple_mapping (const char *filename,
+		       unsigned int (*func) (unsigned int),
+		       const char *version)
+{
+  FILE *stream;
+  unsigned int ch, i;
+  struct mapping_table t;
+  unsigned int level1_offset, level2_offset, level3_offset;
+
+  stream = fopen (filename, "w");
+  if (stream == NULL)
+    {
+      fprintf (stderr, "cannot open '%s' for writing\n", filename);
+      exit (1);
+    }
+
+  fprintf (stream, "/* DO NOT EDIT! GENERATED AUTOMATICALLY! */\n");
+  fprintf (stream, "/* Simple character mapping of Unicode characters.  */\n");
+  fprintf (stream, "/* Generated automatically by gen-case.c for Unicode %s.  */\n",
+	   version);
+
+  t.p = 7;
+  t.q = 9;
+  mapping_table_init (&t);
+
+  for (ch = 0; ch < 0x110000; ch++)
+    {
+      int value = (int) func (ch) - (int) ch;
+
+      mapping_table_add (&t, ch, value);
+    }
+
+  mapping_table_finalize (&t);
+
+  /* Offsets in t.result, in memory of this process.  */
+  level1_offset =
+    5 * sizeof (uint32_t);
+  level2_offset =
+    5 * sizeof (uint32_t)
+    + t.level1_size * sizeof (uint32_t);
+  level3_offset =
+    5 * sizeof (uint32_t)
+    + t.level1_size * sizeof (uint32_t)
+    + (t.level2_size << t.q) * sizeof (uint32_t);
+
+  for (i = 0; i < 5; i++)
+    fprintf (stream, "#define mapping_header_%d %d\n", i,
+	     ((uint32_t *) t.result)[i]);
+  fprintf (stream, "static const\n");
+  fprintf (stream, "struct\n");
+  fprintf (stream, "  {\n");
+  fprintf (stream, "    int level1[%zu];\n", t.level1_size);
+  fprintf (stream, "    short level2[%zu << %d];\n", t.level2_size, t.q);
+  fprintf (stream, "    int level3[%zu << %d];\n", t.level3_size, t.p);
+  fprintf (stream, "  }\n");
+  fprintf (stream, "u_mapping =\n");
+  fprintf (stream, "{\n");
+  fprintf (stream, "  {");
+  if (t.level1_size > 8)
+    fprintf (stream, "\n   ");
+  for (i = 0; i < t.level1_size; i++)
+    {
+      uint32_t offset;
+      if (i > 0 && (i % 8) == 0)
+	fprintf (stream, "\n   ");
+      offset = ((uint32_t *) (t.result + level1_offset))[i];
+      if (offset == 0)
+	fprintf (stream, " %5d", -1);
+      else
+	fprintf (stream, " %5zu",
+		 (offset - level2_offset) / sizeof (uint32_t));
+      if (i+1 < t.level1_size)
+	fprintf (stream, ",");
+    }
+  if (t.level1_size > 8)
+    fprintf (stream, "\n ");
+  fprintf (stream, " },\n");
+  fprintf (stream, "  {");
+  if (t.level2_size << t.q > 8)
+    fprintf (stream, "\n   ");
+  for (i = 0; i < t.level2_size << t.q; i++)
+    {
+      uint32_t offset;
+      if (i > 0 && (i % 8) == 0)
+	fprintf (stream, "\n   ");
+      offset = ((uint32_t *) (t.result + level2_offset))[i];
+      if (offset == 0)
+	fprintf (stream, " %5d", -1);
+      else
+	fprintf (stream, " %5zu",
+		 (offset - level3_offset) / sizeof (int32_t));
+      if (i+1 < t.level2_size << t.q)
+	fprintf (stream, ",");
+    }
+  if (t.level2_size << t.q > 8)
+    fprintf (stream, "\n ");
+  fprintf (stream, " },\n");
+  fprintf (stream, "  {");
+  if (t.level3_size << t.p > 8)
+    fprintf (stream, "\n   ");
+  for (i = 0; i < t.level3_size << t.p; i++)
+    {
+      if (i > 0 && (i % 8) == 0)
+	fprintf (stream, "\n   ");
+      fprintf (stream, " %5d", ((int32_t *) (t.result + level3_offset))[i]);
+      if (i+1 < t.level3_size << t.p)
+	fprintf (stream, ",");
+    }
+  if (t.level3_size << t.p > 8)
+    fprintf (stream, "\n ");
+  fprintf (stream, " }\n");
+  fprintf (stream, "};\n");
+
+  if (ferror (stream) || fclose (stream))
+    {
+      fprintf (stderr, "error writing to '%s'\n", filename);
+      exit (1);
+    }
+}
+
+/* ========================================================================= */
+
 int
 main (int argc, char * argv[])
 {
@@ -6792,6 +7011,13 @@ main (int argc, char * argv[])
   debug_output_org_wbrk_tables ("uniwbrk/wbrkprop_org.txt");
   output_wbrk_tables ("uniwbrk/wbrkprop.h", version);
 
+  output_simple_mapping_test ("../tests/unicase/test-uc_toupper.c", "uc_toupper", to_upper, version);
+  output_simple_mapping_test ("../tests/unicase/test-uc_tolower.c", "uc_tolower", to_lower, version);
+  output_simple_mapping_test ("../tests/unicase/test-uc_totitle.c", "uc_totitle", to_title, version);
+  output_simple_mapping ("unicase/toupper.h", to_upper, version);
+  output_simple_mapping ("unicase/tolower.h", to_lower, version);
+  output_simple_mapping ("unicase/totitle.h", to_title, version);
+
   return 0;
 }