Merge pull request #56 from rth/unicode-10

Upgrade to Unicode 10
unicode-rs · Oct 22, 2019 · 0e09a6b · 0e09a6b
2 parents e2e111e + e765bea
commit 0e09a6b
Show file tree

Hide file tree

Showing 4 changed files with 1,276 additions and 1,216 deletions.
diff --git a/scripts/unicode.py b/scripts/unicode.py
@@ -59,7 +59,7 @@ def is_surrogate(n):
 
 def fetch(f):
     if not os.path.exists(os.path.basename(f)):
-        os.system("curl -O http://www.unicode.org/Public/9.0.0/ucd/%s"
+        os.system("curl -O http://www.unicode.org/Public/10.0.0/ucd/%s"
                   % f)
 
     if not os.path.exists(os.path.basename(f)):
@@ -357,4 +357,4 @@ def emit_break_module(f, break_table, break_cats, name):
         for cat in sentence_cats:
             sentence_table.extend([(x, y, cat) for (x, y) in sentence_cats[cat]])
         sentence_table.sort(key=lambda w: w[0])
-        emit_break_module(rf, sentence_table, list(sentence_cats.keys()), "sentence")
+        emit_break_module(rf, sentence_table, list(sentence_cats.keys()), "sentence")
diff --git a/scripts/unicode_gen_breaktests.py b/scripts/unicode_gen_breaktests.py
@@ -172,7 +172,7 @@ def create_grapheme_data(f):
     stype = "&'static [(&'static str, &'static [&'static str])]"
     dtype = "&'static [(&'static str, &'static [&'static str], &'static [&'static str])]"
     f.write("    // official Unicode test data\n")
-    f.write("    // http://www.unicode.org/Public/9.0.0/ucd/auxiliary/GraphemeBreakTest.txt\n")
+    f.write("    // http://www.unicode.org/Public/10.0.0/ucd/auxiliary/GraphemeBreakTest.txt\n")
     unicode.emit_table(f, "TEST_SAME", test_same, stype, True, showfun, True)
     unicode.emit_table(f, "TEST_DIFF", test_diff, dtype, True, showfun, True)
 
@@ -187,7 +187,7 @@ def create_words_data(f):
 
     wtype = "&'static [(&'static str, &'static [&'static str])]"
     f.write("    // official Unicode test data\n")
-    f.write("    // http://www.unicode.org/Public/9.0.0/ucd/auxiliary/WordBreakTest.txt\n")
+    f.write("    // http://www.unicode.org/Public/10.0.0/ucd/auxiliary/WordBreakTest.txt\n")
     unicode.emit_table(f, "TEST_WORD", test, wtype, True, showfun, True)
 
 def create_sentence_data(f):
@@ -201,7 +201,7 @@ def create_sentence_data(f):
 
     wtype = "&'static [(&'static str, &'static [&'static str])]"
     f.write("    // official Unicode test data\n")
-    f.write("    // http://www.unicode.org/Public/UNIDATA/auxiliary/SentenceBreakTest.txt\n")
+    f.write("    // http://www.unicode.org/Public/10.0.0/ucd/auxiliary/SentenceBreakTest.txt\n")
     unicode.emit_table(f, "TEST_SENTENCE", test, wtype, True, showfun, True)
 
 if __name__ == "__main__":