AMK's regex test suite

2 files changed, 331 insertions, 0 deletions

diff --git a/Lib/test/regex_tests.py b/Lib/test/regex_tests.py
new file mode 100644
index 0000000000..70ecdab9d3
--- /dev/null
+++ b/Lib/test/regex_tests.py
@@ -0,0 +1,283 @@
+
+# Regex test suite and benchmark suite v1.5a2
+# Due to the use of r"aw" strings, this file will 
+# only work with Python 1.5 or higher.
+
+# The 3 possible outcomes for each pattern
+[SUCCEED, FAIL, SYNTAX_ERROR] = range(3)
+
+# Benchmark suite (needs expansion)
+#
+# The benchmark suite does not test correctness, just speed.  The
+# first element of each tuple is the regex pattern; the second is a 
+# string to match it against.  The benchmarking code will embed the
+# second string inside several sizes of padding, to test how regex 
+# matching performs on large strings.
+
+benchmarks = [
+	('Python', 'Python'),                     # Simple text literal
+	('.*Python', 'Python'),                   # Bad text literal
+        ('.*Python.*', 'Python'),                 # Worse text literal
+	('.*\\(Python\\)', 'Python'),             # Bad text literal with grouping
+
+	('(Python\\|Perl\\|Tcl', 'Perl'),          # Alternation
+	('\\(Python\\|Perl\\|Tcl\\)', 'Perl'),     # Grouped alternation
+	('\\(Python\\)\\1', 'PythonPython'),       # Backreference
+#	('\\([0a-z][a-z]*,\\)+', 'a5,b7,c9,'),     # Disable the fastmap optimization
+	('\\([a-z][a-z0-9]*,\\)+', 'a5,b7,c9,')    # A few sets
+]
+
+# Test suite (for verifying correctness)
+#
+# The test suite is a list of 5- or 3-tuples.  The 5 parts of a
+# complete tuple are:
+# element 0: a string containing the pattern
+#         1: the string to match against the pattern
+#         2: the expected result (SUCCEED, FAIL, SYNTAX_ERROR)
+#         3: a string that will be eval()'ed to produce a test string.
+#            This is an arbitrary Python expression; the available
+#            variables are "found" (the whole match), and "g1", "g2", ...
+#            up to "g10" contain the contents of each group, or the
+#            string 'None' if the group wasn't given a value.
+#         4: The expected result of evaluating the expression.
+#            If the two don't match, an error is reported.
+#
+# If the regex isn't expected to work, the latter two elements can be omitted.
+
+tests = [
+('abc', 'abc', SUCCEED,
+ 'found', 'abc'),
+('abc', 'xbc', FAIL),
+('abc', 'axc', FAIL),
+('abc', 'abx', FAIL),
+('abc', 'xabcy', SUCCEED,
+ 'found', 'abc'),
+('abc', 'ababc', SUCCEED,
+ 'found', 'abc'),
+('ab*c', 'abc', SUCCEED,
+ 'found', 'abc'),
+('ab*bc', 'abc', SUCCEED,
+ 'found', 'abc'),
+('ab*bc', 'abbc', SUCCEED,
+ 'found', 'abbc'),
+('ab*bc', 'abbbbc', SUCCEED,
+ 'found', 'abbbbc'),
+('ab+bc', 'abbc', SUCCEED,
+ 'found', 'abbc'),
+('ab+bc', 'abc', FAIL),
+('ab+bc', 'abq', FAIL),
+('ab+bc', 'abbbbc', SUCCEED,
+ 'found', 'abbbbc'),
+('ab?bc', 'abbc', SUCCEED,
+ 'found', 'abbc'),
+('ab?bc', 'abc', SUCCEED,
+ 'found', 'abc'),
+('ab?bc', 'abbbbc', FAIL),
+('ab?c', 'abc', SUCCEED,
+ 'found', 'abc'),
+('^abc$', 'abc', SUCCEED,
+ 'found', 'abc'),
+('^abc$', 'abcc', FAIL),
+('^abc', 'abcc', SUCCEED,
+ 'found', 'abc'),
+('^abc$', 'aabc', FAIL),
+('abc$', 'aabc', SUCCEED,
+ 'found', 'abc'),
+('^', 'abc', SUCCEED,
+ 'found+"-"', '-'),
+('$', 'abc', SUCCEED,
+ 'found+"-"', '-'),
+('a.c', 'abc', SUCCEED,
+ 'found', 'abc'),
+('a.c', 'axc', SUCCEED,
+ 'found', 'axc'),
+('a.*c', 'axyzc', SUCCEED,
+ 'found', 'axyzc'),
+('a.*c', 'axyzd', FAIL),
+('a[bc]d', 'abc', FAIL),
+('a[bc]d', 'abd', SUCCEED,
+ 'found', 'abd'),
+('a[b-d]e', 'abd', FAIL),
+('a[b-d]e', 'ace', SUCCEED,
+ 'found', 'ace'),
+('a[b-d]', 'aac', SUCCEED,
+ 'found', 'ac'),
+('a[-b]', 'a-', SUCCEED,
+ 'found', 'a-'),
+('a[b-]', 'a-', SUCCEED,
+ 'found', 'a-'),
+('a[]b', '-', SYNTAX_ERROR),
+('a[', '-', SYNTAX_ERROR),
+('a\\', '-', SYNTAX_ERROR),
+('abc\\)', '-', SYNTAX_ERROR),
+('\\(abc', '-', SYNTAX_ERROR),
+('a]', 'a]', SUCCEED,
+ 'found', 'a]'),
+('a[]]b', 'a]b', SUCCEED,
+ 'found', 'a]b'),
+('a[^bc]d', 'aed', SUCCEED,
+ 'found', 'aed'),
+('a[^bc]d', 'abd', FAIL),
+('a[^-b]c', 'adc', SUCCEED,
+ 'found', 'adc'),
+('a[^-b]c', 'a-c', FAIL),
+('a[^]b]c', 'a]c', FAIL),
+('a[^]b]c', 'adc', SUCCEED,
+ 'found', 'adc'),
+('\\ba\\b', 'a-', SUCCEED,
+ '"-"', '-'),
+('\\ba\\b', '-a', SUCCEED,
+ '"-"', '-'),
+('\\ba\\b', '-a-', SUCCEED,
+ '"-"', '-'),
+('\\by\\b', 'xy', FAIL),
+('\\by\\b', 'yz', FAIL),
+('\\by\\b', 'xyz', FAIL),
+('ab\\|cd', 'abc', SUCCEED,
+ 'found', 'ab'),
+('ab\\|cd', 'abcd', SUCCEED,
+ 'found', 'ab'),
+('\\(\\)ef', 'def', SUCCEED,
+ 'found+"-"+g1', 'ef-'),
+('$b', 'b', FAIL),
+('a(b', 'a(b', SUCCEED,
+ 'found+"-"+g1', 'a(b-None'),
+('a(*b', 'ab', SUCCEED,
+ 'found', 'ab'),
+('a(*b', 'a((b', SUCCEED,
+ 'found', 'a((b'),
+('a\\\\b', 'a\\b', SUCCEED,
+ 'found', 'a\\b'),
+('\\(\\(a\\)\\)', 'abc', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'a-a-a'),
+('\\(a\\)b\\(c\\)', 'abc', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'abc-a-c'),
+('a+b+c', 'aabbabc', SUCCEED,
+ 'found', 'abc'),
+('\\(a+\\|b\\)*', 'ab', SUCCEED,
+ 'found+"-"+g1', 'ab-b'),
+('\\(a+\\|b\\)+', 'ab', SUCCEED,
+ 'found+"-"+g1', 'ab-b'),
+('\\(a+\\|b\\)?', 'ab', SUCCEED,
+ 'found+"-"+g1', 'a-a'),
+('\\)\\(', '-', SYNTAX_ERROR),
+('[^ab]*', 'cde', SUCCEED,
+ 'found', 'cde'),
+('abc', '', FAIL),
+('a*', '', SUCCEED,
+ 'found', ''),
+('a\\|b\\|c\\|d\\|e', 'e', SUCCEED,
+ 'found', 'e'),
+('\\(a\\|b\\|c\\|d\\|e\\)f', 'ef', SUCCEED,
+ 'found+"-"+g1', 'ef-e'),
+('abcd*efg', 'abcdefg', SUCCEED,
+ 'found', 'abcdefg'),
+('ab*', 'xabyabbbz', SUCCEED,
+ 'found', 'ab'),
+('ab*', 'xayabbbz', SUCCEED,
+ 'found', 'a'),
+('\\(ab\\|cd\\)e', 'abcde', SUCCEED,
+ 'found+"-"+g1', 'cde-cd'),
+('[abhgefdc]ij', 'hij', SUCCEED,
+ 'found', 'hij'),
+('^\\(ab\\|cd\\)e', 'abcde', FAIL,
+ 'xg1y', 'xy'),
+('\\(abc\\|\\)ef', 'abcdef', SUCCEED,
+ 'found+"-"+g1', 'ef-'),
+('\\(a\\|b\\)c*d', 'abcd', SUCCEED,
+ 'found+"-"+g1', 'bcd-b'),
+('\\(ab\\|ab*\\)bc', 'abc', SUCCEED,
+ 'found+"-"+g1', 'abc-a'),
+('a\\([bc]*\\)c*', 'abc', SUCCEED,
+ 'found+"-"+g1', 'abc-bc'),
+('a\\([bc]*\\)\\(c*d\\)', 'abcd', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'abcd-bc-d'),
+('a\\([bc]+\\)\\(c*d\\)', 'abcd', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'abcd-bc-d'),
+('a\\([bc]*\\)\\(c+d\\)', 'abcd', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'abcd-b-cd'),
+('a[bcd]*dcdcde', 'adcdcde', SUCCEED,
+ 'found', 'adcdcde'),
+('a[bcd]+dcdcde', 'adcdcde', FAIL),
+('\\(ab\\|a\\)b*c', 'abc', SUCCEED,
+ 'found+"-"+g1', 'abc-ab'),
+('\\(\\(a\\)\\(b\\)c\\)\\(d\\)', 'abcd', SUCCEED,
+ 'g1+"-"+g2+"-"+g3+"-"+g4', 'abc-a-b-d'),
+('[a-zA-Z_][a-zA-Z0-9_]*', 'alpha', SUCCEED,
+ 'found', 'alpha'),
+('^a\\(bc+\\|b[eh]\\)g\\|.h$', 'abh', SUCCEED,
+ 'found+"-"+g1', 'bh-None'),
+('\\(bc+d$\\|ef*g.\\|h?i\\(j\\|k\\)\\)', 'effgz', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'effgz-effgz-None'),
+('\\(bc+d$\\|ef*g.\\|h?i\\(j\\|k\\)\\)', 'ij', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'ij-ij-j'),
+('\\(bc+d$\\|ef*g.\\|h?i\\(j\\|k\\)\\)', 'effg', FAIL),
+('\\(bc+d$\\|ef*g.\\|h?i\\(j\\|k\\)\\)', 'bcdd', FAIL),
+('\\(bc+d$\\|ef*g.\\|h?i\\(j\\|k\\)\\)', 'reffgz', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'effgz-effgz-None'),
+('\\(\\(\\(\\(\\(\\(\\(\\(\\(a\\)\\)\\)\\)\\)\\)\\)\\)\\)', 'a', SUCCEED,
+ 'found', 'a'),
+('multiple words of text', 'uh-uh', FAIL),
+('multiple words', 'multiple words, yeah', SUCCEED,
+ 'found', 'multiple words'),
+('\\(.*\\)c\\(.*\\)', 'abcde', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'abcde-ab-de'),
+('(\\(.*\\), \\(.*\\))', '(a, b)', SUCCEED,
+ 'g2+"-"+g1', 'b-a'),
+('[k]', 'ab', FAIL),
+('a[-]?c', 'ac', SUCCEED,
+ 'found', 'ac'),
+('\\(abc\\)\\1', 'abcabc', SUCCEED,
+ 'g1', 'abc'),
+('\\([a-c]*\\)\\1', 'abcabc', SUCCEED,
+ 'g1', 'abc'),
+('^\\(.+\\)?B', 'AB', SUCCEED,
+ 'g1', 'A'),
+('\\(a+\\).\\1$', 'aaaaa', SUCCEED,
+ 'found+"-"+g1', 'aaaaa-aa'),
+('^\\(a+\\).\\1$', 'aaaa', FAIL),
+('\\(abc\\)\\1', 'abcabc', SUCCEED,
+ 'found+"-"+g1', 'abcabc-abc'),
+('\\([a-c]+\\)\\1', 'abcabc', SUCCEED,
+ 'found+"-"+g1', 'abcabc-abc'),
+('\\(a\\)\\1', 'aa', SUCCEED,
+ 'found+"-"+g1', 'aa-a'),
+('\\(a+\\)\\1', 'aa', SUCCEED,
+ 'found+"-"+g1', 'aa-a'),
+('\\(a+\\)+\\1', 'aa', SUCCEED,
+ 'found+"-"+g1', 'aa-a'),
+('\\(a\\).+\\1', 'aba', SUCCEED,
+ 'found+"-"+g1', 'aba-a'),
+('\\(a\\)ba*\\1', 'aba', SUCCEED,
+ 'found+"-"+g1', 'aba-a'),
+('\\(aa\\|a\\)a\\1$', 'aaa', SUCCEED,
+ 'found+"-"+g1', 'aaa-a'),
+('\\(a\\|aa\\)a\\1$', 'aaa', SUCCEED,
+ 'found+"-"+g1', 'aaa-a'),
+('\\(a+\\)a\\1$', 'aaa', SUCCEED,
+ 'found+"-"+g1', 'aaa-a'),
+('\\([abc]*\\)\\1', 'abcabc', SUCCEED,
+ 'found+"-"+g1', 'abcabc-abc'),
+('\\(a\\)\\(b\\)c\\|ab', 'ab', SUCCEED,
+ 'found+"-"+g1+"-"+g2', 'ab-None-None'),
+('\\(a\\)+x', 'aaax', SUCCEED,
+ 'found+"-"+g1', 'aaax-a'),
+('\\([ac]\\)+x', 'aacx', SUCCEED,
+ 'found+"-"+g1', 'aacx-c'),
+('\\([^/]*/\\)*sub1/', 'd:msgs/tdir/sub1/trial/away.cpp', SUCCEED,
+ 'found+"-"+g1', 'd:msgs/tdir/sub1/-tdir/'),
+('\\([^.]*\\)\\.\\([^:]*\\):[T ]+\\(.*\\)', 'track1.title:TBlah blah blah', SUCCEED,
+ 'found+"-"+g1+"-"+g2+"-"+g3', 'track1.title:TBlah blah blah-track1-title-Blah blah blah'),
+('\\([^N]*N\\)+', 'abNNxyzN', SUCCEED,
+ 'found+"-"+g1', 'abNNxyzN-xyzN'),
+('\\([^N]*N\\)+', 'abNNxyz', SUCCEED,
+ 'found+"-"+g1', 'abNN-N'),
+('\\([abc]*\\)x', 'abcx', SUCCEED,
+ 'found+"-"+g1', 'abcx-abc'),
+('\\([abc]*\\)x', 'abc', FAIL),
+('\\([xyz]*\\)x', 'abcx', SUCCEED,
+ 'found+"-"+g1', 'x-'),
+('\\(a\\)+b\\|aac', 'aac', SUCCEED,
+ 'found+"-"+g1', 'aac-None')
+]
+
diff --git a/Lib/test/test_regex.py b/Lib/test/test_regex.py
index 078ca406f4..4be6026419 100644
--- a/Lib/test/test_regex.py
+++ b/Lib/test/test_regex.py
@@ -60,3 +60,51 @@ except regex.error:
     print 'caught expected exception'
 else:
     print 'expected regex.error not raised'
+
+from regex_tests import *
+if verbose: print 'Running regex_tests test suite'
+
+for t in tests:
+    pattern=s=outcome=repl=expected=None
+    if len(t)==5:
+	pattern, s, outcome, repl, expected = t
+    elif len(t)==3:
+	pattern, s, outcome = t 
+    else:
+	raise ValueError, ('Test tuples should have 3 or 5 fields',t)
+
+    try:
+	obj=regex.compile(pattern)
+    except regex.error:
+	if outcome==SYNTAX_ERROR: pass    # Expected a syntax error
+	else: 
+	    # Regex syntax errors aren't yet reported, so for 
+	    # the official test suite they'll be quietly ignored.
+	    pass
+	    #print '=== Syntax error:', t
+    else:
+	try:
+	    result=obj.search(s)
+	except regex.error, msg:
+	    print '=== Unexpected exception', t, repr(msg)
+	if outcome==SYNTAX_ERROR:
+	    # This should have been a syntax error; forget it.
+	    pass
+	elif outcome==FAIL:
+	    if result==-1: pass   # No match, as expected
+	    else: print '=== Succeeded incorrectly', t
+	elif outcome==SUCCEED:
+	    if result!=-1:
+		# Matched, as expected, so now we compute the
+		# result string and compare it to our expected result.
+		start, end = obj.regs[0]
+		found=s[start:end]
+		groups=obj.group(1,2,3,4,5,6,7,8,9,10)
+		vardict=vars()
+		for i in range(len(groups)):
+		    vardict['g'+str(i+1)]=str(groups[i])
+		repl=eval(repl)
+		if repl!=expected:
+		    print '=== grouping error', t, repr(repl)+' should be '+repr(expected)
+	    else:
+		print '=== Failed incorrectly', t