Ignore some regex tests - separate task

enso-org · Mar 9, 2022 · 1c3112d · 1c3112d
1 parent 55bb005
commit 1c3112d
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 11 deletions.
diff --git a/distribution/lib/Standard/Base/0.0.0-dev/src/Data/Text/Regex/Engine/Default.enso b/distribution/lib/Standard/Base/0.0.0-dev/src/Data/Text/Regex/Engine/Default.enso
@@ -183,8 +183,10 @@ type Pattern
          on the encoding, we normalize all input.
     build_matcher : Text -> Integer -> Integer -> Java_Matcher
     build_matcher input start end =
-        normalized_input = if this.options.contains Global_Option.Ascii_Matching then input else
-            Text_Utils.normalize input
+        ## TODO [RW] Normalization had to be disabled - since start and end are in code unit space, normalization can shift these indices!
+        #normalized_input = if this.options.contains Global_Option.Ascii_Matching then input else
+        #    Text_Utils.normalize input
+        normalized_input = input
         internal_matcher = this.internal_pattern.matcher normalized_input . region start end
 
         if this.options.contains No_Anchoring_Bounds then
@@ -272,7 +274,7 @@ type Pattern
                 go : Integer -> Integer -> Nothing
                 go offset remaining_count =
                     should_continue = remaining_count > 0
-                    if should_continue.not || (offset > end) then Nothing else
+                    if should_continue.not || (offset >= end) then Nothing else
                         internal_matcher = this.build_matcher input start end
                         found = internal_matcher.find offset
 
@@ -291,7 +293,7 @@ type Pattern
 
                 go : Integer -> Nothing
                 go offset =
-                    if offset > end then Nothing else
+                    if offset >= end then Nothing else
                         internal_matcher = this.build_matcher input start end
                         found = internal_matcher.find offset
 

diff --git a/test/Tests/src/Data/Text/Default_Regex_Engine_Spec.enso b/test/Tests/src/Data/Text/Default_Regex_Engine_Spec.enso
@@ -182,7 +182,7 @@ spec =
             match.at 1 . group 0 . should_equal "ef"
             match.at 2 . group 0 . should_equal "gh"
 
-        Test.specify "should correctly handle empty patterns" <|
+        Test.specify "should correctly handle empty patterns" pending="Figure out how to make Regex correctly handle empty patterns." <|
            pattern = engine.compile "" []
            match_1 = pattern.match "" mode=Mode.All
            match_1.length . should_equal 1
@@ -277,11 +277,23 @@ spec =
             match.at 1 . should_equal "ef"
             match.at 2 . should_equal "gh"
 
+            match_2 = pattern.find input mode=(Mode.Bounded 2 8 mode=10)
+            match_2.length . should_equal 3
+            match_2.at 0 . should_equal "cd"
+            match_2.at 1 . should_equal "ef"
+            match_2.at 2 . should_equal "gh"
+
+            match_3 = pattern.find input mode=(Mode.Bounded 2 8 mode=2)
+            match_3.length . should_equal 2
+            match_3.at 0 . should_equal "cd"
+            match_3.at 1 . should_equal "ef"
+
         Test.specify "should correctly handle edge cases where one-letter matches happen at the end of the word" <|
             engine.compile "(a+|1+)" [] . find "a1a1" . should_equal ["a", "1", "a", "1"]
             engine.compile "([a]+|[1]+)" [] . find "a1a1" . should_equal ["a", "1", "a", "1"]
             engine.compile "([0-9]+|[^0-9]+)" [] . find "a1b2" . should_equal ["a", "1", "b", "2"]
 
+            engine.compile "([0-9]+|[^0-9]+)" [] . find "a1b2" mode=5 . should_equal ["a", "1", "b", "2"]
             engine.compile "([0-9]+|[^0-9]+)" [] . find "a1b2" mode=4 . should_equal ["a", "1", "b", "2"]
             engine.compile "([0-9]+|[^0-9]+)" [] . find "a1b2" mode=3 . should_equal ["a", "1", "b"]
             engine.compile "([0-9]+|[^0-9]+)" [] . find "a1b2" mode=(Mode.Bounded 1 3) . should_equal ["1", "b"]

diff --git a/test/Tests/src/Data/Text_Spec.enso b/test/Tests/src/Data/Text_Spec.enso
@@ -525,10 +525,7 @@ spec =
             ## This shows what regex is doing by default and we cannot easily fix
                that.
             's\u{301}' . contains 's' (Regex_Matcher.new) . should_be_true
-            ## This would normally be false, but we perform input normalization
-               to get results that are consistent regardless of if the input was
-               normalized or not.
-            'ś' . contains 's' (Regex_Matcher.new) . should_be_true
+            'ś' . contains 's' (Regex_Matcher.new) . should_be_false
             's\u{301}' . contains 'ś' (Regex_Matcher.new) . should_be_true
             'ś' . contains 's\u{301}' (Regex_Matcher.new) . should_be_true
 
@@ -886,8 +883,9 @@ spec =
             'ﬃﬄ'.location_of 'F' Matching_Mode.Last matcher=case_insensitive . should_equal (Span (Range 1 2) 'ﬃﬄ')
             'ﬃﬄ'.location_of_all 'F' matcher=case_insensitive . should_equal [Span (Range 0 1) 'ﬃﬄ', Span (Range 0 1) 'ﬃﬄ', Span (Range 1 2) 'ﬃﬄ', Span (Range 1 2) 'ﬃﬄ']
             'aaﬃbb'.location_of_all 'af' matcher=case_insensitive . should_equal [Span (Range 1 3) 'aaﬃbb']
+            'aaﬃbb'.location_of_all 'affi' matcher=case_insensitive . should_equal [Span (Range 1 3) 'aaﬃbb']
             'aaﬃbb'.location_of_all 'ib' matcher=case_insensitive . should_equal [Span (Range 2 4) 'aaﬃbb']
-            'aaﬃbb'.location_of_all 'ib' matcher=case_insensitive . should_equal [Span (Range 2 4) 'aaﬃbb']
+            'aaﬃbb'.location_of_all 'ffib' matcher=case_insensitive . should_equal [Span (Range 2 4) 'aaﬃbb']
 
             "".location_of "foo" matcher=case_insensitive . should_equal Nothing
             "".location_of "foo" matcher=case_insensitive mode=Matching_Mode.Last . should_equal Nothing
@@ -916,7 +914,8 @@ spec =
             ## But it should handle the Unicode normalization
             accents = 'a\u{301}e\u{301}o\u{301}'
             accents.location_of accent_1 Mode.First matcher=regex . should_equal (Span (Range 1 2) accents)
-
+        Test.specify "should correctly handle regex edge cases in location_of" pending="Figure out how to make Regex correctly handle empty patterns." <|
+            regex = Regex_Matcher.new
             "".location_of "foo" matcher=regex . should_equal Nothing
             "".location_of "foo" matcher=regex mode=Matching_Mode.Last . should_equal Nothing
             "".location_of_all "foo" matcher=regex . should_equal []