Minor enhancement to Word generation of internal regular expression when characters are 2 consecutive chars

4 files changed, 38 insertions, 23 deletions

diff --git a/CHANGES b/CHANGES
index bbf643d..48e8fba 100644
--- a/CHANGES
+++ b/CHANGES
@@ -38,6 +38,9 @@ Version 3.0.7 -
 
   In previous versions, the second call to `bool()` would return `False`.
 
+- Minor enhancement to Word generation of internal regular expression, to
+  emit consecutive characters in range, such as "ab", as "ab", not "a-b".
+
 
 Version 3.0.6 -
 ---------------
diff --git a/pyparsing/__init__.py b/pyparsing/__init__.py
index e152fdd..e83ab52 100644
--- a/pyparsing/__init__.py
+++ b/pyparsing/__init__.py
@@ -126,7 +126,7 @@ class version_info(NamedTuple):
 
 
 __version_info__ = version_info(3, 0, 7, "final", 0)
-__version_time__ = "13 Nov 2021 18:44 UTC"
+__version_time__ = "03 Dec 2021 11:21 UTC"
 __version__ = __version_info__.__version__
 __versionTime__ = __version_time__
 __author__ = "Paul McGuire <ptmcg.gm+pyparsing@gmail.com>"
diff --git a/pyparsing/util.py b/pyparsing/util.py
index 1309ad6..34ce092 100644
--- a/pyparsing/util.py
+++ b/pyparsing/util.py
@@ -213,9 +213,10 @@ def _collapse_string_to_ranges(
             if first == last:
                 ret.append(escape_re_range_char(first))
             else:
+                sep = "" if ord(last) == ord(first) + 1 else "-"
                 ret.append(
-                    "{}-{}".format(
-                        escape_re_range_char(first), escape_re_range_char(last)
+                    "{}{}{}".format(
+                        escape_re_range_char(first), sep, escape_re_range_char(last)
                     )
                 )
     else:
diff --git a/tests/test_unit.py b/tests/test_unit.py
index dbfdfe5..95183a3 100644
--- a/tests/test_unit.py
+++ b/tests/test_unit.py
@@ -8177,28 +8177,39 @@ class Test02_WithoutPackrat(ppt.TestParseResultsAsserts, TestCase):
                 "__diag__.{} not set to True".format(diag_name),
             )
 
-    def testWordInternalReRangesKnownSets(self):
-        self.assertEqual(
-            "[!-~]+",
-            pp.Word(pp.printables).reString,
-            "failed to generate correct internal re",
-        )
-        self.assertEqual(
-            "[0-9A-Za-z]+",
-            pp.Word(pp.alphanums).reString,
-            "failed to generate correct internal re",
-        )
-        self.assertEqual(
-            "[!-~¡-ÿ]+",
-            pp.Word(pp.pyparsing_unicode.Latin1.printables).reString,
-            "failed to generate correct internal re",
-        )
-        self.assertEqual(
-            "[À-ÖØ-öø-ÿ]+",
-            pp.Word(pp.alphas8bit).reString,
-            "failed to generate correct internal re",
+    def testWordInternalReRangeWithConsecutiveChars(self):
+        self.assertParseAndCheckList(
+            pp.Word("ABCDEMNXYZ"),
+            "ABCDEMNXYZABCDEMNXYZABCDEMNXYZ",
+            ["ABCDEMNXYZABCDEMNXYZABCDEMNXYZ"]
         )
 
+    def testWordInternalReRangesKnownSet(self):
+        tests = [
+            ("ABCDEMNXYZ", "[A-EMNX-Z]+"),
+            (pp.printables, "[!-~]+"),
+            (pp.alphanums, "[0-9A-Za-z]+"),
+            (pp.pyparsing_unicode.Latin1.printables, "[!-~¡-ÿ]+"),
+            (pp.pyparsing_unicode.Latin1.alphanums, "[0-9A-Za-zª²³µ¹ºÀ-ÖØ-öø-ÿ]+"),
+            (pp.alphas8bit, "[À-ÖØ-öø-ÿ]+"),
+        ]
+        failed = []
+        for word_string, expected_re in tests:
+            try:
+                msg = "failed to generate correct internal re for {!r}".format(word_string)
+                resultant_re = pp.Word(word_string).reString
+                self.assertEqual(
+                    expected_re,
+                    resultant_re,
+                    msg + "; expected {!r} got {!r}".format(expected_re, resultant_re)
+                )
+            except AssertionError:
+                failed.append(msg)
+
+        if failed:
+            print("Errors:\n{}".format("\n".join(failed)))
+            self.fail("failed to generate correct internal re's")
+
     def testWordInternalReRanges(self):
         import random