allow to ignore line breaks in doctest output

Author: picnixz

Doc/library/doctest.rst

@@ -616,7 +616,8 @@ doctest decides whether actual output matches an example's expected output:
    sequence of whitespace within the actual output. By default, whitespace must
    match exactly. :const:`NORMALIZE_WHITESPACE` is especially useful when a line of
    expected output is very long, and you want to wrap it across multiple lines in
-   your source.
+   your source. If the expected output does not contain any whitespace, consider
+   using :data:`IGNORE_LINEBREAK` or :data:`ELLIPSIS`.
 
 
 .. index:: single: ...; in doctests
@@ -666,6 +667,31 @@ doctest decides whether actual output matches an example's expected output:
       to the module containing the exception under test.
 
 
+.. data:: IGNORE_LINEBREAK
+
+   When specified, single line breaks in the expected output are eliminated,
+   thereby allowing strings without whitespaces to span multiple lines.
+
+   .. doctest::
+      :no-trim-doctest-flags:
+
+      >>> "foobar123456"            # doctest: +IGNORE_LINEBREAK
+      'foobar
+      123456'
+
+   Consider using :data:`NORMALIZE_WHITESPACE` when strings with whitespaces
+   need to be split across multiple lines:
+
+   .. doctest::
+      :no-trim-doctest-flags:
+
+      >>> "the string to split"     # doctest: +NORMALIZE_WHITESPACE
+      'the string
+      to split'
+
+   .. versionadded:: next
+
+
 .. data:: SKIP
 
    When specified, do not run the example at all.  This can be useful in contexts

Doc/whatsnew/3.15.rst

@@ -311,6 +311,23 @@ difflib
   (Contributed by Jiahao Li in :gh:`134580`.)
 
 
+doctest
+-------
+
+* Add :data:`~doctest.IGNORE_LINEBREAK` option to allow breaking expected
+  output strings without whitespaces into multiple lines:
+
+  .. doctest::
+     :no-trim-doctest-flags:
+
+     >>> import string
+     >>> print(string.ascii_letters)  # doctest: +IGNORE_LINEBREAK
+     abcdefghijklmnopqrstuvwxyz
+     ABCDEFGHIJKLMNOPQRSTUVWXYZ
+
+  (Contributed by Bénédikt Tran in :gh:`138135`.)
+
+
 hashlib
 -------
 

Lib/doctest.py

@@ -53,6 +53,7 @@ def _test():
     'DONT_ACCEPT_TRUE_FOR_1',
     'DONT_ACCEPT_BLANKLINE',
     'NORMALIZE_WHITESPACE',
+    'IGNORE_LINEBREAK',
     'ELLIPSIS',
     'SKIP',
     'IGNORE_EXCEPTION_DETAIL',
@@ -156,6 +157,7 @@ def register_optionflag(name):
 DONT_ACCEPT_TRUE_FOR_1 = register_optionflag('DONT_ACCEPT_TRUE_FOR_1')
 DONT_ACCEPT_BLANKLINE = register_optionflag('DONT_ACCEPT_BLANKLINE')
 NORMALIZE_WHITESPACE = register_optionflag('NORMALIZE_WHITESPACE')
+IGNORE_LINEBREAK = register_optionflag('IGNORE_LINEBREAK')
 ELLIPSIS = register_optionflag('ELLIPSIS')
 SKIP = register_optionflag('SKIP')
 IGNORE_EXCEPTION_DETAIL = register_optionflag('IGNORE_EXCEPTION_DETAIL')
@@ -1751,9 +1753,18 @@ def check_output(self, want, got, optionflags):
             if got == want:
                 return True
 
+        # This flag causes doctest to ignore '\n' in `want`.
+        # Note that this can be used in conjunction with
+        # the NORMALIZE_WHITESPACE and ELLIPSIS flags.
+        if optionflags & IGNORE_LINEBREAK:
+            # `want` originally ends with '\n' so we add it back
+            want = ''.join(want.split('\n')) + '\n'
+            if got == want:
+                return True
+
         # This flag causes doctest to ignore any differences in the
         # contents of whitespace strings.  Note that this can be used
-        # in conjunction with the ELLIPSIS flag.
+        # in conjunction with the IGNORE_LINEBREAK and ELLIPSIS flags.
         if optionflags & NORMALIZE_WHITESPACE:
             got = ' '.join(got.split())
             want = ' '.join(want.split())
@@ -2268,7 +2279,7 @@ def set_unittest_reportflags(flags):
       >>> doctest.set_unittest_reportflags(ELLIPSIS)
       Traceback (most recent call last):
       ...
-      ValueError: ('Only reporting flags allowed', 8)
+      ValueError: ('Only reporting flags allowed', 16)
 
       >>> doctest.set_unittest_reportflags(old) == (REPORT_NDIFF |
       ...                                   REPORT_ONLY_FIRST_FAILURE)
@@ -2924,6 +2935,13 @@ def get(self):
                      15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26,
                      27, 28, 29]
             """,
+
+            "line break elimination": r"""
+                >>> "foobar"  # doctest: +IGNORE_LINEBREAK
+                'foo
+                bar
+                '
+            """,
            }
 
 

Lib/test/test_doctest/test_doctest.py

@@ -204,7 +204,7 @@ def test_Example(): r"""
     ...                           options={doctest.ELLIPSIS: True})
     >>> (example.source, example.want, example.exc_msg,
     ...  example.lineno, example.indent, example.options)
-    ('[].pop()\n', '', 'IndexError: pop from an empty list\n', 5, 4, {8: True})
+    ('[].pop()\n', '', 'IndexError: pop from an empty list\n', 5, 4, {16: True})
 
 The constructor normalizes the `source` string to end in a newline:
 
@@ -1396,6 +1396,55 @@ def optionflags(): r"""
     [0,   1,  2,  3,  4,  5,  6,  7,  8,  9,
     10,  11, 12, 13, 14, 15, 16, 17, 18, 19]
 
+The IGNORE_LINEBREAK flag causes all sequences of newlines to be removed:
+
+    >>> def f(x):
+    ...     '\n>>> "foobar"\n\'foo\nbar\''
+
+    >>> # Without the flag:
+    >>> test = doctest.DocTestFinder().find(f)[0]
+    >>> doctest.DocTestRunner(verbose=False).run(test)
+    ... # doctest: +ELLIPSIS
+    **********************************************************************
+    File ..., line 3, in f
+    Failed example:
+        "foobar"
+    Expected:
+        'foo
+        bar'
+    Got:
+        'foobar'
+    TestResults(failed=1, attempted=1)
+
+    >>> # With the flag:
+    >>> test = doctest.DocTestFinder().find(f)[0]
+    >>> flags = doctest.IGNORE_LINEBREAK
+    >>> doctest.DocTestRunner(verbose=False, optionflags=flags).run(test)
+    TestResults(failed=0, attempted=1)
+
+    ... ignore surrounding new lines
+
+    >>> "foobar"  # doctest: +IGNORE_LINEBREAK
+    '
+    foo
+    bar'
+    >>> "foobar"  # doctest: +IGNORE_LINEBREAK
+    'foo
+    bar
+    '
+    >>> "foobar"  # doctest: +IGNORE_LINEBREAK
+    '
+    foo
+    bar
+    '
+
+    ... non-quoted output:
+
+    >>> import string
+    >>> print(string.ascii_letters)  # doctest: +IGNORE_LINEBREAK
+    abcdefghijklmnopqrstuvwxyz
+    ABCDEFGHIJKLMNOPQRSTUVWXYZ
+
 The ELLIPSIS flag causes ellipsis marker ("...") in the expected
 output to match any substring in the actual output:
 

Misc/NEWS.d/next/Library/2025-08-25-13-31-25.gh-issue-138135.yo5w-T.rst

@@ -0,0 +1,3 @@
+:mod:`doctest`: Add :data:`~doctest.IGNORE_LINEBREAK` option to allow
+breaking expected output strings without whitespaces into multiple lines.
+Patch by Bénédikt Tran.