bpo-32308: Replace empty matches adjacent to a previous non-empty match in re.sub(). (#4846)

3 files changed, 22 insertions, 9 deletions

diff --git a/Doc/howto/regex.rst b/Doc/howto/regex.rst
index fa8c693..87a6b1a 100644
--- a/Doc/howto/regex.rst
+++ b/Doc/howto/regex.rst
@@ -1140,12 +1140,12 @@ new string value and the number of replacements  that were performed::
    >>> p.subn('colour', 'no colours at all')
    ('no colours at all', 0)
 
-Empty matches are replaced only when they're not adjacent to a previous match.
+Empty matches are replaced only when they're not adjacent to a previous empty match.
 ::
 
    >>> p = re.compile('x*')
    >>> p.sub('-', 'abxd')
-   '-a-b-d-'
+   '-a-b--d-'
 
 If *replacement* is a string, any backslash escapes in it are processed.  That
 is, ``\n`` is converted to a single newline character, ``\r`` is converted to a
diff --git a/Doc/library/re.rst b/Doc/library/re.rst
index dae1d7e..9b175f4 100644
--- a/Doc/library/re.rst
+++ b/Doc/library/re.rst
@@ -708,12 +708,15 @@ form.
    That way, separator components are always found at the same relative
    indices within the result list.
 
-   The pattern can match empty strings. ::
+   Empty matches for the pattern split the string only when not adjacent
+   to a previous empty match.
 
       >>> re.split(r'\b', 'Words, words, words.')
       ['', 'Words', ', ', 'words', ', ', 'words', '.']
+      >>> re.split(r'\W*', '...words...')
+      ['', '', 'w', 'o', 'r', 'd', 's', '', '']
       >>> re.split(r'(\W*)', '...words...')
-      ['', '...', 'w', '', 'o', '', 'r', '', 'd', '', 's', '...', '']
+      ['', '...', '', '', 'w', '', 'o', '', 'r', '', 'd', '', 's', '...', '', '', '']
 
    .. versionchanged:: 3.1
       Added the optional flags argument.
@@ -778,8 +781,8 @@ form.
    The optional argument *count* is the maximum number of pattern occurrences to be
    replaced; *count* must be a non-negative integer.  If omitted or zero, all
    occurrences will be replaced. Empty matches for the pattern are replaced only
-   when not adjacent to a previous match, so ``sub('x*', '-', 'abc')`` returns
-   ``'-a-b-c-'``.
+   when not adjacent to a previous empty match, so ``sub('x*', '-', 'abxd')`` returns
+   ``'-a-b--d-'``.
 
    In string-type *repl* arguments, in addition to the character escapes and
    backreferences described above,
@@ -805,6 +808,9 @@ form.
       Unknown escapes in *repl* consisting of ``'\'`` and an ASCII letter
       now are errors.
 
+      Empty matches for the pattern are replaced when adjacent to a previous
+      non-empty match.
+
 
 .. function:: subn(pattern, repl, string, count=0, flags=0)
 
diff --git a/Doc/whatsnew/3.7.rst b/Doc/whatsnew/3.7.rst
index 1924881..1311e9e 100644
--- a/Doc/whatsnew/3.7.rst
+++ b/Doc/whatsnew/3.7.rst
@@ -881,8 +881,9 @@ Changes in the Python API
 * The result of splitting a string on a :mod:`regular expression <re>`
   that could match an empty string has been changed.  For example
   splitting on ``r'\s*'`` will now split not only on whitespaces as it
-  did previously, but also between any pair of non-whitespace
-  characters.  The previous behavior can be restored by changing the pattern
+  did previously, but also on empty strings before all non-whitespace
+  characters and just before the end of the string.
+  The previous behavior can be restored by changing the pattern
   to ``r'\s+'``.  A :exc:`FutureWarning` was emitted for such patterns since
   Python 3.5.
 
@@ -893,7 +894,13 @@ Changes in the Python API
   positions 2--3.  To match only blank lines, the pattern should be rewritten
   as ``r'(?m)^[^\S\n]*$'``.
 
-  (Contributed by Serhiy Storchaka in :issue:`25054`.)
+  :func:`re.sub()` now replaces empty matches adjacent to a previous
+  non-empty match.  For example ``re.sub('x*', '-', 'abxd')`` returns now
+  ``'-a-b--d-'`` instead of ``'-a-b--d-'`` (the first minus between 'b' and
+  'd' replaces 'x', and the second minus replaces an empty string between
+  'x' and 'd').
+
+  (Contributed by Serhiy Storchaka in :issue:`25054` and :issue:`32308`.)
 
 * :class:`tracemalloc.Traceback` frames are now sorted from oldest to most
   recent to be more consistent with :mod:`traceback`.