bpo-25054, bpo-1647489: Added support of splitting on zerowidth patterns. (#4471)

Also fixed searching patterns that could match an empty string.

Doc/library/re.rst

@@ -708,37 +708,19 @@ form.
    That way, separator components are always found at the same relative
    indices within the result list.
 
-   .. note::
-
-      :func:`split` doesn't currently split a string on an empty pattern match.
-      For example::
-
-         >>> re.split('x*', 'axbc')
-         ['a', 'bc']
+   The pattern can match empty strings. ::
 
-      Even though ``'x*'`` also matches 0 'x' before 'a', between 'b' and 'c',
-      and after 'c', currently these matches are ignored.  The correct behavior
-      (i.e. splitting on empty matches too and returning ``['', 'a', 'b', 'c',
-      '']``) will be implemented in future versions of Python, but since this
-      is a backward incompatible change, a :exc:`FutureWarning` will be raised
-      in the meanwhile.
-
-      Patterns that can only match empty strings currently never split the
-      string.  Since this doesn't match the expected behavior, a
-      :exc:`ValueError` will be raised starting from Python 3.5::
-
-         >>> re.split("^$", "foo\n\nbar\n", flags=re.M)
-         Traceback (most recent call last):
-           File "<stdin>", line 1, in <module>
-           ...
-         ValueError: split() requires a non-empty pattern match.
+      >>> re.split(r'\b', 'Words, words, words.')
+      ['', 'Words', ', ', 'words', ', ', 'words', '.']
+      >>> re.split(r'(\W*)', '...words...')
+      ['', '...', 'w', '', 'o', '', 'r', '', 'd', '', 's', '...', '']
 
    .. versionchanged:: 3.1
       Added the optional flags argument.
 
-   .. versionchanged:: 3.5
-      Splitting on a pattern that could match an empty string now raises
-      a warning.  Patterns that can only match empty strings are now rejected.
+   .. versionchanged:: 3.7
+      Added support of splitting on a pattern that could match an empty string.
+
 
 .. function:: findall(pattern, string, flags=0)
 
@@ -746,8 +728,10 @@ form.
    strings.  The *string* is scanned left-to-right, and matches are returned in
    the order found.  If one or more groups are present in the pattern, return a
    list of groups; this will be a list of tuples if the pattern has more than
-   one group.  Empty matches are included in the result unless they touch the
-   beginning of another match.
+   one group.  Empty matches are included in the result.
+
+   .. versionchanged:: 3.7
+      Non-empty matches can now start just after a previous empty match.
 
 
 .. function:: finditer(pattern, string, flags=0)
@@ -755,8 +739,10 @@ form.
    Return an :term:`iterator` yielding :ref:`match objects <match-objects>` over
    all non-overlapping matches for the RE *pattern* in *string*.  The *string*
    is scanned left-to-right, and matches are returned in the order found.  Empty
-   matches are included in the result unless they touch the beginning of another
-   match.
+   matches are included in the result.
+
+   .. versionchanged:: 3.7
+      Non-empty matches can now start just after a previous empty match.
 
 
 .. function:: sub(pattern, repl, string, count=0, flags=0)

Doc/whatsnew/3.7.rst

@@ -364,6 +364,10 @@ The flags :const:`re.ASCII`, :const:`re.LOCALE` and :const:`re.UNICODE`
 can be set within the scope of a group.
 (Contributed by Serhiy Storchaka in :issue:`31690`.)
 
+:func:`re.split` now supports splitting on a pattern like ``r'\b'``,
+``'^$'`` or ``(?=-)`` that matches an empty string.
+(Contributed by Serhiy Storchaka in :issue:`25054`.)
+
 string
 ------
 
@@ -768,6 +772,23 @@ Changes in the Python API
   avoid a warning escape them with a backslash.
   (Contributed by Serhiy Storchaka in :issue:`30349`.)
 
+* 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
+  to ``r'\s+'``.  A :exc:`FutureWarning` was emitted for such patterns since
+  Python 3.5.
+
+  For patterns that match both empty and non-empty strings, the result of
+  searching for all matches may also be changed in other cases.  For example
+  in the string ``'a\n\n'``, the pattern ``r'(?m)^\s*?$'`` will not only
+  match empty strings at positions 2 and 3, but also the string ``'\n'`` at
+  positions 2--3.  To match only blank lines, the pattern should be rewritten
+  as ``r'(?m)^[^\S\n]*$'``.
+
+  (Contributed by Serhiy Storchaka in :issue:`25054`.)
+
 * :class:`tracemalloc.Traceback` frames are now sorted from oldest to most
   recent to be more consistent with :mod:`traceback`.
   (Contributed by Jesse Bakker in :issue:`32121`.)

Lib/doctest.py

@@ -1611,7 +1611,7 @@ class OutputChecker:
                           '', want)
             # If a line in got contains only spaces, then remove the
             # spaces.
-            got = re.sub(r'(?m)^\s*?$', '', got)
+            got = re.sub(r'(?m)^[^\S\n]+$', '', got)
             if got == want:
                 return True
 

Lib/test/test_re.py

@@ -331,21 +331,21 @@ class ReTests(unittest.TestCase):
                          ['', 'a', '', '', 'c'])
 
         for sep, expected in [
-            (':*', ['', 'a', 'b', 'c']),
-            ('(?::*)', ['', 'a', 'b', 'c']),
-            ('(:*)', ['', ':', 'a', ':', 'b', '::', 'c']),
-            ('(:)*', ['', ':', 'a', ':', 'b', ':', 'c']),
+            (':*', ['', 'a', 'b', 'c', '']),
+            ('(?::*)', ['', 'a', 'b', 'c', '']),
+            ('(:*)', ['', ':', 'a', ':', 'b', '::', 'c', '', '']),
+            ('(:)*', ['', ':', 'a', ':', 'b', ':', 'c', None, '']),
         ]:
-            with self.subTest(sep=sep), self.assertWarns(FutureWarning):
+            with self.subTest(sep=sep):
                 self.assertTypedEqual(re.split(sep, ':a:b::c'), expected)
 
         for sep, expected in [
-            ('', [':a:b::c']),
-            (r'\b', [':a:b::c']),
-            (r'(?=:)', [':a:b::c']),
-            (r'(?<=:)', [':a:b::c']),
+            ('', ['', ':', 'a', ':', 'b', ':', ':', 'c', '']),
+            (r'\b', [':', 'a', ':', 'b', '::', 'c', '']),
+            (r'(?=:)', ['', ':a', ':b', ':', ':c']),
+            (r'(?<=:)', [':', 'a:', 'b:', ':', 'c']),
         ]:
-            with self.subTest(sep=sep), self.assertRaises(ValueError):
+            with self.subTest(sep=sep):
                 self.assertTypedEqual(re.split(sep, ':a:b::c'), expected)
 
     def test_qualified_re_split(self):
@@ -356,7 +356,6 @@ class ReTests(unittest.TestCase):
                          ['', ':', 'a', ':', 'b::c'])
         self.assertEqual(re.split("(:+)", ":a:b::c", maxsplit=2),
                          ['', ':', 'a', ':', 'b::c'])
-        with self.assertWarns(FutureWarning):
         self.assertEqual(re.split("(:*)", ":a:b::c", maxsplit=2),
                          ['', ':', 'a', ':', 'b::c'])
 
@@ -1751,6 +1750,25 @@ class ReTests(unittest.TestCase):
                          "span=(3, 5), match='bb'>" %
                          (type(second).__module__, type(second).__qualname__))
 
+    def test_zerowidth(self):
+        # Issues 852532, 1647489, 3262, 25054.
+        self.assertEqual(re.split(r"\b", "a::bc"), ['', 'a', '::', 'bc', ''])
+        self.assertEqual(re.split(r"\b|:+", "a::bc"), ['', 'a', '', 'bc', ''])
+        self.assertEqual(re.split(r"(?<!\w)(?=\w)|:+", "a::bc"), ['', 'a', 'bc'])
+        self.assertEqual(re.split(r"(?<=\w)(?!\w)|:+", "a::bc"), ['a', '', 'bc', ''])
+
+        self.assertEqual(re.sub(r"\b", "-", "a::bc"), '-a-::-bc-')
+        self.assertEqual(re.sub(r"\b|:+", "-", "a::bc"), '-a--bc-')
+        self.assertEqual(re.sub(r"(\b|:+)", r"[\1]", "a::bc"), '[]a[][::]bc[]')
+
+        self.assertEqual(re.findall(r"\b|:+", "a::bc"), ['', '', '::', '', ''])
+        self.assertEqual(re.findall(r"\b|\w+", "a::bc"),
+                         ['', 'a', '', '', 'bc', ''])
+
+        self.assertEqual([m.span() for m in re.finditer(r"\b|:+", "a::bc")],
+                         [(0, 0), (1, 1), (1, 3), (3, 3), (5, 5)])
+        self.assertEqual([m.span() for m in re.finditer(r"\b|\w+", "a::bc")],
+                         [(0, 0), (0, 1), (1, 1), (3, 3), (3, 5), (5, 5)])
 
     def test_bug_2537(self):
         # issue 2537: empty submatches

Misc/NEWS.d/next/Library/2017-11-20-01-01-01.bpo-25054.rOlRV6.rst

+Added support of splitting on a pattern that could match an empty string.

Misc/NEWS.d/next/Library/2017-11-20-01-29-46.bpo-1647489.-ZNNkh.rst

+Fixed searching regular expression patterns that could match an empty
+string. Non-empty string can now be correctly found after matching an empty
+string.

Modules/_sre.c

@@ -446,6 +446,8 @@ state_init(SRE_STATE* state, PatternObject* pattern, PyObject* string,
 
     state->isbytes = isbytes;
     state->charsize = charsize;
+    state->match_all = 0;
+    state->must_advance = 0;
 
     state->beginning = ptr;
 
@@ -559,14 +561,14 @@ pattern_dealloc(PatternObject* self)
 }
 
 LOCAL(Py_ssize_t)
-sre_match(SRE_STATE* state, SRE_CODE* pattern, int match_all)
+sre_match(SRE_STATE* state, SRE_CODE* pattern)
 {
     if (state->charsize == 1)
-        return sre_ucs1_match(state, pattern, match_all);
+        return sre_ucs1_match(state, pattern, 1);
     if (state->charsize == 2)
-        return sre_ucs2_match(state, pattern, match_all);
+        return sre_ucs2_match(state, pattern, 1);
     assert(state->charsize == 4);
-    return sre_ucs4_match(state, pattern, match_all);
+    return sre_ucs4_match(state, pattern, 1);
 }
 
 LOCAL(Py_ssize_t)
@@ -606,7 +608,7 @@ _sre_SRE_Pattern_match_impl(PatternObject *self, PyObject *string,
 
     TRACE(("|%p|%p|MATCH\n", PatternObject_GetCode(self), state.ptr));
 
-    status = sre_match(&state, PatternObject_GetCode(self), 0);
+    status = sre_match(&state, PatternObject_GetCode(self));
 
     TRACE(("|%p|%p|END\n", PatternObject_GetCode(self), state.ptr));
     if (PyErr_Occurred()) {
@@ -645,7 +647,8 @@ _sre_SRE_Pattern_fullmatch_impl(PatternObject *self, PyObject *string,
 
     TRACE(("|%p|%p|FULLMATCH\n", PatternObject_GetCode(self), state.ptr));
 
-    status = sre_match(&state, PatternObject_GetCode(self), 1);
+    state.match_all = 1;
+    status = sre_match(&state, PatternObject_GetCode(self));
 
     TRACE(("|%p|%p|END\n", PatternObject_GetCode(self), state.ptr));
     if (PyErr_Occurred()) {
@@ -808,11 +811,8 @@ _sre_SRE_Pattern_findall_impl(PatternObject *self, PyObject *string,
         if (status < 0)
             goto error;
 
-        if (state.ptr == state.start)
-            state.start = (void*) ((char*) state.ptr + state.charsize);
-        else
+        state.must_advance = (state.ptr == state.start);
         state.start = state.ptr;
-
     }
 
     state_fini(&state);
@@ -901,17 +901,6 @@ _sre_SRE_Pattern_split_impl(PatternObject *self, PyObject *string,
     void* last;
 
     assert(self->codesize != 0);
-    if (self->code[0] != SRE_OP_INFO || self->code[3] == 0) {
-        if (self->code[0] == SRE_OP_INFO && self->code[4] == 0) {
-            PyErr_SetString(PyExc_ValueError,
-                            "split() requires a non-empty pattern match.");
-            return NULL;
-        }
-        if (PyErr_WarnEx(PyExc_FutureWarning,
-                         "split() requires a non-empty pattern match.",
-                         1) < 0)
-            return NULL;
-    }
 
     if (!state_init(&state, self, string, 0, PY_SSIZE_T_MAX))
         return NULL;
@@ -942,14 +931,6 @@ _sre_SRE_Pattern_split_impl(PatternObject *self, PyObject *string,
             goto error;
         }
 
-        if (state.start == state.ptr) {
-            if (last == state.end || state.ptr == state.end)
-                break;
-            /* skip one character */
-            state.start = (void*) ((char*) state.ptr + state.charsize);
-            continue;
-        }
-
         /* get segment before this match */
         item = getslice(state.isbytes, state.beginning,
             string, STATE_OFFSET(&state, last),
@@ -974,7 +955,7 @@ _sre_SRE_Pattern_split_impl(PatternObject *self, PyObject *string,
         }
 
         n = n + 1;
-
+        state.must_advance = 1;
         last = state.start = state.ptr;
 
     }
@@ -1101,9 +1082,7 @@ pattern_subx(PatternObject* self, PyObject* ptemplate, PyObject* string,
             if (status < 0)
                 goto error;
 
-        } else if (i == b && i == e && n > 0)
-            /* ignore empty match on latest position */
-            goto next;
+        }
 
         if (filter_is_callable) {
             /* pass match object through filter */
@@ -1130,16 +1109,8 @@ pattern_subx(PatternObject* self, PyObject* ptemplate, PyObject* string,
 
         i = e;
         n = n + 1;
-
-next:
-        /* move on */
-        if (state.ptr == state.end)
-            break;
-        if (state.ptr == state.start)
-            state.start = (void*) ((char*) state.ptr + state.charsize);
-        else
+        state.must_advance = 1;
         state.start = state.ptr;
-
     }
 
     /* get segment following last match */
@@ -2450,7 +2421,7 @@ _sre_SRE_Scanner_match_impl(ScannerObject *self)
 
     state->ptr = state->start;
 
-    status = sre_match(state, PatternObject_GetCode(self->pattern), 0);
+    status = sre_match(state, PatternObject_GetCode(self->pattern));
     if (PyErr_Occurred())
         return NULL;
 
@@ -2459,12 +2430,10 @@ _sre_SRE_Scanner_match_impl(ScannerObject *self)
 
     if (status == 0)
         state->start = NULL;
-    else if (state->ptr != state->start)
+    else {
+        state->must_advance = (state->ptr == state->start);
         state->start = state->ptr;
-    else if (state->ptr != state->end)
-        state->start = (void*) ((char*) state->ptr + state->charsize);
-    else
-        state->start = NULL;
+    }
 
     return match;
 }
@@ -2499,12 +2468,10 @@ _sre_SRE_Scanner_search_impl(ScannerObject *self)
 
     if (status == 0)
         state->start = NULL;
-    else if (state->ptr != state->start)
+    else {
+        state->must_advance = (state->ptr == state->start);
         state->start = state->ptr;
-    else if (state->ptr != state->end)
-        state->start = (void*) ((char*) state->ptr + state->charsize);
-    else
-        state->start = NULL;
+    }
 
     return match;
 }

Modules/sre.h

@@ -67,6 +67,7 @@ typedef struct {
     void* end; /* end of original string */
     /* attributes for the match object */
     PyObject* string;
+    Py_buffer buffer;
     Py_ssize_t pos, endpos;
     int isbytes;
     int charsize; /* character size */
@@ -74,11 +75,12 @@ typedef struct {
     Py_ssize_t lastindex;
     Py_ssize_t lastmark;
     void** mark;
+    int match_all;
+    int must_advance;
     /* dynamically allocated stuff */
     char* data_stack;
     size_t data_stack_size;
     size_t data_stack_base;
-    Py_buffer buffer;
     /* current repeat context */
     SRE_REPEAT *repeat;
 } SRE_STATE;

Modules/sre_lib.h

@@ -199,7 +199,7 @@ SRE(charset_loc_ignore)(SRE_STATE* state, SRE_CODE* set, SRE_CODE ch)
     return up != lo && SRE(charset)(state, set, up);
 }
 
-LOCAL(Py_ssize_t) SRE(match)(SRE_STATE* state, SRE_CODE* pattern, int match_all);
+LOCAL(Py_ssize_t) SRE(match)(SRE_STATE* state, SRE_CODE* pattern, int toplevel);
 
 LOCAL(Py_ssize_t)
 SRE(count)(SRE_STATE* state, SRE_CODE* pattern, Py_ssize_t maxcount)
@@ -510,12 +510,12 @@ do { \
 #define JUMP_ASSERT          12
 #define JUMP_ASSERT_NOT      13
 
-#define DO_JUMPX(jumpvalue, jumplabel, nextpattern, matchall) \
+#define DO_JUMPX(jumpvalue, jumplabel, nextpattern, toplevel_) \
     DATA_ALLOC(SRE(match_context), nextctx); \
     nextctx->last_ctx_pos = ctx_pos; \
     nextctx->jump = jumpvalue; \
     nextctx->pattern = nextpattern; \
-    nextctx->match_all = matchall; \
+    nextctx->toplevel = toplevel_; \
     ctx_pos = alloc_pos; \
     ctx = nextctx; \
     goto entrance; \
@@ -523,7 +523,7 @@ do { \
     while (0) /* gcc doesn't like labels at end of scopes */ \
 
 #define DO_JUMP(jumpvalue, jumplabel, nextpattern) \
-    DO_JUMPX(jumpvalue, jumplabel, nextpattern, ctx->match_all)
+    DO_JUMPX(jumpvalue, jumplabel, nextpattern, ctx->toplevel)
 
 #define DO_JUMP0(jumpvalue, jumplabel, nextpattern) \
     DO_JUMPX(jumpvalue, jumplabel, nextpattern, 0)
@@ -540,13 +540,13 @@ typedef struct {
         SRE_CODE chr;
         SRE_REPEAT* rep;
     } u;
-    int match_all;
+    int toplevel;
 } SRE(match_context);
 
 /* check if string matches the given pattern.  returns <0 for
    error, 0 for failure, and 1 for success */
 LOCAL(Py_ssize_t)
-SRE(match)(SRE_STATE* state, SRE_CODE* pattern, int match_all)
+SRE(match)(SRE_STATE* state, SRE_CODE* pattern, int toplevel)
 {
     SRE_CHAR* end = (SRE_CHAR *)state->end;
     Py_ssize_t alloc_pos, ctx_pos = -1;
@@ -563,7 +563,7 @@ SRE(match)(SRE_STATE* state, SRE_CODE* pattern, int match_all)
     ctx->last_ctx_pos = -1;
     ctx->jump = JUMP_NONE;
     ctx->pattern = pattern;
-    ctx->match_all = match_all;
+    ctx->toplevel = toplevel;
     ctx_pos = alloc_pos;
 
 entrance:
@@ -636,11 +636,14 @@ entrance:
         case SRE_OP_SUCCESS:
             /* end of pattern */
             TRACE(("|%p|%p|SUCCESS\n", ctx->pattern, ctx->ptr));
-            if (!ctx->match_all || ctx->ptr == state->end) {
+            if (ctx->toplevel &&
+                ((state->match_all && ctx->ptr != state->end) ||
+                 (state->must_advance && ctx->ptr == state->start)))
+            {
+                RETURN_FAILURE;
+            }
             state->ptr = ctx->ptr;
             RETURN_SUCCESS;
-            }
-            RETURN_FAILURE;
 
         case SRE_OP_AT:
             /* match at given position */
@@ -856,7 +859,9 @@ entrance:
                 RETURN_FAILURE;
 
             if (ctx->pattern[ctx->pattern[0]] == SRE_OP_SUCCESS &&
-                ctx->ptr == state->end) {
+                ctx->ptr == state->end &&
+                !(ctx->toplevel && state->must_advance && ctx->ptr == state->start))
+            {
                 /* tail is empty.  we're finished */
                 state->ptr = ctx->ptr;
                 RETURN_SUCCESS;
@@ -941,7 +946,10 @@ entrance:
             }
 
             if (ctx->pattern[ctx->pattern[0]] == SRE_OP_SUCCESS &&
-                (!match_all || ctx->ptr == state->end)) {
+                !(ctx->toplevel &&
+                  ((state->match_all && ctx->ptr != state->end) ||
+                   (state->must_advance && ctx->ptr == state->start))))
+            {
                 /* tail is empty.  we're finished */
                 state->ptr = ctx->ptr;
                 RETURN_SUCCESS;
@@ -1417,6 +1425,7 @@ SRE(search)(SRE_STATE* state, SRE_CODE* pattern)
             return 0; /* literal can't match: doesn't fit in char width */
 #endif
         end = (SRE_CHAR *)state->end;
+        state->must_advance = 0;
         while (ptr < end) {
             while (*ptr != c) {
                 if (++ptr >= end)
@@ -1458,6 +1467,7 @@ SRE(search)(SRE_STATE* state, SRE_CODE* pattern)
                 return 0;
 
             i = 1;
+            state->must_advance = 0;
             do {
                 if (*ptr == (SRE_CHAR) prefix[i]) {
                     if (++i != prefix_len) {
@@ -1487,6 +1497,7 @@ SRE(search)(SRE_STATE* state, SRE_CODE* pattern)
     if (charset) {
         /* pattern starts with a character from a known set */
         end = (SRE_CHAR *)state->end;
+        state->must_advance = 0;
         for (;;) {
             while (ptr < end && !SRE(charset)(state, charset, *ptr))
                 ptr++;
@@ -1503,13 +1514,15 @@ SRE(search)(SRE_STATE* state, SRE_CODE* pattern)
     } else {
         /* general case */
         assert(ptr <= end);
-        while (1) {
         TRACE(("|%p|%p|SEARCH\n", pattern, ptr));
         state->start = state->ptr = ptr;
-            status = SRE(match)(state, pattern, 0);
-            if (status != 0 || ptr >= end)
-                break;
+        status = SRE(match)(state, pattern, 1);
+        state->must_advance = 0;
+        while (status == 0 && ptr < end) {
             ptr++;
+            TRACE(("|%p|%p|SEARCH\n", pattern, ptr));
+            state->start = state->ptr = ptr;
+            status = SRE(match)(state, pattern, 0);
         }
     }