Remove atomic restriction on capture groups containing recursive back

references, as since 10.30 it has been unnecessary.

ChangeLog

@@ -11,6 +11,17 @@ Version 10.35
 3. A JIT bug is fixed which allowed to read the fields of the compiled
 pattern before its existence is checked.
 
+4. Back in the PCRE1 day, capturing groups that contained recursive back 
+references to themselves were made atomic (version 8.01, change 18) because 
+after the end a repeated group, the captured substrings had their values from 
+the final repetition, not from an earlier repetition that might be the 
+destination of a backtrack. This feature was documented, and was carried over 
+into PCRE2. However, it has now been realized that the major refactoring that 
+was done for 10.30 has made this atomicizing unnecessary, and it is confusing 
+when users are unaware of it, making some patterns appear not to be working as 
+expected. Capture values of recursive back references in repeated groups are
+now correctly backtracked, so this unnecessary restriction has been removed.
+
 
 Version 10.34 21-November-2019
 ------------------------------

configure.ac

@@ -9,9 +9,9 @@ dnl The PCRE2_PRERELEASE feature is for identifying release candidates. It might
 dnl be defined as -RC2, for example. For real releases, it should be empty.
 
 m4_define(pcre2_major, [10])
-m4_define(pcre2_minor, [34])
+m4_define(pcre2_minor, [35])
-m4_define(pcre2_prerelease, [])
+m4_define(pcre2_prerelease, [-RC1])
-m4_define(pcre2_date, [2019-11-21])
+m4_define(pcre2_date, [2019-11-27])
 
 # NOTE: The CMakeLists.txt file searches for the above variables in the first
 # 50 lines of this file. Please update that if the variables above are moved.

doc/html/pcre2pattern.html

@@ -2349,11 +2349,11 @@ using alternation, as in the example above, or by a quantifier with a minimum
 of zero.
 </P>
 <P>
-Backreferences of this type cause the group that they reference to be treated
+For versions of PCRE2 less than 10.25, backreferences of this type used to
-as an
+cause the group that they reference to be treated as an
 <a href="#atomicgroup">atomic group.</a>
-Once the whole group has been matched, a subsequent matching failure cannot
+This restriction no longer applies, and backtracking into such groups can occur 
-cause backtracking into the middle of the group.
+as normal.
 <a name="bigassertions"></a></P>
 <br><a name="SEC20" href="#TOC1">ASSERTIONS</a><br>
 <P>
@@ -3833,7 +3833,7 @@ Cambridge, England.
 </P>
 <br><a name="SEC32" href="#TOC1">REVISION</a><br>
 <P>
-Last updated: 29 July 2019
+Last updated: 18 December 2019
 <br>
 Copyright &copy; 1997-2019 University of Cambridge.
 <br>

doc/pcre2.txt

@@ -180,8 +180,8 @@ REVISION
        Last updated: 17 September 2018
        Copyright (c) 1997-2018 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2API(3)                Library Functions Manual                PCRE2API(3)
 
 
@@ -3724,8 +3724,8 @@ REVISION
        Last updated: 02 September 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2BUILD(3)              Library Functions Manual              PCRE2BUILD(3)
 
 
@@ -4296,8 +4296,8 @@ REVISION
        Last updated: 03 March 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2CALLOUT(3)            Library Functions Manual            PCRE2CALLOUT(3)
 
 
@@ -4726,8 +4726,8 @@ REVISION
        Last updated: 03 February 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2COMPAT(3)             Library Functions Manual             PCRE2COMPAT(3)
 
 
@@ -4935,8 +4935,8 @@ REVISION
        Last updated: 13 July 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2JIT(3)                Library Functions Manual                PCRE2JIT(3)
 
 
@@ -5360,8 +5360,8 @@ REVISION
        Last updated: 23 May 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2LIMITS(3)             Library Functions Manual             PCRE2LIMITS(3)
 
 
@@ -5430,8 +5430,8 @@ REVISION
        Last updated: 02 February 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2MATCHING(3)           Library Functions Manual           PCRE2MATCHING(3)
 
 
@@ -5654,8 +5654,8 @@ REVISION
        Last updated: 23 May 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2PARTIAL(3)            Library Functions Manual            PCRE2PARTIAL(3)
 
 
@@ -6034,8 +6034,8 @@ REVISION
        Last updated: 04 September 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2PATTERN(3)            Library Functions Manual            PCRE2PATTERN(3)
 
 
@@ -8078,10 +8078,10 @@ BACKREFERENCES
        the backreference. This can be done using alternation, as in the  exam-
        ple above, or by a quantifier with a minimum of zero.
 
-       Backreferences  of  this type cause the group that they reference to be
+       For versions of PCRE2 less than 10.25, backreferences of this type used
-       treated as an atomic group.  Once the whole group has been  matched,  a
+       to cause the group that they reference  to  be  treated  as  an  atomic
-       subsequent  matching  failure cannot cause backtracking into the middle
+       group.   This restriction no longer applies, and backtracking into such
-       of the group.
+       groups can occur as normal.
 
 
 ASSERTIONS
@@ -9463,11 +9463,11 @@ AUTHOR
 
 REVISION
 
-       Last updated: 29 July 2019
+       Last updated: 18 December 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2PERFORM(3)            Library Functions Manual            PCRE2PERFORM(3)
 
 
@@ -9701,8 +9701,8 @@ REVISION
        Last updated: 03 February 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2POSIX(3)              Library Functions Manual              PCRE2POSIX(3)
 
 
@@ -10031,8 +10031,8 @@ REVISION
        Last updated: 30 January 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2SAMPLE(3)             Library Functions Manual             PCRE2SAMPLE(3)
 
 
@@ -10310,8 +10310,8 @@ REVISION
        Last updated: 27 June 2018
        Copyright (c) 1997-2018 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2SYNTAX(3)             Library Functions Manual             PCRE2SYNTAX(3)
 
 
@@ -10823,8 +10823,8 @@ REVISION
        Last updated: 29 July 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 
 PCRE2UNICODE(3)            Library Functions Manual            PCRE2UNICODE(3)
 
 
@@ -11256,5 +11256,5 @@ REVISION
        Last updated: 24 May 2019
        Copyright (c) 1997-2019 University of Cambridge.
 ------------------------------------------------------------------------------
-
+ 
-
+ 

doc/pcre2pattern.3

@@ -1,4 +1,4 @@
-.TH PCRE2PATTERN 3 "29 July 2019" "PCRE2 10.34"
+.TH PCRE2PATTERN 3 "18 December 2019" "PCRE2 10.35"
 .SH NAME
 PCRE2 - Perl-compatible regular expressions (revised API)
 .SH "PCRE2 REGULAR EXPRESSION DETAILS"
@@ -2346,14 +2346,14 @@ the first iteration does not need to match the backreference. This can be done
 using alternation, as in the example above, or by a quantifier with a minimum
 of zero.
 .P
-Backreferences of this type cause the group that they reference to be treated
+For versions of PCRE2 less than 10.25, backreferences of this type used to
-as an
+cause the group that they reference to be treated as an
 .\" HTML <a href="#atomicgroup">
 .\" </a>
 atomic group.
 .\"
-Once the whole group has been matched, a subsequent matching failure cannot
+This restriction no longer applies, and backtracking into such groups can occur 
-cause backtracking into the middle of the group.
+as normal.
 .
 .
 .\" HTML <a name="bigassertions"></a>
@@ -3874,6 +3874,6 @@ Cambridge, England.
 .rs
 .sp
 .nf
-Last updated: 29 July 2019
+Last updated: 18 December 2019
 Copyright (c) 1997-2019 University of Cambridge.
 .fi

src/pcre2_compile.c

@@ -6671,23 +6671,11 @@ for (;; pptr++)
             }
 
           /* For a back reference, update the back reference map and the
-          maximum back reference. Then, for each group, we must check to
+          maximum back reference. */
-          see if it is recursive, that is, it is inside the group that it
-          references. A flag is set so that the group can be made atomic.
-          */
 
           cb->backref_map |= (groupnumber < 32)? (1u << groupnumber) : 1;
           if (groupnumber > cb->top_backref)
             cb->top_backref = groupnumber;
-
-          for (oc = cb->open_caps; oc != NULL; oc = oc->next)
-            {
-            if (oc->number == groupnumber)
-              {
-              oc->flag = TRUE;
-              break;
-              }
-            }
           }
         }
 
@@ -7682,19 +7670,6 @@ for (;; pptr++)
 
     cb->backref_map |= (meta_arg < 32)? (1u << meta_arg) : 1;
     if (meta_arg > cb->top_backref) cb->top_backref = meta_arg;
-
-    /* Check to see if this back reference is recursive, that it, it
-    is inside the group that it references. A flag is set so that the
-    group can be made atomic. */
-
-    for (oc = cb->open_caps; oc != NULL; oc = oc->next)
-      {
-      if (oc->number == meta_arg)
-        {
-        oc->flag = TRUE;
-        break;
-        }
-      }
     break;
 
 
@@ -8035,7 +8010,6 @@ and skip over the pattern offset. */
 lookbehind = *code == OP_ASSERTBACK ||
              *code == OP_ASSERTBACK_NOT ||
              *code == OP_ASSERTBACK_NA;
-
 if (lookbehind)
   {
   lookbehindlength = META_DATA(pptr[-1]);
@@ -8053,7 +8027,6 @@ if (*code == OP_CBRA)
   capnumber = GET2(code, 1 + LINK_SIZE);
   capitem.number = capnumber;
   capitem.next = cb->open_caps;
-  capitem.flag = FALSE;
   capitem.assert_depth = cb->assert_depth;
   cb->open_caps = &capitem;
   }
@@ -8182,26 +8155,9 @@ for (;;)
     PUT(code, 1, (int)(code - start_bracket));
     code += 1 + LINK_SIZE;
 
-    /* If it was a capturing subpattern, check to see if it contained any
+    /* If it was a capturing subpattern, remove the block from the chain. */
-    recursive back references. If so, we must wrap it in atomic brackets. In
-    any event, remove the block from the chain. */
 
-    if (capnumber > 0)
+    if (capnumber > 0) cb->open_caps = cb->open_caps->next;
-      {
-      if (cb->open_caps->flag)
-        {
-        (void)memmove(start_bracket + 1 + LINK_SIZE, start_bracket,
-          CU2BYTES(code - start_bracket));
-        *start_bracket = OP_ONCE;
-        code += 1 + LINK_SIZE;
-        PUT(start_bracket, 1, (int)(code - start_bracket));
-        *code = OP_KET;
-        PUT(code, 1, (int)(code - start_bracket));
-        code += 1 + LINK_SIZE;
-        length += 2 + 2*LINK_SIZE;
-        }
-      cb->open_caps = cb->open_caps->next;
-      }
 
     /* Set values to pass back */
 

src/pcre2_internal.h

@@ -1759,13 +1759,11 @@ typedef struct pcre2_memctl {
 
 /* Structure for building a chain of open capturing subpatterns during
 compiling, so that instructions to close them can be compiled when (*ACCEPT) is
-encountered. This is also used to identify subpatterns that contain recursive
+encountered. */
-back references to themselves, so that they can be made atomic. */
 
 typedef struct open_capitem {
   struct open_capitem *next;    /* Chain link */
   uint16_t number;              /* Capture number */
-  uint16_t flag;                /* Set TRUE if recursive back ref */
   uint16_t assert_depth;        /* Assertion depth when opened */
 } open_capitem;
 

testdata/testinput1

@@ -6386,4 +6386,11 @@ ef) x/x,mark
 /^(?<A>a)(?(<A>)b)((?<=b).*)$/
     abc
 
+/^(a\1?){4}$/
+    aaaa
+    aaaaaa
+
+/^((\1+)|\d)+133X$/
+    111133X
+
 # End of testinput1 

testdata/testinput2

@@ -324,16 +324,7 @@
 \= Expect no match
     fooabar
 
-# This one is here because Perl behaves differently; see also the following.
+# Perl does not fail these two for the final subjects.
-
-/^(a\1?){4}$/I
-\= Expect no match
-    aaaa
-    aaaaaa
-
-# Perl does not fail these two for the final subjects. Neither did PCRE until
-# release 8.01. The problem is in backtracking into a subpattern that contains
-# a recursive reference to itself. PCRE has now made these into atomic patterns.
 
 /^(xa|=?\1a){2}$/
     xa=xaa
@@ -5772,4 +5763,13 @@ a)"xI
 /(a)?a/I
     manm
 
+/^(?|(\*)(*napla:\S*_(\2?+.+))|(\w)(?=\S*_(\2?+\1)))+_\2$/
+    *abc_12345abc
+
+/^(?|(\*)(*napla:\S*_(\3?+.+))|(\w)(?=\S*_((\2?+\1))))+_\2$/
+    *abc_12345abc
+
+/^((\1+)(?C)|\d)+133X$/
+    111133X\=callout_capture
+
 # End of testinput2

testdata/testoutput1

@@ -10112,4 +10112,18 @@ No match
  1: a
  2: c
 
+/^(a\1?){4}$/
+    aaaa
+ 0: aaaa
+ 1: a
+    aaaaaa
+ 0: aaaaaa
+ 1: aa
+
+/^((\1+)|\d)+133X$/
+    111133X
+ 0: 111133X
+ 1: 11
+ 2: 11
+
 # End of testinput1 

testdata/testoutput2

@@ -809,24 +809,7 @@ Subject length lower bound = 3
     fooabar
 No match
 
-# This one is here because Perl behaves differently; see also the following.
+# Perl does not fail these two for the final subjects.
-
-/^(a\1?){4}$/I
-Capture group count = 1
-Max back reference = 1
-Compile options: <none>
-Overall options: anchored
-First code unit = 'a'
-Subject length lower bound = 4
-\= Expect no match
-    aaaa
-No match
-    aaaaaa
-No match
-
-# Perl does not fail these two for the final subjects. Neither did PCRE until
-# release 8.01. The problem is in backtracking into a subpattern that contains
-# a recursive reference to itself. PCRE has now made these into atomic patterns.
 
 /^(xa|=?\1a){2}$/
     xa=xaa
@@ -10060,7 +10043,6 @@ No match
 ------------------------------------------------------------------
         Bra
         ^
-        Once
         CBra 1
         ab
         CBra 2
@@ -10071,8 +10053,6 @@ No match
         Alt
         x
         Ket
-        Ket
-        Once
         CBra 1
         ab
         CBra 2
@@ -10083,7 +10063,6 @@ No match
         Alt
         x
         Ket
-        Ket
         $
         Ket
         End
@@ -10479,27 +10458,23 @@ Failed: error 168 at offset 3: \c must be followed by a printable ASCII characte
 /(?P<abn>(?P=abn)xxx)/B
 ------------------------------------------------------------------
         Bra
-        Once
         CBra 1
         \1
         xxx
         Ket
         Ket
-        Ket
         End
 ------------------------------------------------------------------
 
 /(a\1z)/B
 ------------------------------------------------------------------
         Bra
-        Once
         CBra 1
         a
         \1
         z
         Ket
         Ket
-        Ket
         End
 ------------------------------------------------------------------
 
@@ -11299,27 +11274,23 @@ No match
 /(?P<abn>(?P=abn)xxx)/B
 ------------------------------------------------------------------
         Bra
-        Once
         CBra 1
         \1
         xxx
         Ket
         Ket
-        Ket
         End
 ------------------------------------------------------------------
 
 /(a\1z)/B
 ------------------------------------------------------------------
         Bra
-        Once
         CBra 1
         a
         \1
         z
         Ket
         Ket
-        Ket
         End
 ------------------------------------------------------------------
 
@@ -13319,7 +13290,6 @@ Failed: error 144 at offset 5: subpattern name must start with a non-digit
         Bra
         Brazero
         SCBra 1
-        Once
         CBra 2
         CBra 3
         a
@@ -13331,7 +13301,6 @@ Failed: error 144 at offset 5: subpattern name must start with a non-digit
         Ket
         Recurse
         Ket
-        Ket
         KetRmax
         a?+
         Ket
@@ -13999,7 +13968,6 @@ Matched, but too many substrings
 /((?+1)(\1))/B
 ------------------------------------------------------------------
         Bra
-        Once
         CBra 1
         Recurse
         CBra 2
@@ -14007,7 +13975,6 @@ Matched, but too many substrings
         Ket
         Ket
         Ket
-        Ket
         End
 ------------------------------------------------------------------
 
@@ -14425,7 +14392,6 @@ Subject length lower bound = 1
 ------------------------------------------------------------------
         Bra
         Any
-        Once
         CBra 1
         Recurse
         Recurse
@@ -14434,7 +14400,6 @@ Subject length lower bound = 1
         Alt
         $
         Ket
-        Ket
         CBra 2
         Ket
         Ket
@@ -14445,7 +14410,6 @@ Subject length lower bound = 1
 ------------------------------------------------------------------
         Bra
         Any
-        Once
         CBra 1
         Recurse
         Recurse
@@ -14457,7 +14421,6 @@ Subject length lower bound = 1
         Alt
         $
         Ket
-        Ket
         CBra 3
         Ket
         Ket
@@ -17435,6 +17398,45 @@ Subject length lower bound = 1
     manm
  0: a
 
+/^(?|(\*)(*napla:\S*_(\2?+.+))|(\w)(?=\S*_(\2?+\1)))+_\2$/
+    *abc_12345abc
+ 0: *abc_12345abc
+ 1: c
+ 2: 12345abc
+
+/^(?|(\*)(*napla:\S*_(\3?+.+))|(\w)(?=\S*_((\2?+\1))))+_\2$/
+    *abc_12345abc
+ 0: *abc_12345abc
+ 1: c
+ 2: 12345abc
+ 3: 12345abc
+
+/^((\1+)(?C)|\d)+133X$/
+    111133X\=callout_capture
+Callout 0: last capture = 2
+ 1: 1
+ 2: 111
+--->111133X
+    ^   ^       |
+Callout 0: last capture = 2
+ 1: 3
+ 2: 3
+--->111133X
+    ^     ^     |
+Callout 0: last capture = 2
+ 1: 1
+ 2: 11
+--->111133X
+    ^  ^        |
+Callout 0: last capture = 2
+ 1: 3
+ 2: 3
+--->111133X
+    ^     ^     |
+ 0: 111133X
+ 1: 11
+ 2: 11
+
 # End of testinput2
 Error -70: PCRE2_ERROR_BADDATA (unknown error number)
 Error -62: bad serialized data

testdata/testoutput8-16-2

@@ -720,41 +720,37 @@ Memory allocation (code space): 14
 
 /(((a\2)|(a*)\g<-1>))*a?/
 ------------------------------------------------------------------
-  0  39 Bra
+  0  35 Bra
   2     Brazero
-  3  32 SCBra 1
+  3  28 SCBra 1
-  6  27 Once
+  6  12 CBra 2
-  8  12 CBra 2
+  9   7 CBra 3
- 11   7 CBra 3
+ 12     a
- 14     a
+ 14     \2
- 16     \2
+ 16   7 Ket
- 18   7 Ket
+ 18  11 Alt
- 20  11 Alt
+ 20   5 CBra 4
- 22   5 CBra 4
+ 23     a*
- 25     a*
+ 25   5 Ket
- 27   5 Ket
+ 27  20 Recurse
- 29  22 Recurse
+ 29  23 Ket
- 31  23 Ket
+ 31  28 KetRmax
- 33  27 Ket
+ 33     a?+
- 35  32 KetRmax
+ 35  35 Ket
- 37     a?+
+ 37     End
- 39  39 Ket
- 41     End
 ------------------------------------------------------------------
 
 /((?+1)(\1))/
 ------------------------------------------------------------------
-  0  20 Bra
+  0  16 Bra
-  2  16 Once
+  2  12 CBra 1
-  4  12 CBra 1
+  5   7 Recurse
-  7   9 Recurse
+  7   5 CBra 2
-  9   5 CBra 2
+ 10     \1
- 12     \1
+ 12   5 Ket
- 14   5 Ket
+ 14  12 Ket
- 16  12 Ket
+ 16  16 Ket
- 18  16 Ket
+ 18     End
- 20  20 Ket
- 22     End
 ------------------------------------------------------------------
 
 "(?1)(?#?'){2}(a)"
@@ -771,45 +767,41 @@ Memory allocation (code space): 14
 
 /.((?2)(?R)|\1|$)()/
 ------------------------------------------------------------------
-  0  28 Bra
+  0  24 Bra
   2     Any
-  3  18 Once
+  3   7 CBra 1
-  5   7 CBra 1
+  6  19 Recurse
-  8  23 Recurse
+  8   0 Recurse
- 10   0 Recurse
+ 10   4 Alt
- 12   4 Alt
+ 12     \1
- 14     \1
+ 14   3 Alt
- 16   3 Alt
+ 16     $
- 18     $
+ 17  14 Ket
- 19  14 Ket
+ 19   3 CBra 2
- 21  18 Ket
+ 22   3 Ket
- 23   3 CBra 2
+ 24  24 Ket
- 26   3 Ket
+ 26     End
- 28  28 Ket
- 30     End
 ------------------------------------------------------------------
 
 /.((?3)(?R)()(?2)|\1|$)()/
 ------------------------------------------------------------------
-  0  35 Bra
+  0  31 Bra
   2     Any
-  3  25 Once
+  3  14 CBra 1
-  5  14 CBra 1
+  6  26 Recurse
-  8  30 Recurse
+  8   0 Recurse
- 10   0 Recurse
+ 10   3 CBra 2
- 12   3 CBra 2
+ 13   3 Ket
- 15   3 Ket
+ 15  10 Recurse
- 17  12 Recurse
+ 17   4 Alt
- 19   4 Alt
+ 19     \1
- 21     \1
+ 21   3 Alt
- 23   3 Alt
+ 23     $
- 25     $
+ 24  21 Ket
- 26  21 Ket
+ 26   3 CBra 3
- 28  25 Ket
+ 29   3 Ket
- 30   3 CBra 3
+ 31  31 Ket
- 33   3 Ket
+ 33     End
- 35  35 Ket
- 37     End
 ------------------------------------------------------------------
 
 /(?1)()((((((\1++))\x85)+)|))/

testdata/testoutput8-32-2

@@ -720,41 +720,37 @@ Memory allocation (code space): 28
 
 /(((a\2)|(a*)\g<-1>))*a?/
 ------------------------------------------------------------------
-  0  39 Bra
+  0  35 Bra
   2     Brazero
-  3  32 SCBra 1
+  3  28 SCBra 1
-  6  27 Once
+  6  12 CBra 2
-  8  12 CBra 2
+  9   7 CBra 3
- 11   7 CBra 3
+ 12     a
- 14     a
+ 14     \2
- 16     \2
+ 16   7 Ket
- 18   7 Ket
+ 18  11 Alt
- 20  11 Alt
+ 20   5 CBra 4
- 22   5 CBra 4
+ 23     a*
- 25     a*
+ 25   5 Ket
- 27   5 Ket
+ 27  20 Recurse
- 29  22 Recurse
+ 29  23 Ket
- 31  23 Ket
+ 31  28 KetRmax
- 33  27 Ket
+ 33     a?+
- 35  32 KetRmax
+ 35  35 Ket
- 37     a?+
+ 37     End
- 39  39 Ket
- 41     End
 ------------------------------------------------------------------
 
 /((?+1)(\1))/
 ------------------------------------------------------------------
-  0  20 Bra
+  0  16 Bra
-  2  16 Once
+  2  12 CBra 1
-  4  12 CBra 1
+  5   7 Recurse
-  7   9 Recurse
+  7   5 CBra 2
-  9   5 CBra 2
+ 10     \1
- 12     \1
+ 12   5 Ket
- 14   5 Ket
+ 14  12 Ket
- 16  12 Ket
+ 16  16 Ket
- 18  16 Ket
+ 18     End
- 20  20 Ket
- 22     End
 ------------------------------------------------------------------
 
 "(?1)(?#?'){2}(a)"
@@ -771,45 +767,41 @@ Memory allocation (code space): 28
 
 /.((?2)(?R)|\1|$)()/
 ------------------------------------------------------------------
-  0  28 Bra
+  0  24 Bra
   2     Any
-  3  18 Once
+  3   7 CBra 1
-  5   7 CBra 1
+  6  19 Recurse
-  8  23 Recurse
+  8   0 Recurse
- 10   0 Recurse
+ 10   4 Alt
- 12   4 Alt
+ 12     \1
- 14     \1
+ 14   3 Alt
- 16   3 Alt
+ 16     $
- 18     $
+ 17  14 Ket
- 19  14 Ket
+ 19   3 CBra 2
- 21  18 Ket
+ 22   3 Ket
- 23   3 CBra 2
+ 24  24 Ket
- 26   3 Ket
+ 26     End
- 28  28 Ket
- 30     End
 ------------------------------------------------------------------
 
 /.((?3)(?R)()(?2)|\1|$)()/
 ------------------------------------------------------------------
-  0  35 Bra
+  0  31 Bra
   2     Any
-  3  25 Once
+  3  14 CBra 1
-  5  14 CBra 1
+  6  26 Recurse
-  8  30 Recurse
+  8   0 Recurse
- 10   0 Recurse
+ 10   3 CBra 2
- 12   3 CBra 2
+ 13   3 Ket
- 15   3 Ket
+ 15  10 Recurse
- 17  12 Recurse
+ 17   4 Alt
- 19   4 Alt
+ 19     \1
- 21     \1
+ 21   3 Alt
- 23   3 Alt
+ 23     $
- 25     $
+ 24  21 Ket
- 26  21 Ket
+ 26   3 CBra 3
- 28  25 Ket
+ 29   3 Ket
- 30   3 CBra 3
+ 31  31 Ket
- 33   3 Ket
+ 33     End
- 35  35 Ket
- 37     End
 ------------------------------------------------------------------
 
 /(?1)()((((((\1++))\x85)+)|))/

testdata/testoutput8-8-2

@@ -720,41 +720,37 @@ Memory allocation (code space): 10
 
 /(((a\2)|(a*)\g<-1>))*a?/
 ------------------------------------------------------------------
-  0  57 Bra
+  0  51 Bra
   3     Brazero
-  4  48 SCBra 1
+  4  42 SCBra 1
-  9  40 Once
+  9  18 CBra 2
- 12  18 CBra 2
+ 14  10 CBra 3
- 17  10 CBra 3
+ 19     a
- 22     a
+ 21     \2
- 24     \2
+ 24  10 Ket
- 27  10 Ket
+ 27  16 Alt
- 30  16 Alt
+ 30   7 CBra 4
- 33   7 CBra 4
+ 35     a*
- 38     a*
+ 37   7 Ket
- 40   7 Ket
+ 40  30 Recurse
- 43  33 Recurse
+ 43  34 Ket
- 46  34 Ket
+ 46  42 KetRmax
- 49  40 Ket
+ 49     a?+
- 52  48 KetRmax
+ 51  51 Ket
- 55     a?+
+ 54     End
- 57  57 Ket
- 60     End
 ------------------------------------------------------------------
 
 /((?+1)(\1))/
 ------------------------------------------------------------------
-  0  31 Bra
+  0  25 Bra
-  3  25 Once
+  3  19 CBra 1
-  6  19 CBra 1
+  8  11 Recurse
- 11  14 Recurse
+ 11   8 CBra 2
- 14   8 CBra 2
+ 16     \1
- 19     \1
+ 19   8 Ket
- 22   8 Ket
+ 22  19 Ket
- 25  19 Ket
+ 25  25 Ket
- 28  25 Ket
+ 28     End
- 31  31 Ket
- 34     End
 ------------------------------------------------------------------
 
 "(?1)(?#?'){2}(a)"
@@ -771,45 +767,41 @@ Memory allocation (code space): 10
 
 /.((?2)(?R)|\1|$)()/
 ------------------------------------------------------------------
-  0  42 Bra
+  0  36 Bra
   3     Any
-  4  27 Once
+  4  11 CBra 1
-  7  11 CBra 1
+  9  28 Recurse
- 12  34 Recurse
+ 12   0 Recurse
- 15   0 Recurse
+ 15   6 Alt
- 18   6 Alt
+ 18     \1
- 21     \1
+ 21   4 Alt
- 24   4 Alt
+ 24     $
- 27     $
+ 25  21 Ket
- 28  21 Ket
+ 28   5 CBra 2
- 31  27 Ket
+ 33   5 Ket
- 34   5 CBra 2
+ 36  36 Ket
- 39   5 Ket
+ 39     End
- 42  42 Ket
- 45     End
 ------------------------------------------------------------------
 
 /.((?3)(?R)()(?2)|\1|$)()/
 ------------------------------------------------------------------
-  0  53 Bra
+  0  47 Bra
   3     Any
-  4  38 Once
+  4  22 CBra 1
-  7  22 CBra 1
+  9  39 Recurse
- 12  45 Recurse
+ 12   0 Recurse
- 15   0 Recurse
+ 15   5 CBra 2
- 18   5 CBra 2
+ 20   5 Ket
- 23   5 Ket
+ 23  15 Recurse
- 26  18 Recurse
+ 26   6 Alt
- 29   6 Alt
+ 29     \1
- 32     \1
+ 32   4 Alt
- 35   4 Alt
+ 35     $
- 38     $
+ 36  32 Ket
- 39  32 Ket
+ 39   5 CBra 3
- 42  38 Ket
+ 44   5 Ket
- 45   5 CBra 3
+ 47  47 Ket
- 50   5 Ket
+ 50     End
- 53  53 Ket
- 56     End
 ------------------------------------------------------------------
 
 /(?1)()((((((\1++))\x85)+)|))/