Fix \K within recursion bug in interpreter.

ChangeLog

@@ -35,6 +35,10 @@ now runs clean and the resulting "configure" seems to work, so I hope nothing
 is broken. Later: the requirement for autoconf 2.70 broke some automatic test 
 robots. It doesn't seem to be necessary: trying a reduction to 2.60.
 
+6. The pattern /a\K.(?0)*/ when matched against "abac" by the interpreter gave
+the answer "bac", whereas Perl and JIT both yield "c". This was because the 
+effect of \K was not propagating back from the recursion. This is now fixed.
+
 
 Version 10.36 04-December-2020
 ------------------------------

src/pcre2_match.c

@@ -818,10 +818,12 @@ fprintf(stderr, "++ op=%d\n", *Fecode);
 
       /* N is now the frame of the recursion; the previous frame is at the
       OP_RECURSE position. Go back there, copying the current subject position
-      and mark, and move on past the OP_RECURSE. */
+      and mark, and the start_match position (\K might have changed it), and
+      then move on past the OP_RECURSE. */
 
       P->eptr = Feptr;
       P->mark = Fmark;
+      P->start_match = Fstart_match; 
       F = P;
       Fecode += 1 + LINK_SIZE;
       continue;

testdata/testinput1

@@ -6429,4 +6429,7 @@ ef) x/x,mark
 /a{65536/
     >a{65536<
 
+/a\K.(?0)*/
+    abac
+
 # End of testinput1 

testdata/testoutput1

@@ -10188,4 +10188,8 @@ No match
     >a{65536<
  0: a{65536
 
+/a\K.(?0)*/
+    abac
+ 0: c
+
 # End of testinput1