diff options
author | Karl Williamson <public@khwilliamson.com> | 2011-12-31 21:48:53 -0700 |
---|---|---|
committer | Karl Williamson <public@khwilliamson.com> | 2011-12-31 21:48:53 -0700 |
commit | 167e651f9a7445679df315ec089c3adf7b383552 (patch) | |
tree | 066353db068aa280fcaa6f44cf01e2f6a9b29bfd | |
parent | 2373c1b5d1361ab8bbe954fc8234512fd554e7e7 (diff) | |
download | perl-khw/newtricky.tar.gz |
regcomp.c: remove invalid assertionkhw/newtricky
It turns out that a node can be an EXACTFU_SS upon entry to join_exact()
contrary to what I thought. One way, anyway, for this to happen, is
through the (?&NAME) construct which copies a pre-existing node.
-rw-r--r-- | regcomp.c | 3 |
1 files changed, 1 insertions, 2 deletions
@@ -2631,10 +2631,9 @@ S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan, IV *min_change, boo #endif DEBUG_PEEP("join",scan,depth); - /* These opcode should only be on output from this routine, never on input + /* This opcode should only be on output from this routine, never on input */ assert(OP(scan) != EXACTFU_NO_TRIE); - assert(OP(scan) != EXACTFU_SS); /* Look through the subsequent nodes in the chain. Skip NOTHING, merge * EXACT ones that are mergeable to the current one. */ |