From spp@ds.net

Sorry for the duplicate message, I want this to go into perlbug.

This looks to me like a forgotten 'break' statement in regcomp.c. I came
across it while looking at the earlier question of vertical tab.

This is against 5.6.0 stock.

-spp

--- perl-5.6.0/regcomp.c        Tue Mar 14 17:19:44 2000
+++ perl-5.6.work/regcomp.c     Tue Jul 18 16:12:54 2000
@@ -2802,6 +2802,7 @@
                        if (strnEQ(posixcc, "space", 5))
                            namedclass =
                                complement ? ANYOF_NSPACE : ANYOF_SPACE;
+                       break;
                    case 'u':
                        if (strnEQ(posixcc, "upper", 5))
                            namedclass =

From @jhi

On Tue, Jul 18, 2000 at 04​:25​:17PM -0400, Stephen P. Potter wrote​:

Sorry for the duplicate message, I want this to go into perlbug.

This looks to me like a forgotten 'break' statement in regcomp.c. I came
across it while looking at the earlier question of vertical tab.

This is against 5.6.0 stock.

Thanks. (It seems that the test suite is not wide enough, grumble...)

From @vanstyn

In <20000724051355.E27450@​beta.hut.fi>, Jarkko Hietaniemi writes​:
:On Tue, Jul 18, 2000 at 04​:25​:17PM -0400, Stephen P. Potter wrote​:
:> Sorry for the duplicate message, I want this to go into perlbug.
:>
:> This looks to me like a forgotten 'break' statement in regcomp.c. I came
:> across it while looking at the earlier question of vertical tab.
:>
:> This is against 5.6.0 stock.
:
:Thanks. (It seems that the test suite is not wide enough, grumble...)

Er, what test would have caught this? The missing break allowed it to
drop through to another string compare that would always fail - slightly
maloptimal, but not such as to cause misbehaviour.

Hugo

From @jhi

On Mon, Jul 24, 2000 at 03​:46​:36AM +0100, Hugo wrote​:

In <20000724051355.E27450@​beta.hut.fi>, Jarkko Hietaniemi writes​:
:On Tue, Jul 18, 2000 at 04​:25​:17PM -0400, Stephen P. Potter wrote​:
:> Sorry for the duplicate message, I want this to go into perlbug.
:>
:> This looks to me like a forgotten 'break' statement in regcomp.c. I came
:> across it while looking at the earlier question of vertical tab.
:>
:> This is against 5.6.0 stock.
:
:Thanks. (It seems that the test suite is not wide enough, grumble...)

Er, what test would have caught this? The missing break allowed it to
drop through to another string compare that would always fail - slightly
maloptimal, but not such as to cause misbehaviour.

Ahhh. Perhaps I should just get some sleep.

@iabyn - Status changed from 'open' to 'resolved'