On 6/22/06, Steve Hay <[EMAIL PROTECTED]> wrote:
demerphq wrote: > On 6/21/06, Steve Peters <[EMAIL PROTECTED]> wrote: >> On Wed, Jun 21, 2006 at 09:07:35AM +0100, Steve Hay wrote: >> > Steve Hay wrote: >> > >Automated smoke report for 5.9.4 patch 28410 >> > >Mugwump.uk.radan.com: Intel(R) Pentium(R) 4 CPU 3.40GHz(~3391 MHz) >> (x86/2 >> > >cpu) >> > > on MSWin32 - WinXP/.Net SP2 >> > > using cl version 12.00.8804 >> > > smoketime 6 hours 57 minutes (average 10 minutes 26 seconds) >> > > >> > >Summary: FAIL(Fm) >> > > >> > >..\regcomp.c(5183) : error C2143: syntax error : missing ';' before >> 'type' >> > >> > Broken by change 28405, fixed by change 28412. >> > >> > "PERL_UNUSED_ARG(depth);" expands to "((void)depth);", so it's the old >> > code before declarations thing which VC++ doesn't allow again. >> > >> > I thought warnings levels for compilers on (some) other systems had >> been >> > set to spot these? Did I dream that? >> > >> >> gcc should, but didn't seem to when I compiled. gcc must not be able to >> recognize this kind of case. > > PERL_UNUSED_ARG cant be used before a GET_RE_DEBUG_FLAGS_DECL;This patch isn't strictly required. They're all cases like: #ifndef DEBUGGING PERL_UNUSED_ARG(arg); #endif GET_RE_DEBUG_FLAGS_DECL; When DEBUGGING that expands to: IV re_debug_flags = 0; GET_RE_DEBUG_FLAGS; and when not DEBUGGING it expands to: ((void)arg); so there's no problem :-)
Yeah, actually i clicked to that this morning. (See yesterdays #irc chat log for why it took till this morning)
However, I agree that it is a little risky having it work purely on the basis of how the #ifdef's work out so putting things the right way round seems like a good move.
That was going to be my "umm, I'm not entirely insane" response so im happy you bring it up first. :-)
Applied as #28414. Thanks!
No, thank you. :-) Cheers, Yves -- perl -Mre=debug -e "/just|another|perl|hacker/"
