Aevar,

BTW - I wasn't saying that something didn't need fixing, only that it may not 
be just a 5.10+ issue.

-- 
Richard Foley
Ciao - shorter than aufwiedersehen

http://www.rfi.net/

On Saturday 19 July 2008 12:43:59 Richard Foley wrote:
>  On Saturday 19 July 2008 10:07:45 �var Arnfjörð Bjarmason wrote:
>  > 
>  > $ perl5.10.0 -d /tmp/xpl-testcase.pl
>  > main::(/tmp/xpl-testcase.pl:71):        setHandlers();
>  >   DB<1> c
>  > Out of memory!
>  > 
>  > perl5.11.0 (blead) will hang given the same task, both 5.10 and blead
>  > will segfault if "t" (trace) is turned on before "c" (continue).
>  > 
>  I'm not sure if this helps, but I get almost the exact opposite result, in 
>  that with 't' and 'c' under 5.8.8, the program hangs, and with with 5.11.0 
>  with 't' and 'c' the program runs through.  And with neither do I get 
an 'Out 
>  of memory!' message.
>  
>  > I forgot to mention that none of this happens in 5.8.x, i.e. it's a
>  > regression in >=5.10.x.
>  >
>  It may not be as simple as that (unless it's a regression >= 5.8.x ;)
>  Here's a resume of what happens to me (in several cases using the debugger 
>  (perl -d) on your test program:
>  
>  	5.8.8 + c -> runs through.
>  	5.8.8 + t + c -> hangs.
>  
>  	5.11.0 + c -> SEGV at line 65.
>  	5.11.0 + t + c -> runs through to line 119 (end of prog).
>  
>  I've attached the -V and -Dr outputs from both 5.8.8 and 5.11.0, from both 
>  perls, and the output of several runs, for reference.
>    
>  -- 
>  Richard Foley
>  Ciao - shorter than aufwiedersehen
>  
>  http://www.rfi.net/
>