nth eval of same code creates errors

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Hi Everyone
Here is a head scratcher. I am using log4perl which contains a few eval statements and sometimes see strange activity - the script (which runs as a POE process) exits after a while (clarification, runs perfectly for a while and then exits) with the error:

Global symbol xxxx requires explicit package name at (eval 226185663) line yyyy.

(xxx being an internal log4perl variable but not always the same one)

Thing is that _many_ iterations of evaluating the code have gone by (like 226185663 above) and then the error is thrown. I can only conclude that log4perl is fine and something else is affecting this.

My questions:
1) I know that the number 226185663 is the eval number (i.e. incremented each time an eval is run), but I can't see how it gets so high - that amount of iterations would take months - is there any way to print this number on the console to see if something else increments it
2) Any idea on the problem at hand? I have lots of scripts using log4perl and only this particular one fails (but only after a long running time)

Thanks

Warren

Comment on nth eval of same code creates errors

Replies are listed 'Best First'.
Re: nth eval of same code creates errors by dave_the_m (Monsignor) on Apr 01, 2008 at 16:43 UTC
If you have the ability to attach a debugger like gdb to the process, then you can look at the value of the PL_evalseq variable (or on a threaded Perl, the value of my_perl->Ievalseq). For example in the following, perl588 and perl588t are unthreaded and threaded versions of perl: $ perl588 -e 'eval "1;" for 1..10; sleep 9999' & [1] 8491 $ gdb `which perl588` 8491 ... (gdb) p PL_evalseq $1 = 0xa (gdb) quit $ perl588t -e 'eval "1;" for 1..10; sleep 9999' & [1] 8525 $ gdb `which perl588t` 8525 ... (gdb) bt #0 0x00110402 in __kernel_vsyscall () #1 0x003f2e20 in __nanosleep_nocancel () from /lib/libc.so.6 #2 0x003f2c6f in sleep () from /lib/libc.so.6 #3 0x08133397 in Perl_pp_sleep (my_perl=0x8837008) at pp_sys.c:4589 ... (gdb) frame 3 (gdb) p my_perl->Ievalseq $1 = 0xa (gdb) [download] Dave.	[reply] [d/l]
Re^2: nth eval of same code creates errors by warrenvw (Initiate) on Apr 02, 2008 at 08:00 UTC
Thanks Dave. I don't have gdb, but that gave me an idea - I'll report my findings back soon. Warren	[reply]
Re: nth eval of same code creates errors by warrenvw (Initiate) on Apr 01, 2008 at 12:02 UTC
Doh! Anoymous Monk would be me not logged in...	[reply]