Re^4: bad die behaviour?

I also don't know why it returns 0. I have tried to force

$! = 1;
[download]

just before the die and the script ends with return code 0 to the shell.
The functions Log and CloseLog are very simple. They write a formatted log line and close the log file.

thanks
Xavi

Comment on Re^4: bad die behaviour? Download Code

Replies are listed 'Best First'.
Re^5: bad die behaviour? by ikegami (Patriarch) on Sep 10, 2007 at 13:26 UTC
Do you have a `$SIG{__DIE__}` handler that calls `exit`? They write a formatted log line and close the log file. Both of which can change `$!`. It's unrelated to the problem you asked about, but a bug nonetheless.	[reply] [d/l] [select]
Re^6: bad die behaviour? by Anonymous Monk on Sep 10, 2007 at 13:38 UTC
No I don't have any signal handler defined. I also tried to print the whole %SIG hash just before the die, to see if any module had set any signal handler that I wasn't aware of, and the only signal handler active was the floating point exception handler. thanks Xavi	[reply]
Re^7: bad die behaviour? by nimdokk (Vicar) on Sep 10, 2007 at 13:58 UTC
You might want to consider using $SIG{__DIE__} instead of the FatalError subroutine. Put what you are doing in the FatalError into the subroutine for the DIE signal handler, evaluate $? (instead of $!) before die'ing. Something like: `$SIG{__DIE__}=sub{ print STDERR "I'm Dyin!: \n"; }; some_command; $error_code=$?>>8; print "Finished running some_command"; die if $error_code > 0;` [download]	[reply] [d/l]