|Subject:||Setting a sane default stack_depth|
|Date:||Mon, 02 Jul 2007 00:58:33 -0400|
|To:||bug-WWW-Mechanize [...] rt.cpan.org|
|From:||Jmax <jmax [...] gnaa.us>|
I think that the stack_depth option of mechanize (which controls the amount of data held in memory regarding previous requests and the state of the mech object after that request) should be set to a value considerably more sane that 0 (infinite). One reason for this is that (on FreeBSD at least), the resulting crashes and error messages caused by the process using up too much memory are rather cryptic, and can lead the developer (or possibly end user) frustrated. I'm sure there are many other reasons. Although I have done no research into a possible value, I can't imagine that any person would need to $mech->back or debug further than the past 10 or 15 requests; 20 at the most. Thanks.