Subject: | limiting execution time |
Thanks for the awesome code!
We (blekko) use Devel::Size in an in-process debugger in all of our
daemons, which lets our devs look at data structures in our running
production daemons. Occasionally devs do something dumb like asking for
the size of a multi-gigabyte hash with lots of tiny things in it. This
takes long enough that the daemon stops processing events for too long
and bad stuff happens.
It's easy to avoid this by limiting the number of things that a single
call to Devel::Size looks at, signaling overflow if the limit is reached.
We're happy to write up a patch, but if you guys have an Opinion as to
what the interface should look like, we're all ears.