My testing shows no problems when alarm is called with my script, so I'm going to assume my likelihood of running into trouble is right up there with the spontaneous appearance of monkeys.That proves nothing. Signals are not delivered in a deterministic fashion, that is, you don't know the (internal) state of the program when the signal is delivered. It may run without problems a thousand times. You may cross the street blindfolded a thousand times and survive - but that doesn't prove crossing streets blindfolded is safe, or even that the risk is acceptable.
However, the non-Perl route is possible and trivial, so I have good enough reason to code a backup solution using Korn Shell's alarm function.Or install a recent version of Perl.
Abigail
In reply to Re: Safety of using alarm
by Abigail-II
in thread Safety of using alarm
by delirium
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |