in reply to Mail::Internet memory leaks??

A few more peices of information would be helpful:

  1. Does the script ever error out with the "COuldn't send $!" message?
  2. Does the script exit gracefully, or at all?
  3. What are the contents of $newbody, and how large is it?
  4. Relevant to #3, have you tried minimizing a test case where you can send any message at all, by hardcoding all the values?
--Chris

e-mail jcwren

Replies are listed 'Best First'.
RE: (jcwren) RE: Mail::Internet memory leaks??
by geektron (Curate) on Oct 11, 2000 at 22:00 UTC
    A few more pieces of information would be helpful:
    1. Does the script ever error out with the "COuldn't send $!" message?
    Not to my knowledge. All the time I spent testing and debugging (yes, by hand. . . I should probably write something to stress-test it via automatic submissions.)
    2.Does the script exit gracefully, or at all?
    I don't have any error messages from the script. The one I posted comes from sendmail.
    3.What are the contents of $newbody, and how large is it?
    Just text, and the script handles 3 responders, each with text messages ~4K.
    4.Relevant to #3, have you tried minimizing a test case where you can send any message at all, by hardcoding all the values?
    back during coding/testing ( a few months ago now ) yes.

    i should probably, as noted in my response to #1, bang up quick stress-tester and fire ~50 messages at it. see if i can force it to choke. . .

    i really wondered if anyone else has heard of this particular error, so i'd have an extra clue or two . . .

RE: (jcwren) RE: Mail::Internet memory leaks??
by geektron (Curate) on Oct 12, 2000 at 02:42 UTC
    update # 2

    i threw together a quick stress tester (basically a for loop that sends mail ), and after 150 near-simultaneous messages ( with a ~4K body ) all still runs like a charm.

    so, now i'm looking at the processes on the box, i guess. . .