On Mon, 2 Sep 2002 20:07:53 UTC [EMAIL PROTECTED] (eric w) wrote: > good programming practices would dictate a way out should there be memory > leaks or similar type bugs in the future. this is apparantly conflicting > with non-programmer expectations.
Not to be a smartass, but good programming practice tests for and eliminates memory leaks. Now, in something the size of Mozilla that's not a simple task or a single iteration of the code but there are enough eager testers here to make it a worthwhile and achievable goal. As long as the result can be achieved manually (via -kill or even a rude termination) I see no advantage to automagically killing the system only to restart it. Even if I were severely memory constrained I would prefer to accept reasonable leakage of resources for the duration of, say, a day in return for the previously demonstrated startup speed so long as a shutdown at the end of a period of activity would clear the losses. On the other hand, if a manual shutdown and restart won't clear the leakage, then it's doubtful that the automatic restart is going to be of any use anyway. As it is, the near total hang upon close rather than at my convenience makes the turbo option essentially useless as it becomes more of a nuisance than the slow start it's supposed to eliminate. Having done considerable programming for both Windows and OS/2 I can appreciate the necessity under Windows due to the way resources are managed but I would appreciate an explanation of the value of the automatic vs. manual re-start under OS/2 - I may be missing something that would be of value to me in other ways. -- Will Honea <[EMAIL PROTECTED]>
