[futurebasic] Re: [FB] memory leak

Message: < previous - next > : Reply : Subscribe : Cleanse
Home   : October 2001 : Group Archive : Group : All Groups

From: Pete <furbies@...>
Date: Fri, 19 Oct 2001 06:15:23 +1000
tedd on 19/10/01 12:26 AM  scrawled:

>> This article defines the term "memory leak."
>> ------------------------------------------------------------------------
>> A memory leak refers to memory that is not released by an application
>> program when the program no longer needs it. This can lead to a crash or to
>> the program displaying an out of memory message.
>> To resolve a memory leak, increase the amount of memory that a program uses
>> (to delay the effect of the leak) or get an updated version of the program
>> that has been corrected.
>> --
>> Pete...                     (the other one)
> Pete:
> Thus far, I have not seen a good written definition (mine, Heather's,
> or your's) for a memory leak -- let alone a good way to counter it.

I was just 'quoting Apple'

Pete...                     (the other one)

> I first came to understand and introduced to what a memory leak was
> when I was programming in C and attended another list. The problem
> arose when the language we were programming in, namely Think C, would
> allow one to allocate memory (malloc) but then when the same memory
> was released (free), the language would not return ALL of the
> allocation -- there would be a minor amount of memory still reserved
> for some unknown reason.
> Usually, this was not a problem -- however, when the operation
> (allocate/release) was done numerous times (i.e., loop), the minor
> amount of memory reserved would eat up available memory and cause a
> crash. This error process was coined a "memory leak" and was beyond
> the programmers ability to fit it. It was a problem with the language
> and hence, my objection to it's use regarding FB.
> Now, that term has apparently broaden to include things far beyond
> the original definition -- however, I believe that it has done so
> because people didn't fully understand the error in the first place.
> enuf said.
> tedd