on 7/31/02 3:47 AM, Seth Willits at seth at freaksw dot com wrote:
> On 7/31/02 12:28 AM, the NSA intercepted the following message from "Pepe"
> <DaVincent at addcom dot de>:
>
>> I didn't try the suggested tips listed above, because before I do that I
>> want to aks you if someone has made the same experience than me.
>> Additionaly it would be interesting to hear if someone has made it to
>> make these crashes disappear with a workaround or so.
>>
>> PS: I'm sure the crashes don't have anything to do with my code.
>> Moreover I nontheless plan to release a software title compiled with RB
>> 4.5. A good idea or not? Are compiled versions also that unstable?
>
> Good idea. Otherwise you just wasted a bunch of money. Compile applications
> are much much much more stable than the IDE. The IDE crashes a lot because
> of memory leaks that really add up over multiple runs (or simply if a leak
> occurs many times), and for other misc. stuff that no one really knows
> about. There is no way for us to make REALbasic crash proof.
>
RB could be made practically crash proof but it would have to change
substantially. The reason RB crashes is that you debug applications in the
same memory/address space as RB. How many times does Codewarrior crash
because you are doing repeated edit/debug cycles.
Beyond that I think that RB doesn't crash on all memory leaks only those
involving circular references. I know that the number of crashes went down
dramatically when I fixed reference leaks due to the new event bug.
Chris
|