| From: | Martin Baute |
| Date: | 1 Sep 2000 at 12:36:55 |
| Subject: | Re: AMIOPEN: NO MORE "SLEEP 1"!!!! |
Hello Patrick
On 01-Sep-00, you wrote:
> Deryk Robosson wrote:
>> eww..that's more ugly than the sleep(1); :)
> It may seem like it at first, but here's what I'm thinking the AVE object
> values are:
>
> ip+64 is some sort of pointer to another object, possibly the
> main AVE window. It's always NULL when AVE hasn't been initialized
> then stays constant throughout the session for any program.
>
> ip+52...
<snip>
> So basically, it's:
>
> Is there a pointer to the main window?
> YES: Continue
> NO: Wait until the background process creates all 9 of the
> main window targets before going on.
No, basically it�s (IMHO) just the kind of code that made Kickstart v2.0
such a pain in the first time... :-/
I don�t know that much about VP coding yet, but shouldn�t everybody
keep his / her hands off such internal data structures?
Yes, there may be no other way (yet) to achieve certain effects. But
if we do "ugly" (hardcoded) workarounds like these, doesn�t that mean
our code will stop working correctly the very moment such data
structures are changed internally?
Is VP so much different from C++? If I have a pointer to a data structure
(object), but no way (method) to access data members of that structure,
doesn�t that mean I am not allowed to access them?
Just to put things back into perspective...
> The only ugly side effect is that as the AVE versions change, there may
> be more or less than 9 targets allocated...
Couldn�t the data structure itself be changed, rendering your code
absolutely unusable?
I might be absolutely wrong on this, so please don�t flame...
Regards
Subscribe/Unsubscribe: open-request@amiga.com
Amiga FAQ: http://www.amiga.com/faq.html