Well I was mistaken, it seemed to work when the
program that I loaded was small enough to preserve the
previous memory data. So Jim was correct in saying
that the contents get unloaded.
Jim - you mentioned the undocumented extension
services any hints on how I could go about digging
into these, some pointers might help.
I will try allocating the function on the heap and see
if that works in the meantime.
Jay
--- Jay Shroff <[EMAIL PROTECTED]> wrote:
> Jim and everyone else, I suspected that the my
> replacement for the trap was probably being unloaded
> from memory, however I am not sure if that is
> entirely
> true.....
>
> Now I might be mistaken but, I just got my code to
> work now (I had a slight problem - in that I was
> really hammering away at the keyqueue and that's
> what
> was causing the crash. I have since fixed that
> problem.)
>
> I don't have the code handy but Monday I will post
> what "seems" t work and would appreciate any
> comments
> on why this works. As based on your's and everyone
> elses comments it should not work.
>
> Thanks again for all the info, it's been really
> helpful.
>
> Jay
> --- Jim Schram <[EMAIL PROTECTED]> wrote:
> > At 12:55 AM +0100 2001/07/07, Igor Mozolevsky
> wrote:
> > >So, how does one get the OS not to unload the
> code
> > for a particular function when the app quits?
> >
> > You don't. And you don't leave traps patched after
> > your application exits.
> >
> > However, there are several methods available to
> work
> > around this. For example, copy the patch
> routine(s)
> > into an allocated [dynamic heap or feature]
> pointer
> > so that neither the code nor the allocated memory
> > block is part of your application. Better yet,
> write
> > a system extension (not documented, sorry) which
> > loads during the boot sequence and permanently
> > patches the trap (like System Updates do).
> >
> > The latter is the best way to avoid problems
> without
> > using a patch chain manager such as Hackmaster. To
> > understand why, you merely need to think through
> the
> > following [extremely common] situation: A and B
> are
> > patches to the same trap. A is applied first. Then
> B
> > is applied. Then A is removed. The trap can no
> > longer be called without crashing. Hackmaster (et
> > al.) remap the patch chain to avoid this
> situation.
> > This is why patching should *only* be done using
> one
> > of the third-party trap managers.
> >
> > Regards,
> >
> > Jim Schram
> > Palm Incorporated
> > Partner Engineering
> >
> >
> > --
> > For information on using the Palm Developer
> Forums,
> > or to unsubscribe, please see
> http://www.palmos.com/dev/tech/support/forums/
>
>
> --
> For information on using the Palm Developer Forums,
> or to unsubscribe, please see
http://www.palmos.com/dev/tech/support/forums/
--
For information on using the Palm Developer Forums, or to unsubscribe, please see
http://www.palmos.com/dev/tech/support/forums/
