Re: pexec bug in MiNT?

[Katherine L Ellis <kellis@primenet.com>]

On Tue, 4 Aug 1998, Brent Thirsk wrote:

>Katherine L Ellis wrote:
>> 
>> hi, was fiddling with pexec in my program, and can't get it to work
>> stable. But the weird thing is that under magic pexec works like it
>> should. No matter the code, (0, or 100) don't matter to me so far it
>> if waits after the child to finish. Am I the only one programming for
>> MiNT?
>Is this a problem with the new kernel.
> Which version.
>I use 1.14.7 with no pexec problems. I have 5 menu programs that each
>spawn a child via pexec, one on each terminal. 
>If this is a problem that occurs after running a child process and 
>exiting, and you then get bombed when trying to run another process,
>the problem is likely that the child did not terminate properly.
>Must be pterm(). 
>If you are using any built in exit functions, 
>depending on what language you are programing in, DON'T. GFA for 
>example, does not use pterm() and will not release all memory control
>blocks back to the system. 
>If you terminate the parent process, restart it, and can pexec a child 
>ok once, that is probably your problem, improper child termination.

Hi Brent, 
1) I use beta 1.15 latest . Might be a diff with 1.14.7, Do I need to
force a Pterm after the pexec?
Do I need to really check the return status of the child? if yes with
pwait?
2) The child I pexec is munpack 1.5. Which seeems to be buggy anyway.
Stack problem. This was compiled with purec I think. I couldn't fixstk
it. Anyone has a gnuc compiled version of it?
3) Do I need to allocate extra ram for the child? even if I use pexec
mode 0?
Thanks.
________________________________________________________________________________
Email:Kellis@primenet.com
Web http://www.primenet.com/~kellis/
________________________________________________________________________________