🚀 go-pugleaf

In article ralpht@meaddata.com (Ralph W. Trickey) writes:
>In article vincent@datmuc.UUCP (vincent) writes:
            vincent@datmuc.dat.de
>>In article smithbm@craft.camp.clarkson.edu (Azrak) writes:
>>>c.brennan@student.utas.edu.au (chris brennan) writes:
>>>>In article smithbm@craft.camp.clarkson.edu (Azrak) writes:
>>>>> vincent@datmuc.UUCP (vincent) writes:

>Suggestions
>1) My attempted mail to you bounced.
	I write in my articles and mails that my address is
	vincent@datmuc.dat.de and that my network manager said he can't
	build me a proper return path.
>2) you don't need the CODESEG in the tsrstack declaration, it may be
>   confusing the compiler.
	CODESEG if just so all my data is in the same segment as my CODE
	and so I don't have to keep track of DS (for now). I also heard
	that for a tiny model everything is in the same segment at the end
	anyway witch is my goal (.COM format).
>3) you need to keep a flag or counter for int 8, it is quite possible
>   for an int 8 to happen while your int 8 code is executing.
	Are you shure I tough that the interupts were disable when inside
	one already. And anyway 1/18 of a second is a lot a time.
>4) your calculation for malloc is off, you allocate tsrstack to be
>   0..1023, then assign the stack to be at 1024.
	Took the code from Schulam and I'm not shure if I/you understand
	the way the stack works (top->down I think) Ok maybe you're wright
	and the stack should be at 1023. I'll will try it.
>5) you might want to simply declare an array or use some other method
>   for keeping the stack, the malloc method doesn't strike me as being
>   very stable accross compilers and compiler releases.
	I did try, same problem. But I think also that malloc is free
	to allocate memory where it wants and not necesary next to you,
	so to calculate the size of staying resident migth be off. MS
	in there TSR example show a different way.  And off course the
	program may pop up with the current foreground stack. You just
	don't know the what you get and it's not recomended .
>6) alt.msdos.programmer is not a valid group
	Ho realy, fooled me. Lots of article go there. Ok Less volume.

>Good Luck.
	Thanks, I'm currently debuging the disinstall routines and I need it.


vincent@datmuc.dat.de

ciao!