BBS: Inland Empire Archive
Date: 11-30-92 (06:04) Number: 323
From: BILL BEELER Refer#: NONE
To: PETER BARNEY Recvd: NO
Subj: Error Level Conf: (2) Quik_Bas
In a message to Bill Beeler <26 Nov 92 11:50> Peter Barney wrote:
>> PB> Regs.ax = &H4C00 + (s AND 255)
>> PB> CALL Interrupt(&H21, Regs, Regs)
>> PB> END SUB
>> I wouldn't....this has been discussed to death over the last two
>> years...you're going to leave some of QB's interrupts hooked and
>> sooner or later you'll end up in lala land...
PB> If it's been discussed to death then I guess there's no way I
PB> could persuade you to tell me why?
Just like I said above...you're doing your clean-up through
DOS and not through QB...QB hooks several vectors on
initialization and these -should- be released before any
bail-out. If any subsquent program attempts to hook any of
those interrupts they'll be going to no where (or some
where but not where you want them to).
Try this...Brent Ashley weeded it out quite a while
back...while MS doesn't condone or condem it, it does work
just fine.
DECLARE SUB ExitWithErrLvl ALIAS "_exit" (BYVAL ErrorLevel%)
ExitWithErrLvl 255
...Bill...
--- GEcho 1.00/beta+
* Origin: Amber Shadow BBS-MultiLine, MultiProblems-HST/V32b (1:203/988)
