Qb hooks int vectors

 BBS: Inland Empire Archive
Date: 03-11-92 (02:20)             Number: 88
From: STEVE HALKO                  Refer#: NONE
  To: ALL                           Recvd: NO  
Subj: Qb hooks int vectors           Conf: (2) Quik_Bas

  There's been some discussion here lately about which interrupt vectors
  are hooked by QB programs.  Since I would much rather code than sleep,
  I did a little research tonight.  I wrote a QB program to dump the
  entire interrupt vector table to a file.  I ran this program both in
  the IDE and as an EXE.  I then created a 3rd file by porting the ouput
  of a RAM viewing utility I have. (For this utility, I know which
  interrupts are hooked, since I have the source.)

  By comparing the three resulting files, I came up with these results:

  Program run from the IDE hooks INTs 0, 2, 4, 9, 10, 23, 24, 34 thru
  3D, 3F, EF, and F0.

  Program run as a stand alone EXE hooks INTs 0, 4, 24, EF, and F0.

  Note that there may be other interrupts that may be hooked by
  different programs, e.g. programs using TIMER may very well hook INT 8
  (just a guess).

  Also, my program did not use floating point.  If it did, then even the
  stand alone EXE would have hooked INTs 34-3D and 3F, (the floating
  point emulation interrupts).

  So it looks like using INT 21h, Function 4Ch to exit a QB program
  would do the most harm if you ran it from the IDE.  If running it as
  an EXE, you would only hang the machine if another program called INT
  0 (Divide by Zero), INT 4 (Overflow), or INT 24 (Critical Error).
  INTs EF and F0 are only used by BASIC programs to store the old values
  of INTs 9 and 8, respectively, so its probably not important whether
  they are restored or not when BASIC is through with them.

 * SLMR 2.1a * Got kleptomania? Take something for it.

--- DB B1061/002487
 * Origin: Gulf Coast BBS -QuickSHARE #2- (904)563-2547 HST/V.32bis (1:365/12)