local variables lose values ?

Hello,

I have a project (C166/HUGE memory model), where some local variables seems to lose their values. Everything worked fine, until I added some more code to a specific function (and one more local variable):

unsigned char Ptmp=0,FSzu=0,IsCancelled=1,errors=0,
FSDin,f3=0,f4=0,Abort=0,ii,done=0;

"done=0" was added to the list. Now, the compiled program seems to have a problem with the variable IsCancelled (set to 1 initially, but whenever referred to it in the function, it returns 0).

The .LST file shows this now (IsCancelled is not "working" anymore):

NAME            CLASS    SPACE  TYPE   OFFSET   SIZE
----------------------------------------------------
  Ptmp . . . .  --reg--         uchar  -----    1
  FSzu . . . .  auto            uchar     2H    1
  IsCancelled.  auto            uchar     0H    1
  errors . . .  auto            uchar     6H    1
  FSDin. . . .  auto            uchar     8H    1
  f3 . . . . .  auto            uchar     AH    1
  f4 . . . . .  auto            uchar     CH    1
  Abort. . . .  auto            uchar     EH    1
  ii . . . . .  --reg--         uchar  -----    1
  done . . . .  auto            uchar     2H    1

The old .LST file, when everything was still working, showed this:

NAME            CLASS    SPACE  TYPE   OFFSET   SIZE
----------------------------------------------------
  Ptmp . . . .  auto            uchar     0H    1
  FSzu . . . .  auto            uchar     2H    1
  IsCancelled.  --reg--         uchar  -----    1
  errors . . .  auto            uchar     6H    1
  FSDin. . . .  auto            uchar     8H    1
  f3 . . . . .  auto            uchar     AH    1
  f4 . . . . .  auto            uchar     CH    1
  Abort. . . .  auto            uchar     EH    1
  ii . . . . .  --reg--         uchar  -----    1

Ok, the class of the variable in question has changed, but it should still work, shouldn't it?
What could cause the problem, and how would one prevent this from happening?

Any help would be great!

Thank you in advance.
Holger

Parents
  • Hello again,

    I forgot to mention that I am using version 3.12 of the C166 package.

    Playing around with some options in MicroVision, I found that it seems to work fine again after using "Use static memory for non-register automatics" in the compiler options.

    This gives me an .LST file showing:

    NAME            CLASS    SPACE  TYPE   OFFSET   SIZE
    ----------------------------
      Ptmp . . . .  auto     NDATA  uchar     0H    1
      FSzu . . . .  auto     FCODE  uchar     2H    1
      IsCancelled.  auto     NDATA  uchar     2H    1
      errors . . .  auto     FCODE  uchar     6H    1
      FSDin. . . .  auto     FCODE  uchar     8H    1
      f3 . . . . .  auto     FCODE  uchar     AH    1
      f4 . . . . .  auto     FCODE  uchar     CH    1
      Abort. . . .  auto     FCODE  uchar     EH    1
      ii . . . . .  auto     NDATA  uchar     4H    1
      done . . . .  auto     NDATA  uchar     6H    1
    

    What happens here, why does it not work in the first version, and more interesting would be: Why does the compiler not complain about anything that causes malfunctions in the compiled program?

    Thank you!
    Holger

Reply
  • Hello again,

    I forgot to mention that I am using version 3.12 of the C166 package.

    Playing around with some options in MicroVision, I found that it seems to work fine again after using "Use static memory for non-register automatics" in the compiler options.

    This gives me an .LST file showing:

    NAME            CLASS    SPACE  TYPE   OFFSET   SIZE
    ----------------------------
      Ptmp . . . .  auto     NDATA  uchar     0H    1
      FSzu . . . .  auto     FCODE  uchar     2H    1
      IsCancelled.  auto     NDATA  uchar     2H    1
      errors . . .  auto     FCODE  uchar     6H    1
      FSDin. . . .  auto     FCODE  uchar     8H    1
      f3 . . . . .  auto     FCODE  uchar     AH    1
      f4 . . . . .  auto     FCODE  uchar     CH    1
      Abort. . . .  auto     FCODE  uchar     EH    1
      ii . . . . .  auto     NDATA  uchar     4H    1
      done . . . .  auto     NDATA  uchar     6H    1
    

    What happens here, why does it not work in the first version, and more interesting would be: Why does the compiler not complain about anything that causes malfunctions in the compiled program?

    Thank you!
    Holger

Children
More questions in this forum