problems with 16F690

[pmuldoon]

I've had some strange problems programming the 16F690 chip using my ver.3.223 compiler. Can anyone tell me if there are some bugs in the ocmpiler specific to that chip? I have programmed many other chips using the same functions, etc and had no problem. In fact, this code was all written and working in a 12f683 before upgrading for more I/O.

The craziness started when I couldn't use:
#use fixed_io(b_outputs=PIN_B7)
the compiler generated an error.

then, I couldn't use:
setup_timer_2(T2_DIV_BY_1, PwmPeriod*2, 1);
setup_ccp1(CCP_PWM);
set_pwm1_duty(PwmOnTimeUs*2);
even though they worked perfectly in the other chip. So I read the datasheet and set it all up thru the registers...old school!

Then my getstring() was misreading input. I finally fixed it by moving
the array "int s[6] " from local to global. The characters were read correctly again.

Then the display of D numbers was garbled, although the LU were fine. Prior to that last change they were displayed correctly. But I was able to get them working by changing from printf("%d",var) to printf("%lu",(int16)var);

I just shipped the demo with my fingers crossed. But I just want to know if these appear to be known compiler issues, or what. As I said, I've done nothing unusual in this program. It was working fine in another chip. And I've written many many programs in many many microchip parts without problems. And I was writing what should have been easy - done it all before - kind of stuff.

Thoughts?

[PCM programmer]

Post a short, but complete test program that demonstrates one of these
problems. Post all #include, #fuses, and #use statements. Post all
variable and constant declarations. The program should be compilable
without errors. Make it very short.

[pmuldoon]

Okay, one at a time. This is the first problem I had, with the error msg in the comment.

#include <16F690.h>
#fuses NOWDT,PUT,NOPROTECT,BROWNOUT,MCLR,INTRC_IO
#device ADC=10
#use delay(clock=8000000)
#use fixed_io(b_outputs=PIN_B7) // this line won't compile for some reason - Line 44(5,74): Option invalid Not a valid port: B_OUTPUTS
#use fixed_io(C_outputs=PIN_C7) // this line works fine

#USE RS232(Baud=19200,XMIT=PIN_B7,RCV=PIN_B5,PARITY=N,BITS=8,ERRORS)

[PCM programmer]

That version is defective. Later versions don't have the problem.
I tested it with vs. 3.249 and it compiles OK.

[pmuldoon]

I was hoping to hear that it was a bad compiler version.
Does it make sense that all of those things could be broken in such a chip specific way? Other chips compiled fine with the exact same code.

Doing the PWM registers by hand and with a datasheet reminds me why I love CCS so much - they let me focus on all my OTHER probems!

[PCM programmer]

That version is defective. See the bug below:

[pmuldoon]

Ha!
Load a register with the correct value... and then clear it!
Why do mistakes always look so much more obvious in hindsight?

Thanks again for the quick research.
You've done in five minutes what I planned on spending an evening doing.
I will upgrade my compiler.