More PIC24 and PCD issues

Markdem · Joined: 24 Jun 2005 Posts: 206

Hi All,

Still trying to port some code to a PIC24 using CCS 4.124. I can't believe it is so hard.

Anyway, I have an Ethernet application working in PIC18. All I have changed in the stack is what timer the tick function uses and what pins are used to connect to the ENC28J60. Nothing else has changed. Now when I try to compile, I get the following error:

asmallri · Joined: 12 Aug 2004 Posts: 1635 Location: Perth, Australia

It may be a case sensitive issue.

Checkout #case
_________________
Regards, Andrew

http://www.brushelectronics.com/software
Home of Ethernet, SD card and Encrypted Serial Bootloaders for PICs!!

Markdem · Joined: 24 Jun 2005 Posts: 206

Hi Asmallri,

Tried with #case just in case (no pun intended) but that did not help.

I have now changed the code back to PIC18 by only changing the device line at the top of main and changing the device in MPLAB. I have left the timer and pin defines configured for the PIC24 just in case that was the issue somehow and it complies fine. I just looks like PCD is doing something silly.

What else can i try to fix the issue? does the SCR=32031 at the end of the error message mean anything useful?

Thanks

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

Markdem · Joined: 24 Jun 2005 Posts: 206

Hi FvM,

The function is defined in arp.c with gets included from stacktsk.c.

Like I said, this project works fine when complied with PCH:

Markdem · Joined: 24 Jun 2005 Posts: 206

I am going to lose it today.

I changed the function name from ARPGet to STUPID and guess what, it now works!

Ttelmah · Joined: 11 Mar 2010 Posts: 19524

OK.
Prototypes have nothing to do with how often a function is used.
They have two separate 'functions'.

The first is to allow the actual definitions to be elsewhere (another module, or further down the code), so that the compiler knows what to 'expect'. In C, all things must be declared before they are used, and having a prototype allows you to make sure this is the case. Original 'C', did not have 'prototypes' as such, requiring the declaration, but not requiring the variable types etc. to be shown.
This is their only real function in CCS.

The other part of their function, is as compilers have got more sophisticated, and support things like overloading, having the variable types declared in the prototype allows checking that types do match. CCS checks that the declarations in the prototype 'match' those in the actual definition (more strictly than is really required), but then does no checking about what is actually passed/used. This is down to the language, rather than CCS, with C, not being an explicitly 'typed' language, so it is perfectly legitimate to pass (say) an int to a function expecting a 'float', while on most more complex languages this has to be explicitly handled.

My guess would be that there is a tiny typing error somewhere in the PCD prototype declaration. Possibly even a 'hidden' character, or a difference in the variable declarations from the real function (the easiest way to get prototypes 'right' is to cut and paste the first line from the declaration, and just add the ';' in place of the '{'). Possibly the most likely thing, is related to the difference in size of a default integer in PCD (signed int16), with something using 'int' in one place and explicit sizing elsewhere....

Best Wishes