Pic16/Pic18/PIC24/dsPIC compiler

[Rocky]

I was ready to buy the compilers for all the pic families but I read a lot about problems with the compiler on this list.

Questions:
-Are those problems (bugs) still not fixed?
-Are the problems across the CCS line of compilers for the PIC processor families?

Thank you.

[FvM]

I'm only aware of PCD issues. PCD is continuously improving, but some bugs related to complex int32 expressions and int8 arithmetics are still continued during the last iterations. I would be also surprized, if all built-in functions related to specific PIC24 hardware, e.g. ADC or CRC unit had become 100% error-free now. Some problems are also related to non obvious restrictions of the Microchip hardware.

When you compare PCD to competitors, particularly C30, the built-in functions part is unrivaled anyway. So if some manual programming is necessary here, it wouldn't be a problem to my opinion. The bad thing is of course, if a function is said to work correctly, but it doesn't. Examples from the last weeks forum discussion (related to previous PCD version bugs, that have been fixed) are e.g. a #pin_select accessing the wrong register. Truely bad, if you just trust the compiler.

For the arithmetic bugs, I unintentionally found the MD5 code (a simple cryptographic application) as a test case. It took me some effort, to find workarounds for all bugs in V4.081. They are almost removed until one. If the original code (from the respective RFC) compiles correctly, I'll issue an update to my post http://www.ccsinfo.com/forum/viewtopic.php?t=36359.

[Rocky]

Thank you FvM, it makes me feel better.
I don't like to change between compilers when changing processors of the same supplier. Bad enough to have to change from one supplier to the other.

If I order the PCWHD with its own IDE, I assume that I can still use the compilers with MPLAB the same way as I would use the commmand line versions. Is this a correct assumption?
_________________
PERSEVERANCE IS THE KEY TO SUCCESS