Problems using CCS loader

[eric]

Hi,
I did have the CCS loader.c working on a 18F525 in 2003. But stopped functioning in the spring 04. Finally got around to looking at it.

What I found in reading back the "bootloaded" hex file that the loader starts writing in FF's. Its pretty consistent where it is doing it.

10016000F5501200FFFF204E4F54204D594C4F4187
:10017000440A0D00FFFFEA6A070EE96EEF50D8B49B
:10018000CFEF00F00A0E016E006A002EFED7012E9E
:10019000FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF6F
:1001A000FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF5F
:1001B000FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF4F

Noting some possible problems with CCS write_program_memory I
rolled my compiler version from 3.191 back to 3.188, but this did not help

Any suggestions welcome

[ckielstra]

What you are showing is part of a hex-file. Where did you capture this hex-file, is this the file before loading it into memory?

In what respect is this file different from the hex-file generated by the compiler?

[Guest]

Yes, this is the read back hex file. On the second line should read

:10018000CFEF00F00A0E016E006A002EFED7012E9E
:10019000FBD75F0E006E002EFED7EF2EF3D7000CBC

not

:10018000CFEF00F00A0E016E006A002EFED7012E9E
:10019000FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF6F

I had noticed other posts regarding bootloaders writing FF's that might be related to the write_program_memory function. Though I would ask before I try to discover other reasons for this problem.

The code for this hex file pretty much just toggles a pin on port C, outputs a few lines of text and then call the boot loader again that is loaded into high mem.

[ckielstra]

Sorry, I still don't understand completely. You say this is 'the read back hex file'. You mean this is the data that is in your PIC's memory? What tool do you use for creating this 'read back'?
I use the ICD2 in-circuit debugger to inspect memory, but that doesn't give me hex-files.

I'm just trying to figure out if the error is in the programming or in the reading.