I agree with Erik and others. I write most code in C, with Assembler where appropriate.

Years ago I thought it would be useful to convert an assembler routine to C so it could be more portable, generic, "modern" or whatever. I don't even really remember why...

After struggling with the code and going back and forth reviewing the .LST file to make sure that the code created by the compiler was going to work with my hardware, I asked myself, "Why am I spending all this time trying to get the code working, WHEN MY ASSEMBLER ROUTINE ALREADY WORKS?" I think that was the point that I stopped working on the C conversion.

Also note that I have spent debugging time after switching C compilers (and even after upgrading the same compiler) but never after changing assemblers. I have had minor issues with assembler directive syntax and such, but the output of an assembler is much easier to verify:
1. Rename existing hex file
2. Make changes to source file headers, directives, comments...
3. Create new hex file.
4. Do a file compare on the two hex files. If they are identical, no debugging is necessary!

However, this is just my opinion, and Your Mileage May Vary.

-- Tom Bachmann