Olin Lathrop wrote...

>Dave Dilatush wrote:
>> But I've found that if I define ANOTHER routine named
>> ___floatsisf in the my_source_2.s file and declare it global with
>> the .global directive, the project will assemble and link without
>> any error message at all.

>Hmm.  I thought it did barf on multiply defined globals.  

It barfs just fine if I create multiply-defined globals in my own
code, but it appears quite content to have me stomp all over
library symbols without even warning me that I'm doing so.

I can understand Dave Tweed's point about permitting selected
library routines to be redefined without having to re-compile the
entire library; but I would think there'd be some protection
against unintentionally doing that-- if nothing other than a
warning that says, "FYI, your symbol 'xxxxx' replicates a global
symbol already defined in library file 'yyyyy'."

>Are the library
>routines possibly defined with the WEAK (something like that) attribute?
>It's purpose is to explicitly allow the behavior you describe.

Nope.  According to the results of a dump with the pic30-nm.exe
utility, all of these are global symbols in code sections; none
are listed as weak.

>> This seems more than a bit "unclean", and I'd appreciate any
>> thoughts on how to deal with it to avoid "running over"
>> library-defined symbols by inadvertently re-defining one or more
>> of them in my own code.

>Don't use the Microchp libraries?

Know a good substitute?

Math libraries are a real boon to me: of all the programming
tasks in assembly language, I find math operations by far the
hardest to code without making a ton of mistakes-- or else
slowing to a crawl to avoid them.  Everything else is relatively
easy for this aging, opamp-jockey brain of mine, but I have a
great deal of difficulty coding math without royally garfing it
up.

If anyone knows of a dsPIC30F math library that's comprehensive,
reasonably well-documented, and doesn't cost an arm and a leg,
I'm certainly interested.  Otherwise I'll use the Microchip
library-- AFTER doing a lot of testing, of course...

Dave D.

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist