> Maarten,
>
> > > On a similar note, chapter 3.12 says on labels in inline assembler, that
> "In
> > > older versions of the compiler all labels defined within inline
> assembler
> > > code had to be of the form nnnnn$ where nnnn is a number less than 100
> > > (which implies a limit of utmost 100 inline assembler labels per
> function)."
> > > OK, which is "older versions of compiler" and assuming I have the
> up-to-date
> > > version, what are the rules now?
> >
> > I don't know what "older versions" means. The label does
> > not need to be of the form nnnnn$ but these are treated
> > as local labels by the assembler. They can reoccur after
> > any other label. The compiler will generate only local
> > labels above 100 and so all <100 are at your disposal.
> > If you need more you might need to rethink your
> > implementation strategy.
>
> I just thought on this further, looking here and there.
> The "locality" of these labels is constrained in a rather strange way... but
> useful, I admit... and it brings up several other questions:
>
> - where does the compiler use non-local labels, which are the boundaries of
> "locality"? I assume, it uses them _only_ on the function boundaries (and
> for variables and constants and stuff like that, but I meant in the raw code
> area). Correct? If yes, it would mean, that the locality of local labels is
> constrained within a function.
The compiler uses non-local labels for function entry
points only.
> - the example given in chapter 3.12 is incorrect then in two ways: the
> compiler would not insert a non-local label ("clabel") inside a function,
> and the $0003 label would not be viewed by the two parts of asm code as one
> symbol, if clabel is in between.
The example is correct because the C label is translated
to a local label in asm. Using C labels is generally
frowned upon. C labels have function scope.
> - if I use an inline assembler function with non-local symbol, and if that
> inline assembler would be inside a C function with C statements before and
> after (e.g. within a loop), such a non-local symbol would break up the local
> symbols which the compiler put inside the function. This is not a question
> but a statement, I verified this. Whence, the "in older versions" shall be
> removed and the statement is still very valid - the only safe "any label"
> condition I am able to formulate concisely is, that the inline assembler
> should be not surrounded by any C-statements in that function (although this
> might be often the case, I admit).
I would not recommend using a non-local label inside any
C function unless it's naked and fully inline asm.
> - also, I'd reformulate "nnnnn$" which implies a 5-digit number (with
> subsequent "nnnn" implying 4-digit number :-) ) into "n$, where n is a
> positive decimal integer (with or without leading zeros), or something like
> that
It doesn't really matter but SDCC always generates 5
digits.
> Just an anecdotal mention, the assembler's documentation says, the maximum
> number ("nnnnn") of local symbols is 255. So I looked into the sources and
> found this wonderful comment just at the right place:
>
> /* sandeep changed to 'int' from 'char' */
> /* this will increase the number temp symbols
> that can be defined from 255 to INT_MAX */
>
> :-)
>
> JW
Yes, that's what's in the asxxxx docs. If you look in
subversion you might notice that nobody ever dared
updating that documentation except me just once. But
there were more changes made to the assembler/linker I
guess.
Maarten
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Sdcc-user mailing list
Sdcc-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sdcc-user
