As Eric Weddington wrote:

> I'm sorry, could you clarify this? Are you suggesting that we will
> have to *add* a command-line option to get the smaller code?

Yes, in the same sense as -mint8 produces smaller code when applied.
Except, -mno-mainframe causes way less portability issues.

Marek's suggestion to use __C_task has some merit as well (and is IMHO
also compatible with other compilers, at least with IAR), but would
even require code changes.  Perhaps we could combine them: allow for
the __C_task attribute, and provide a command-line option that would
treat each implementation of function main() as if __C_task were
applied (so no code changes are needed).

> If so, this certainly
> doesn't help backwards-compatibility.

Right now, we are already incompatible (with no option to the users to
get rid of the incompatibility), and surprisingly few users really
noticed it.

> I like the ideal of standards conformance, but the bottom line is
> that I lean towards practicality and no code bloat.

Both are valid goals.  We also need to be allowed to fix old mistakes.
See above: other compilers also require you to specifically request
that code-saving features.

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)


_______________________________________________
AVR-libc-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/avr-libc-dev

Reply via email to