Re: [avr-libc-dev] I intend to roll avr-libc-1.2.5 by tomorrow

[Joerg Wunsch]

As Björn Haase wrote:

> Open question for issue 1) was if adiw may be safely used in the
> library for all of the devices. Since the present state of eeprom
> functions makes use of adiw in the eeprom_read_word, I'd simply
> assume that the present library implementation is correct and I
> could use adiw for replacing the ld x+ .

Even if the present implementation is incorrect, that would be a
different issue then anyway.  I think it's OK to behave the same way
the present implementation does (nobody complained by now).

If there are other issues, we could handle them as a normal bug later
on.

Btw., I still think we should aim for a slightly different approach in
the long run that allows for per-device sub-libraries, together with
per-architecture libraries as we've got them now.  At the same point,
libm could be dropped and integrated into the main library (maintained
as an empty stub library for some time so people who are use to add
-lm will get some time for transition).

As this requires changes to the compiler as well (even though we can
make it backwards-compatible to older compilers by shipping per-arch
libraries the same way we're doing now), this is most certainly the
time to call it avr-libc-2.0 then.

> Changes for issue 2) would simply involve adding a [test]

> and 3) are straight forward. I'll try my very best to post a patch
> with above corrections until tomorrow morning, say 10 `o clock
> German time.?

Even 1800 UTC (2000 MESZ) suffices, I most likely won't find much time
to work on it over the day anyway.  Daytime is family time. ;-)

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

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