Re: [avrdude-dev] Re: avrdude 4.2.0 Release

[Jan-Hinnerk Reichert]

On Friday 05 September 2003 04:33, Theodore A. Roth wrote:
> On Thu, 4 Sep 2003, E. Weddington wrote:
> :)Hello All,
> :)
> :)[ Any and all apologies. I'm sending this out individually as the
> : avrdude )list has been slow of late.]
> :)
> :)Brian and I were hoping to get an avrdude 4.2.0 release out soon,
> :)preferably this week?
> :)
> :)AFAIK, these are the only issues:
> :)
> :)1. Patch from Jan-Hinnerk Reichert, concerning fuse bits for
> : AVR910. See )post to list on August 28.
>
> I think this patch breaks the encapsulation of the avr910 module.
> It makes calls to avr_write_byte_default() and
> avr_read_byte_default(). I'm not sure if that is the best route to
> take since if may come back to bite us in the future if there are
> changes made to avr.c.

I have done it this way, because I thought it is the right thing to do 
from an OO point of view. It is like calling the method of the 
parent, if the own one fails.

The idea from Ted to check it in "avr.c" is fine, too. It is a bit 
farther away from the OO-modeling, but fits better in the actual 
source code. So, I vote for Ted's way.

-----

Speaking of OO, it will probably be a good idea to allow a programmer 
to extend the data structure somehow. This programmer specific 
details (like version) can be stored during initialisation and used 
later.

However, I don't think that version checking will do any good, with 
that many patched programmers out there ;-(

Nothing for this release, of course.

BTW: Why is avrdude not written in C++? I just want to know, no flame 
war, please.

> Brian, what's the policy here, if any?
>
> I don't think the patch is functionally wrong though.

Don't think so either. However, I only checked on a non-atmel prommer 
with an ATmega161.

 Jan-Hinnerk