[avrdude-dev] Windows port (was: Butterfly proposal: Exit upgrade function)

[Jan-Hinnerk Reichert]

On Tuesday 03 February 2004 17:03, Martin Thomas wrote:

> I've planned to announce it later, but - whatever: I've ported
> AVRDUDE 4.3.0 to work on Windows without the need of
> the cygwin.dll. Some changes where needed in the source code
> and I don't know if everything is correct. Feel free to test it.
> I've tested with the STK500 and a mega16 and have positive
> feedback from someone using the STK500 and mega8515s.
> The source is based on V4.3.0 and does not include the latest
> changes from the savannah cvs done bei "Henni" (Jan-Hinnerk?).
> The source archive together with a Windows binary can be
> found on:
> http://www.siwawi.arubi.uni-kl.de/avr_projects/index.html#avrdudew3
>2

That's great news, at least for the people that use Windows ;-)

> Benefits:
> - faster in MS-Windows
> - no link to cygwin1.dll - easier redistribution like  in WINAVR
> - avoid version mismatch with other cygwin dlls in the system
> - no link to cygwin1.dll - smaller binary installer packages
> - possibly a larger user and tester base because of easier
> installation and faster programming/flashing compared to some other
>   programmers

Yes, we know the benefits. There has an effort to remove 
cygwin-dependencies for a while.

> The changes shouldn't break the unix build but I've only tested
> the comiler/link with cygwin-"standard" and cygwin-"non-dll".

I will test linux build when I have the time.

> Avrdude seems to be faster with "windows native" than with
> "windows cygwin-dll" this may break timings esp. with
> STK200-type interfaces - needs testing.

Could you please tell some numbers.

The time for the _read_ of flash along with the byte size of flash is 
sufficient for me to tell, if SPI-timing gets broken (in theory).

> So far I follow this list from the savannah list-archive please
> "cc" to the adress mentioned on my "avr-projects" page.

Done ;-)

Cheers
  Jan-Hinnerk