James Henry Dodd wrote:
Hi All,
How about a peer-to-peer scheme where, still using the I*C example, there
could be an I*C directory where those who so desire could contribute their
code. Perhaps the only requirement would be a "this works for me" statement.
public folders are a bit difficult to set up, with all the security
implications. but maybe a Wiki would be something?
a Wiki would allow all users to edit the pages add comments, update
code. (entire files can also be attached, but that 'hides' them, you
have to first download it to view, altough otherwise you have to
copy&paste...)
a Wiki would have the advantage that everyone can contribute and its
indexed by google, once its found by its robots.
i'm not yet sure whre i can set up a wiki, if that's even desired by the
people on this list. there should be a possibility on sf.net, but the
webspace is not too large...
[...]
The idea of standard bit-twidling macros (and, why not, other peripherals
too) is also good. It contributes to the "mspgcc way".
altough its mot so easy.
if you define port and bit separately, they are only loosely coupled,
but its more flexible. it also easy to use with (inline) asm.
macros/inline functions that set or clear a bit are a bit less error
prone but then, you also loose flexibility and code may be a bit larger
because of that.
i often prefer the first style. also because i like to see the bit
set/clear code in the source and it has advantages if more that one bit
has to be set/cleared. and i dont like too many macros, esp. if they
look like functions but are not. they can be misleading when you have to
squeze aut the last byte out of the code, so that it fits on the smaller
MCU.
chris
