> Let me just quote you:
> "This is mostly habitual -- this is what the GNU Coding Standard specifies
> for comments and which is enforced for GNU software which I have dealt a
> lot with."
> You didn't say it was common sense. You did say that it was what the
> GNU Coding Standard specified, and as a consequence, what you were used
> to. So please keep your "oh come on" for yourself, you pointed the
> discussion in this direction yourself.
Well, I take no habits that make no sense in the first place. And I have
gone into great lengths to explain and justify what drives me in this case
-- I got it from the GNU standard and got convinced it is good, so I got
to using it. I can write comments according to a different style, no
problem (as long as there is any defined style for a given case), but I
have to put some explicit effort into it.
Similarly, habitually I write code in the Linux indentation style because
I like it, but I can use your hated GNU style (or any other that follows
any recognisable rules) as well, except I have to put some brainpower into
> What matters is not "the pieces of code I am interested in", but the
> pieces of code _you_ are the master of, or not. As explained somewhere
> else in this thread, you are free to use whatever style you like (as
> long as it complies with Documentation/CodingStyle, that is) in new
> code you write and in code you maintain. For all the rest, you should
> stick to the surrounding style. This is common sense, as you'd say.
Well, sorry, but I could only sense the lack of style in this piece of
code, which is why I tried to apply some. You are free to disagree and as
you have undertaken maintenance of this area I am going to respect it.
> BTW, i2c-sibyte should be converted to a proper platform driver, so
> that only platforms with such a device instantiate it.
The whole of SiByte support should eventually get converted to implement
platform initialisation. I started some of this with changes to the
sb1250-mac.c Ethernet driver sometime in 2006, but no further progress has
been made since. I have other priorities higher on the list, but I have
not forgotten about it and will come back at some point unless someone
else does it first.
> Which legacy driver, "eeprom"? You should probably look into David
> Brownell's at24c driver:
> If it gets enough attention and testing, it could go upstream quickly.
I can see if I can find a couple of cycles to spare and give this piece
of code a shot with my SWARM. There is a pair of 128kB EEPROM chips
onboard (one as a bootstrap option and one to store configuration) and I
have two SDRAM modules installed providing another pair of a smaller size.