Branko Čibej <brane_at_xbc.nu> writes:
> Yes. Changing the coding style in a trivial way that doesn't actually
> improve readability for most people isn't reasonable.
Okay. So this is a different point than I thought you were making.
You don't object to enforcing stylistic consistency in general; you
object to *this particular convention*, because you think it doesn't
improve readability for most people -- and that all the strcmp choices
are about equally okay.
For all I know, you might be right; it's hard to take good surveys about
this stuff. The way currently documented in hacking.html is more
readable for me, but I could live with any of the other ways too, it's
not a big deal.
> We did this once with the space-before-paren thing, though that had
> somewhat stronger arguments on its side because most people (who
> weren't used to RMS's creative destructivism) were a bit put off by
> the code.
Yeah, 90% of people don't like GNU-style, beats me why :-). Anyway,
there's a technical advantage to the new way: it's easier to search for
function calls now, because they look less like conditional guards. So
I think that was a good change, even though I voted against it.
> Changing the spots on the bikeshed from mauve to lavender won't make a
> noticeable difference except that the other half of developers are
> going to be unhappy. :(
>
> I'd suggest to leave well enough alone, and let people write
> !strcmp() or '! strcmp()' or '0 == strcmp()' or 'strcmp() == 0' or
> !strcmp() != 0' or whatever suits them as long as the meaning is
> obvious and the code is correct.
Anyone object if I take it out of hacking.html? I don't care that
strongly about it, and some people care strongly that we not care
strongly...
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-06-05 18:20:26 CEST