I'll take a look at this tomorrow, but wanted to respond to some email
questions/comments:
} * Forgot to mention previously: I've changed the date format from
} 18-Dec-2006 to 2006-Dec-18---any objections? (I believe Philip said
} he was OK with it, as long as there was no date ambiguity.)
I object. 2006-12-18 would be ok, but 2006-Dec-18 is just weird. I
quite like 18-Dec-2006 myself, but if you really want to go ISO 8601,
you should just do that, and have numbers everywhere.
} * Went out on a limb here: I've replaced the various
} --enable-newline-is-foo options with a single --enable-newline=NL,
} where NL can be "lf", "cr", "crlf", or "any". I can add back in the
} -is-foo options if need be; do we need to keep the options the same?
I'll let Philip speak to this. Personally, I'd just as soon this be a
separate change: I'd like this patch to not be really user-visible at
all.
} * There's a simplification I'd like to make: When you give
} AC_ARG_ENABLE(foo, ...), Autoconf automatically defines for you a shell
} variable $enable_foo that contains the value of the option ("yes", "no",
} or whatever was specified with --enable-foo=blah). Same deal with
} AC_ARG_WITH and $with_foo.
}
} What do y'all say we ditch the ac_pcre_* variables, and use
} enable_foo/with_foo in their stead? We could then eliminate the third
} argument from most of the AC_ARG_ENABLE/WITH calls, as it would be
} redundant.
This seems like a nice simplification. Any cost to making this
change?
But again, we have something that works now, so any objection to going
with what we have now, and then making what you suggest part of a
future change? It wouldn't be user-visible, would it?
