On Tue, 20 Mar 2001, Tom Kistner wrote:
> On Tue, Mar 20, 2001 at 11:42:21AM +0000, Yann Golanski (yann@???) wrote:
>
> > I'd rather the "begin X [...] end X" syntax. It makes a lot of sense. Of
>
> Hm but one of both is superflous, because a new "begin" already implies the
> "end" of the previous block.
Bits of syntax shouldn't be ruled out merely because they are
"superfluous". Redundency can serve a purpose in some cases. It can make
things easier for software to parse or for people to parse.
(Remember when
"<bar><foo>something</></>
was valid HTML?)
Now I haven't looked at the configure parsing code in exim, but everything
about the configure syntax suggests that it was designed for a quick and
easy finite-state parsing. The specific suggestions I've seen (with one
exception, ala bind) don't actually force a move from finite state
parsing, but introduce notation that looks like context-free. I really
suspect that adding those won't make things any faster for the software.
As for humans, how about if the default configuration file case things
like
############################################
# This is the END of the transport section #
############################################
###########################################
# This is the begining of .... #
###########################################
or whatever. We can put things in comments that will make things easier
for humans (and maybe vim) to parse, that take take away from the
simplicity of the parsing mechanism.
-j
--
Jeffrey Goldberg
I have recently moved, see
http://www.goldmark.org/jeff/contact.html
Relativism is the triumph of authority over truth, convention over justice
