Re: Some more debconf queries
Rob Browning wrote:
>
> Yeah, it's too bad all the config files don't just contain scheme/lisp
> forms. Then you could just have one universal tiny parser that's
> regular, very well defined, and gives people with parenphobia the
> willies :>
>
Well, using scheme/lisp expressions is daunting. The MacOS X people
are using XML for the task but IMHO, that's even crazier.
> (I've often wondered how much useful work has been wasted on mostly
> tangential attempts to define yet another config-file format. I feel
> the same way about the time spent trying to invent and worse yet,
> parse, the C++ language's syntax. Those front-end writers have my
> deepest sympathy :> )
A universal config file format would be nice, though I wouldn't think
C++. I'd do it like this
* class based language
* statements based on common concepts in config files (options, lists
, etc.)
* annotation based two-way translators for each config format
* multiple inheritance
* and a native grammar formalism to define translators
* perhaps some generics would be nice, but a bad generics makes
your language C++, so best to avoid that.
It would seem to me that this would avoid the brain-damage
of an XML approach. Clean, nice and cool. It'd be a programming
language on its own, somewhere between a declarative class based
language, and a documentation markup language.
And yes, writing a C++ front-end is a terrible job. I didn't do it,
but I don't think I want to write a C++ semantic analyzer in plain C.
Congratulations to the g++ team, though they couldn't implement the
export keyword :)
--
++++-+++-+++-++-++-++--+---+----+----- --- -- - -
+ Eray "eXa" Ozkural . . . . . .
+ CS, Bilkent University, Ankara ^ . o . .
| mail: erayo@cs.bilkent.edu.tr . ^ . .
Reply to: