Hi,
Ok, I was seeing this as 3 step process:
1) Split the Configuration class into Configuration and HtConfiguration:
I've almost finsihed that. HtConfiguration would go int htcommon
and Configuration would stay in htlib
2) Think about a good solution for specifying a context for
a configuration...actually I solved the same problem in another
program I wrote ... but that uses STL intensively :-(
3) Find a way to integrate the URL/SERVER configuration contexts
within this framework
After step 1 I will be able to commit my compression code which is the
priority for me. And afterwards, I'll get back to the Configuration
code. It's necessary that htdig/htword's parameters be server-specific
but it's also necesary for htdig/htword to work outside the htdig
context.
> What do you mean?
> I thought about something like this:
>
> <something1: bla-bla>
> xxx:yyy
> zzz:qqq
> <something2: bla2-bla2>
> www:eee
> rrr:ttt
> </something2>
> </something1>
>
> And interface:
> a=conf["something1=bla-bla:xxx"] //a=yyy
> a=conf["something1=bla-bla|something2=bla2-bla2:www"] //a=eee
>
Yup, that's the sort of thing I was thinking of. Loic suggests that using
XML for the configuration file would be the best solution.
But ... I thought we were in feature freeze? ;-)
Cheers,
Marcel.
PS.
> If we are talking about config, I thought about another 2 things:
> 1. stronger syntaxis.
> I mean avoid stupid typo: "max_dic_size" instead "max_doc_size"
> 2. Usage in conf.[yl] String class instead char*.
There's no reason not to use String instead of char *, why should it
be a problem?
----------------------
Marcel Bosc
ECILA
100 av. du Gal Leclerc
93500 Pantin - France
Tel: 33 1 56 96 10 85
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED]
You will receive a message to confirm this.
