Skip to main content.
home | support | download

Back to List Archive

Re: HTTP method config

From: Bill Moseley <moseley(at)not-real.hank.org>
Date: Wed Jan 16 2002 - 18:58:42 GMT
At 08:27 AM 01/16/02 -0800, David L Norris wrote:
>Is there a good list of which config file directives won't work via HTTP
>method?

You mean the (depreciated yet? ;) -S http method?

Well, the http://swish-e.org/2.2/docs/SWISH-CONFIG.html file does have
three sections:

  Directives for the File Access method only 
  Directives for the HTTP Access Method Only 
  Directives for the prog Access Method Only 

And I've tried to note which can be used where, which I think about it.

As I've mentioned I'd like to depreciate the -S switch and switch over to
just using -i or (poorly named) IndexDir.

       file://  
       http:// 
       prog://

(Although, I'm not sure that anyone would really use that -- hence it's way
down on the todo list...)

But, one tiny problem is the some of the config settings are only checked
in the individual modules (http.c, extprog.c, fs.c).  That means if you use
"Delay" for -S fs, you get an error now, which is nice, but would need to
be changed to allow URL-type of sources specified.

>Specifically those related to handling which parser to use for
>a particular file extension.  I don't see a clear answer in the docs. 
>Does prog method get around this problem somehow?

Which problem do you mean?  

Yes, the code for IndexContents (to decide which parser to use) is after
the source input modules (fs, http, prog).  The prog method does allow
passing in a (currently undocumented?) header to override (or rather "set")
the parser to use.

   Document-Type: HTML2

This is useful when spidering because you can set the parser based on the
http Content-Type: header instead of the file extension.

ReplaceRules works for all source input types, but FileRules is only for -S
fs, except for FileRules title, which is check while parsing.

If you are using -S prog, then the smart thing is to only send the data you
want indexed.




-- 
Bill Moseley
mailto:moseley@hank.org
Received on Wed Jan 16 10:58:52 2002