Re: lynx-dev cernrules.txt

[Klaus Weide]

On Sun, 7 Feb 1999, Michael Warner wrote:
> On or about 05 Feb, 1999, Klaus Weide <address@hidden> wrote:
> 
> > # in case you find it useful.  Note that this may change or go away in
> > # future releases of Lynx; if you find it useful, consider describing your
> > # use of it in a message to <address@hidden>.
>
> Well, OK.  It makes a quick and easy replacement for
> "--enable-gzip-help".  I stuck in:
> 
> RULE:Map file://localhost/PATH/lynx_help/*.html \
>       file://localhost/PATH/lynx_help/*.html.gz
> 
> and similarly for lynx_help/keystrokes/*.html, ran 'gzip -r
> lynx_help', and Bob's your uncle (whatever THAT means).
> 
> Another caveat - it seems the specified part of the URL
> must match exactly.  My lynx.cfg had the help-file location
> as file://localhost/users/wsunix/warner/[...], and using
> file://localhost/~/[...] in the RULE didn't seem to work.
> Apparently the URL is matched as-is, without tilde expansion?
> The comment didn't specify either way, I don't think, just said
> the tilde form was OK.  I haven't tried using the tilde in both
> places yet.
> 
> Now I have, and, at least for me on OSF1 V4.0 878 alpha, neither
> /~/ nor /~warner/ seems to work for a file:// URL, even if it's
> identical in both the HELPFILE and RULE entries.

When the URL reaches the point where the rules are applied, the
tilde is already changed to the directory path.  And the URLs
in the rules file are then just compared with that as strings
(plus the '*' matching), there isn't any special interpretation
of '~' at that point.  I probably should have written that
down.

> My uneducated opinion is that it's worth leaving in, if it
> doesn't have too much impact on lynx's size, and I dimly
> recall reading that that's the case.

It's overall not much code, and the size-conscious can exclude
that with -DNO_RULES.  I admit I haven't yet found a use for
it that I *really* need and couldn't do otherwise.  But I go
through a personal proxy, where I can do that kind of stuff
already.

Does anyone here know of a way to access URLs for Usenet messages
by message-id on one of the archiving sites?  (dejanews etc.)
I would find it really nice to have a rule to automatically
access archived copies of news articles that way.  Something
like
  news:*          -> http://wherever.net/find?msg=*
for
  news:address@hidden -> http://wherever.net/address@hidden


   Klaus