🚀 go-pugleaf

On Sunday, in article <847032387snz@wash.demon.co.uk>
     john@wash.demon.co.uk "John Washington" wrote:

> I think that something like the crude ka9q/tw http fetch is
> useful for grabbing things like .zip files, or maybe a home
> page or an index or two, but it's no substitute for true
> web browsing.  To achieve full web browsing offline I think
> needs a proxy server on your own machine.  That can then
> roll the pellets for ka9q to fire, and have sufficient
> intelligence to fake the directory structure and multiple
> sites without needing to clutter your disk with an immense
> tree.

OK, I'll take some of that. Next year ?

> alt.gay.ka9q/dnos?  I'm not sure what you mean by 'recall' the
> http command.  Maybe that's why it's not in that version!
> Sorry, I've not had time to look at dnos yet.

Recall is my shorthand description of a command history facility. Just 
hit the up arrow (DNOS), or whatever, and get back earlier commands for 
re-execution or modification. It wouldn't be specific to http at all. I 
often want to redo diagnostic commands and I'd rather not retype 'em. 
That's why I described it as general.

> IS> Incidentally, ever successfully got any pages from Compuserve with 
> IS> KA9Q/Textwin's http ? If so, what was the format of the URL ? So far as
> 
> Perhaps I subconsciously excluded C/serve (hate hate hate).  :-)

I don't blame you, I dislike CIS too (ex-user with scars to prove it, 
but hate is going a bit far), but there are people (for some reason a 
lot of German folk) who still use it.

> Don't you just hate an outfit that allows its users to access
> the Internet, but encourages use of its own 'forums' and doesn't
> make them available to non-Compuserve users?

Nope, I'm grateful. Most of the forum contents are crud and there is 
quite enough of that on the internet without CIS adding their private 
sea of grief to it. Don't feel excluded, think of it as a lucky escape.

> Seriously though, I hope to find out what's wrong, and/or make
> the fetcher work thru www.cache-demon.

I'd like the fetch sorted, but I'm less sure about the cache. I've only 
done a few experiments, but TW/http sans cache is always quicker than 
browser (be it Netscape, DOSLynx or Minuet) with cache. I can believe 
that popular pages *might* be quicker, but I obviously travel the less 
trodden byways of the web. For a new fetch, the cache is a hindrance 
rather than a help and it may even be so all the time if it acts as a 
bottleneck (which I believe it does). A hint to at least make cache use 
optional if it's added. For what it's worth, turning off the cache does 
speed up the browser fetches somewhat, but they never approach the speed 
of TW/http. For sheer speed, scripted http is in a different league and 
multiple http sessions even more so. Even though speed isn't really the 
browser's thing, take a bow.

> IS> I can see, what works fine for Netscape gets a raspberry when used with 
> IS> http. The KA9Q, or rather the JNOS, www server also waves two fingers at 
> IS> me when using http :(
> 
> Comes with source, doesn't it?  Simple matter of grep two.fingers *.c
> and then you know which line to change?

Yup. Easy, huh ?  See, there's no excuse at all :) 

Ian.
-- 

  |  Ian Smith            | "The Moving Finger writes;
  |  ian@isis.demon.co.uk |  and, having writ, Moves on."