[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: lynx-dev why reload to save? and other questions
From: |
David Woolley |
Subject: |
Re: lynx-dev why reload to save? and other questions |
Date: |
Thu, 2 Jul 1998 08:15:32 +0100 (BST) |
> 1/ When getting to an interesting page, I want to save it in its original
> format. I must reload it using the 'd' binding. I find this most
> inconvenient because the very reason I use lynx is to save costly
> french telecom time.
The code isn't structured to do this and adding generalised caching tends
to bloat it when there are quite a few dedicated caching servers that could
be run on the same machine (free for Unix, probably expensive for Windows).
This is an FAQ and there is a strong lobby to have the code added by
some DOS/Windows users. See the archives for details.
NB make sure you are already using your ISPs caching proxy.
>
> 2/ lynx does not keep open FTP connexions, so it connects every time
> It access a page on a given server. This is a bummer because many
> servers are long to accept a connexion but snappy to answer once
> connected.
I wondered about that. Some versions of the WWW library certainly have code
for doing this, although it can be anti-social to maintain a connection
inefinitely, so there would need to be logic to drop connections after
a short time or if the site seemed no longer to be current. I wonder if
it is the need to purge the cached connections that has prevented its
inclusion.
>
> 3/ Is there a way to start lynx in download mode instead
> with display mode. For example I got an URL that points
> to a binary that I want to download?
-source
>
> 3/ When started from another tool with a wrong URL, it
> exits. I would expect a more gracious fail. Indeed, very often
> everything but the last component are ok.
> say, if I try to access http://www.foo.org/~hacker/rpms/nerdtool-1.0
> and that nerdtool-1.0 has been removed, I would
> like to have http://www.foo.org/~hacker/rpms displayed to
> see whatever new version of nerdtool is available.
More and more on commercial sites this strategy will simply produce a
"not authorised" response. You would probably have to qualify it with
the presence of ~. It will probably break a lot of uses of (3) above and
other command line uses.