Re: lynx-dev dev.16 patch 5 - "internal-links" text

[Klaus Weide]

On Fri, 10 Dec 1999, Leonid Pauzner wrote:

> 9-Dec-99 21:54 Klaus Weide wrote:
> > Maybe this creates some discussion...
> 
> > +  --enable-internal-links    (prevent defining DONT_TRACK_INTERNAL_LINKS)
> 
> Do I understand this right that with --enable-internal-links two kind of
> links are interpreted differently (and may cause resubmitions when
> href="foo.html#frag" activated)?

My attempts to explain this in the past have not been successful.
There are just too many details that folks don't follow, it seems.  So I ask
you to test what lynx does for yourself.  Either with --enable-internal-links
or without it, or, even better, comparing both.

I have put up a test page at
<http://enteract.com/%7Ekweide/test/addition.cgi> (I had it lying around
anyway.)

(There's a link to the script itself.  If doing much testing over the
network is a problem, you should be able to copy it and make it run
locally - even as a "lynxcgi:" if necessary.)

Submit the form, follow the links, explore.  Have a look at the source
if you like.  Pay attention to "where you are" (see INFO page).  The four
different types of links to compare are [4], [5], [6], [7] (please follow
me along in setting Keypad mode to "Links are numbered", so we can refer
to the links by the same numbers).

Does the observed behavior correspond to the expected behavior?

> Well, this is a tiny difference and I
> doubt page authors read RFC so carefully and this way.

Occasionally - not more than once or twice a year - a question has come
up on lynx-dev where --enable-internal-links helped for a real problem.
But this *is* an arcane topic - I doubt that page authors can rely on
any specific browser behavior in this area, even among more popular
browsers.  (But someone may want to test my examples with those and
share their observations.)

> Anyway, if without --enable-internal-links both kind of links are
> assumed equaly to href="#frag" and cause resubmittion, that should be
> fixed to work without resubmition.

Another simpler demo page:
      <http://enteract.com/%7Ekweide/test/test-postlink1.html>

Please check for yourself in which cases resubmission occurs when it
shouldn't (if it does).

(After you've checked out those test, we can proceed to explore even
more aracane situations - like what happens with the LIST page.)

   Klaus