*** Thomas Dickey ([EMAIL PROTECTED]) wrote in the Lynx list on Jan...:
:) On Wed, Jan 03, 2001 at 03:27:23PM -0500, Philip Webb wrote:
:) > the big, big problem is JS, which is taking over like creeping ivy:
:) > at least, couldn't we get Lynx to extract URLs from JS & show them as links?
:)
:) I don't think so. What would be doable (I think) would be to interpret a
:) subset of JS (the parts that generate html on the fly).
I have been looking at the problem since some time ago, you can make easy
subsitution when you have urls of the form
'http://www.cnn.com/' + tail
where tail is some variable, I think this is easy to support but it's too
restricted to call it a contribution to anything.
The problem is that Javascript is a programming language, you need to
support expressions like
if (this condition) then
variable = window.open('http://this.address'+info+'.html', etc, etc)
else
variable = window.open('http://other.address', etc, etc)
and so on. A job like this requires that you have a way to get values of
JS variables from the document itself, for example, when doing forms, so
you need to consider that this means that a lot of new functions must be
added to the library. You would need to keep the document in memory (or on
file), is this already done? I believe that you can get relatively good
success if you get a good way to parse documents that contain expressions
like above and ignore what really does not apply to lynx.
--
Eduardo
http://www.math.washington.edu/~chappa/pine/
; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to [EMAIL PROTECTED]
