Re: lynx-dev JavaScript again (was no subject)

[rjp]

In message <address@hidden>, 
           "Larry W. Virden" writes:
> 
> Anyone have a few man-years they would like to burn on such a project?

I have been doing.  I have just about working javascript: urls, partial
document.write support and that's about it.  Oh, and simple form validation
works.  

However, this is not pleasant.  The number of workarounds I've had to
implement to work around Lynx's internal structures (the way forms are
implemented internally baffles me, for example) and the broken Javascript
out there (javascript: urls are supposed to return the resultant location -
this is handled fine.  Unfortunately 90% of the Javascript out there doesn't
work like this, but works by setting document.location.  This is somewhat
handled, but only on a javascript: link.)

There are a whole host of problems.  A partial list:

Lynx has no internal document structure
Lynx has no event model
Lynx is user-centric; programmatically doing things like jumping to a link
 is, from what I can gather from the code, not supported and would need major
 hacking to get it to work.  The same applies to forms.  I haven't yet made
 enough sense of the forms code to allow programmatic submission of forms
 via Javascript possible yet.
Most of the Javascript code on the web doesn't work like it's supposed to
Most of the Javascript code on the web just isn't applicable to Lynx

If anyone wants my patch (using Netscape's libjs), I can clean it up and mail
it to you.  It isn't pretty, and there are still a lot of problems lurking in
the interactions.

It does work for simple, conforming Javascript though.

Oh, and it always asks for confirmation about following a javascript: link,
just in case.
-- 
rob partington % address@hidden % http://lynx.browser.org/