Re: lynx-dev Scripting LYNX to follow links

[Jeff Haferman]

In lynx-dev, you wrote:
>On Mon, 25 Oct 1999, Jeff Haferman wrote:
>>I am trying to automate web queries, and thought that lynx may be the
>>vehicle to use.  I want to be able to move about an HTML document and
>>follow links of my choosing, per a script, and print pages to files.
>>In particular, I need to go to
>>
>>    http://in-133.infospace.com/info/rtq/login.htm
>>
>>and then interactively command
>>
>>    <tab>....<tab>loginname<tab>password<enter>
>>       (20 tabs)
>>    <tab>...<tab>symbolname<enter>
>..
>>Is lynx capable of what I am attempting?  (I am using lynx v2.8rel2 for
>>UNIX).  If so, how do I do it?  Otherwise, can you suggest a more
>>appropriate tool?  I'm fairly new at this, so I might need a few
>>specifics.
>
>It was discussed here recently -- the unix tool "expect" seems to be
>able to do what you want.  (I haven't read the manpage enough to figure out
>how to get it to let me just INTERACT and _then_ save what I did into a file..
>that's what I mostly want it to do -- just save my keystrokes and then play
>them back..)
>

Okay... what you need to use is the "autoexpect" script that comes with expect. 
 
I've been banging my head all day with expect and lynx, pretty much getting
nowhere.

So, any experts out there?  The page that I'm working with changes
everytime I load it due to advertisements, etc.... so, I have to
get expect to read tab past this unknown number of links in order
to reach the input box I'm looking for...