wrote:
> <form action='loginScript.cgi' method="post" name="loginForm">
> <input name="userid" />
^[1]
> <input name="password" />
^[1]
> <a href="javascript:document.loginForm.submit()" >
`----.,---'`-------.,-------'`---.,--' ^
[2] [3] [4] [5]
> <img src="someImage.gif" >
^[6]
> </a>
> </form>
[1] IE does not support XHTML, so there is no use in serving it to it.
If it was HTML, it was equivalent to <input ...>> which is just
as bad.
And should not the password `input' element be of type="password"?
[2] The pseudo-protocol `javascript
:' does not belong into `href'
attribute values, at least not without a graceful-degrading alternative.
Here, users without enabled client-side script support will not be able
to submit the form. A simple
<input type="image" src="someImage.gif" alt="whatever">
would have sufficed for all UAs newer than Netscape 4.x (199

.
[3] Form/HTMLFormElement objects should be referred to using the `forms'
collection of HTMLDocument objects: document.forms['loginForm'].
[4] Methods should be feature-tested on run-time before they are called:
<URL:http://pointedears.de/scripts/test/whatami#inference>
[5] The HTML 4.01 Specification recommends against whitespace
after the `a' element's start tag and before its end tag:
<URL:http://www.w3.org/TR/html4/struct/text.html#h-9.1>,
"In order to avoid problems with SGML line break rules and
inconsistencies among extant implementations [...]"
[6] The `alt' attribute is missing for this to be Valid HTML.
<URL:http://validator.w3.org/>
> [...]
> Any ideas on how to use a server side language (PERL preferably) to
> submit via this javascript dependent form?
Yes, ask in a Perl newsgroup (comp.lang.perl.*). This does not have
anything to do with J(ava)Script/ECMAScript because what is done there is
nothing but emulating an unnamed submit button; incompetently, if I may
add.
> the only think I can think of is to have my server-side language (PERL)
> to create and run a temporary javascript enabled form/webpage and capture
> the results... but this would be highly inelegant..
Yes, it would, and users would wonder what would happen. However, just for
the sake of completeness, for my RFC Search bookmarklet I happen to use
something similar:
data:text/html;charset=ISO-8859-1
%3C!DOCTYPE%20html%20PUBLIC%20%22-//W3C//DTD%20HTML%204.01//EN%22%20%22http://www.w3.org/TR/html4/strict.dtd%22%3E%3Chtml%3E%3Chead%3E%3Ctitle%3ERFC %20Search%3C/title%3E%3C/head%3E%3Cbody%3E%3Cform%20action=%22http://www.rfc-editor.org/cgi-bin/rfcsearch.pl%22%20method=%22post%22%3E%3Cinput%20t ype=%22hidden%22%20name=%22searchwords%22%20value= %22%s%22%3E%3Cinput%20type=%22hidden%22%20name=%22 filefmt%22%20value=%22txt%22%3E%3Cinput%20type=%22 hidden%22%20name=%22num%22%20value=%2225%22%3E%3C/form%3E%3Cscript%20type=%22text/javascript%22%3Efunction%20isMethodType(s)%20%7B%2 0return%2
(s%20==%20%22function%22%20%7C%7C%20s%20==%20%22ob ject%22)
%20%7D%20var%20o%20=%20document;%20if%20(o%20&&%20 (o%20=%20o.forms)%20&&%2
(o%20=%20o%5B0%5D)%20&&%20isMethodType(typeof%20o. submit))%20o.submit()
%3C/script%3E%3C/body%3E%3C/html%3E
(This is initiated from the client side, of course.)
PointedEars