Sebastian Rahtz wrote:
> I have corrected one error in Roma (the spurious copy of the schema in
> the middle of the form), but cannot understand why IE6 fails to work.
Opera fails too (acting even more strangely than IE), as does Moz on Knoppix
(according to Chuck Muller).
> If anyone out there can suggest what I am doing wrong, I'd
> be very happy. This is really a pretty simple Perl CGI.pm
> cgi script.
I don't think the problem is with the CGI part; it's with the [x]HTML that
it's dishing out. I'd warrant that when it can pass
it will work fine on all browsers.
If you decide to give it a try, you'll probably have to make several trips,
copying the viewed source of the CGI page and making changes as necessary to
the script itself until it spits out clean markup. (I just now used the
"enter your HTML directly" link there to paste the copied source, and I
noticed there was no closing </form> -- that or any number of the other
errors -- singly or in combination -- can cause big problems.)
In deference to the mode and mindset established by Mosaic/Netscape in the
early 90's, browsers do try and grok as much tag soup as they can, but
there's only so much AI they can apply to each and every case. I used to
get bitten badly myself, until I read
which finally convinced me. :)
At the very least, it eliminates one possible cause of problems.
PS. In addition to any number of your familiar command-line validators
(nsgmls, xmllint, ...), there's also the W3C's online one, but for [x]HTML I
prefer the WDG's interface and find the error messages more helpful. Plus
there seems to be no way to copy and paste Viewed Source at the W3C site.