Re: [Htmlvalidator-help] html validator problem

[Marc G. <mg...@sk...>]

Hi Branko,

First to be sure, the extension implements 2 algorithm in 0.8x.
Maybe read this:=20
http://users.skynet.be/mgueury/mozilla/user_guide.html#Tidy_vs_SGML_parse=
r_

The reason why you get no error in the W3c validator or in the extension=20
with the SGML parser
is because these 2 programs are the same and based on DTDs.

A DTDs is not specifying the possible attributes for arguments. It means=20
that they are not able to check if or not

<object type=3D"application/x-shockwave-flash" data=3D"/swf/tekst.swf?tek=
st=3D??%20????" width=3D"340" height=3D"30">=09

is right or wrong. Now the other algorithm, Tidy does it.

And in this case, unhappily the spec are strangely done but Tidy is=20
right. (I think)
I do not follow this spec closely. But I am nearly sure that there is no=20
final specification
for URL containing non ascii character. The ""good"" way is to encode=20
the characters.
The reason is that there is no charset specification of the URL sent in=20
a HTTP request.
So, there is also no way to know how to decode the URL. The big problem=20
of this is that the way to
encode non-ascii character depends of the HTTP server .... For UTF8

/swf/tekst.swf?tekst=3D??%20????

should be something like this:

/swf/tekst.swf?tekst=3D%D0%8A%D1%83%20%D0%88%D0%BE%D1%80%D0%BA

For more info, go here: http://www.ietf.org/rfc/rfc2396.txt and look for=20
ASCII.

Marc


Branko Jevti=E6 wrote:
> Hi!
> First of all, thank you for this great extension! it's priceless!
> Only problem that I've encountered (in both 0.7xxx and beta 0.8=20
> versions) is that if I have cyrillic text in my url for SWF file for=20
> example html validator says it's an malformed URI reference, but if I=20
> check it on w3c site it says it's valid. Here is one site for example:=20
> http://kosjeric-ds.dev.si21.info/
>
> best regards!
> Bane