wdg-html-validator and w3c-dtd-xhtml

[Onkar Shinde]

On Tue, Jun 9, 2009 at 6:39 PM, Andrea Gasparini<gaspa at yattaweb.it> wrote:
> Hi,
> while reporting to its Debian maintainer the merge of ikiwiki, I found the
> real problem of ikiwiki really was a strange behavior of wdg-html-validator
> ( note that 'man validate' tells it is "an Offline validator" )
> While in debian without a working net it works flawlessy:
> $ ping www.w3.org
> ping: unknown host www.w3.org
> $validate html/index.html
> $
>
> In Ubuntu it's broken:
>  $ validate html/index.html
>  *** Errors validating html/index.html: ***
> Error at line 2, character 53:  host "www.w3.org" not found
> Error at line 2, character 53:  DTD did not contain element declaration for
>    document type name
> ...[lot of errors here] ...
>
> That's the prologue. The fact it's that w3c-dtd-xhtml has a revision
> (here's the related patch[1]) different from debian.
>
> And in fact, while searching for a dtd in Ubuntu fails
> $ xmlcatalog /etc/sgml/w3c-dtd-xhtml.cat "-//W3C//DTD XHTML 1.0 Strict//EN"
> No entry for PUBLIC -//W3C//DTD XHTML 1.0 Strict//EN
> ...instaling the package from debian, it works:
> $ xmlcatalog /etc/sgml/w3c-dtd-xhtml.cat "-//W3C//DTD XHTML 1.0 Strict//EN"
> /usr/share/xml/xhtml/schema/dtd/1.0/xhtml1-strict.dtd
>
> ( useless to say that both validate and ikiwiki work now correctly)
>
> Finally, the question: has anyone the clue to check why ubuntu had this
> delta, and could say if it can be dropped, or not?
> And if not: how can we correct this?
> If yes: it'll lead to a sync of a couple of packages.
>
> [1]http://patches.ubuntu.com/w/w3c-dtd-xhtml/w3c-dtd-
> xhtml_1.1-5ubuntu1.patch

The package was changed in hardy for following reasons.
1. I was trying to solve FTBFS in lucene2 where a unit test was trying
access DTD on internet. Since internet access is not allowed form
buildd chroot the test was failing which eventually caused FTBFS.
2. I decided to use DTD from w3c-dtd-xhtml package and patch lucene2
source to refer to local path. [1]
3. While trying to use this solution I found out that the DTD from
w3c-dtd-xhtml was not working as the entity sets it referred were
stored in different path than where DTD was stored. So I had to patch
w3c-dtd-xhtml. [2]

There is already a bug reported in Debian which talks about wrong
nistallation paths for entity sets. But the developers weren't
convinced that it is a problem. [3] Hence we are still carrying this
diff.

I don't think the patch should cause the problem. Try replacing
'/etc/sgml/catalog' with '/etc/xml/catalog'. I am in no way expert in
how the xmlcatalog command is supposed to work and I may be wrong
about '/etc/sgml/catalog'.

I hope this info is sufficient.

[1] https://bugs.edge.launchpad.net/ubuntu/+source/lucene2/+bug/189602
[2] https://bugs.edge.launchpad.net/ubuntu/+source/w3c-dtd-xhtml/+bug/183164
[3] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=423087


Cheers,
Onkar