W3C Validator Not Working Offline

[Vigil]

Hi,

I have downloaded the source code and installed everything on
http://validator.w3.org/source/. I have the validator 'working' offline
with Apache2. I have uploaded a valid XHTML 1 Strict file to the online
validator and it passes with flying colours. When I pass it to my offline
version, it says, "This page is not Valid XHTML 1.0 Strict!". This happens
with both the Address and Local File methods.

I tried the command line '/usr/local/bin/onsgmls -c
/usr/local/share/sgml-lib/sp-1.3/pubtext/xml.soc -E0 -R -wvalid
-wnon-sgml-char-ref -wno-duplicate -wxml' and passed it the test file, but
no matter how I call it, it always gives the error:

cannot find "/usr/local/share/sgml-lib/sp-1.3/pubtext/xml.soc"; tried

I have tried pointing to umpteen possible locations for xml.soc, all that
are valid. I even copy and paste each time to ls to make sure my eyes are
not going funny, but no matter what I do, it always says it cannot find
this.

I have done a google on the web and on newsgroups and have had some close
hits but none that are spot on or have otherwise helped. Is this a known
bug with the latest version of anything? Has anyone got this working? What
should I do? :-/

--

..

[Nick Kew]

In article <pa****************************@privacy.net>, one of infinite monkeys
at the keyboard of Vigil <[e-mail]> wrote:

I tried the command line '/usr/local/bin/onsgmls -c
/usr/local/share/sgml-lib/sp-1.3/pubtext/xml.soc -E0 -R -wvalid
-wnon-sgml-char-ref -wno-duplicate -wxml' and passed it the test file, but
no matter how I call it, it always gives the error:

cannot find "/usr/local/share/sgml-lib/sp-1.3/pubtext/xml.soc"; tried
Firstly (not the cause of your problem), sp 1.3 is seriously old.
Neither that nor even up-to-date OpenSP has a very extensive
entity catalogue. That means your "offline" validator may have to
fetch DTDs/etc over the 'net far more often than you'd expect.

The validator bundles a better [sg|x]ml lib, or you can run the automatic
updating utility available at http://valet.webthing.com/catalogue/
I have tried pointing to umpteen possible locations for xml.soc, all that
are valid.
That message means the environment variable SGML_SEARCH_PATH doesn't
include /usr/local/share/sgml-lib/sp-1.3/pubtext , so it won't look there.

That's caused by the -R flag. Without it you'll get "normal" behaviour,
which opens a potential security hole when someone tries things like
<!ENTITY passwd "/etc/passwd">
I have done a google on the web and on newsgroups and have had some close
hits but none that are spot on or have otherwise helped.

You must've googled the wrong places ....

Actually there's probably not much public record: Liam (who discovered
the security issue and wrote the patch) let us know by private email,
and we kept it there until we'd all closed it - which was much less
simple before the OpenSP-1.5 release.

Is this not documented in the package you downloaded?

BTW, if you really want offline use, you might be interested in
validator-lite, <URL:http://www.webthing.com/software/validator-lite/>
Oh, and (more shameless self-promotion), since you have Apache 2 you
can also use the altogether-more-powerful mod_validator from
http://apache.webthing.com/
--
Nick Kew

In urgent need of paying work - see http://www.webthing.com/~nick/cv.html

[Eric Cartman]

Vigil <[e-mail]> wrote in message news:<pa****************************@privacy.net>. ..

Hi,

I have downloaded the source code and installed everything on
http://validator.w3.org/source/. I have the validator 'working' offline
with Apache2. I have uploaded a valid XHTML 1 Strict file to the online
validator and it passes with flying colours. When I pass it to my offline
version, it says, "This page is not Valid XHTML 1.0 Strict!". This happens
with both the Address and Local File methods.

I had problems at first. Like you I downloaded all the packages,
installed them and tried to run and it didnt work! What I found was
that I already had various packages (especially onsgmls) installed on
my system as default. The new packages I had dowloaded installed to
default locations different to the originals and so I was still using
the outdated packages on my system and not those I had downloaded!

Try doing a 'which onsgmls' and see what you are using, you may find
its not the one you have just installed! Just amend paths in the
scripts to suit or amend your LIBRARY paths.

HTH

EC
..

[Vigil]

Thanks for your help. I have settled on Validator-lite, which is much
cleaner :-)

Might you submit it to Freshmeat so that people can see when new updates
happen?

PS. I tried the automatic catalogue updating utility but every one gave an
unknown HTTP error.

--

..