Beta testers for an XHTML syntax checker sought

[Leo Breebaart]

Hi all,

I have written an XHTML syntax checker, called 'Tidybot'. It is
built on top of the well-known "HTML Tidy" library.

I wrote it because I needed some very specific functionality I
couldn't easily find elsewhere. Now that it exists, I'd like to
make it available to the world, just in case others find it
useful, too.
What it does:

- Traverses one or more source directories on your hard disk
recursively, and runs all .html/.htm files it finds through
TidyLib, collecting all warnings and errors it encounters
and presenting them nicely in an XHTML report.

- You can specify files/directories to exclude from the checks,
you can specify warnings/errors to suppress in the generated
report, and you can specify 'key:value' options to pass
directly to the underlying Tidy engine. You can also tell the
generated report to use a different CSS stylesheet if you want
it to have your own look & feel.

- Comes in both a command-line version (for easy automated
scheduling) and a (functionally equivalent, but more
user-friendly) GUI version.

- Is cross-platform, running on both Unix/Linux and MS Windows
(and I daresay it will run on MacOS as well -- certainly the
command-line version should -- but I haven't been able to test
that). A one-file Installer application is available for
Windows. (On Unix, you will also need to install a number of
prerequisites.)
What it (by design) doesn't do:

- No conversion or editing of files -- it just checks files,
helping you to *keep* things tidy, rather than tidying them for
you.

- Doesn't get pages from a web server -- only static pages
available on the local file system are supported.
What I am looking for:

- People willing to give Tidybot 1.5b2 (the current beta version)
a run on their system, and then send me test reports and
feedback as detailed as they have the time and inclination for.

To clarify: Tidybot may have a rather limited functionality (when
compared to what Tidy is capable of) but it is not a quick hack,
and before I officially release it to the world I really want to
make sure it runs as flawlessly as possible. This is why all
feedback is welcome.

The Tidybot Home Page is:

<http://www.kronto.org/tidybot/>

and you can see daily updated report pages in action at:

<http://library.lspace.org/tidybot/>

Tidybot and its source code are released as free software under
the MIT License.

Many thanks in advance to anybody willing to help me out with
this.

--
Leo Breebaart <le*@kronto.org>

[Andy Dingley]

On 23 Jun 2005 12:49:56 GMT, Leo Breebaart <le*@kronto.org> wrote:

I have written an XHTML syntax checker, called 'Tidybot'. It is
built on top of the well-known "HTML Tidy" library.

Poor choice, IMHO. Tidy is built on HTML and isn't a good basis for an
XML tool. What's its behaviour depending on the content-type returned ?
Does it correctly handle XHTML _as_XML_ ?

[Leo Breebaart]

Andy Dingley <di*****@codesmiths.com> writes:

On 23 Jun 2005 12:49:56 GMT, Leo Breebaart <le*@kronto.org> wrote:

I have written an XHTML syntax checker, called 'Tidybot'. It is
built on top of the well-known "HTML Tidy" library.
Poor choice, IMHO.

Entirely possible -- as I said, I was hesitant about going public
with this utility, because I initially felt it was "just" a
wrapper around a tool that wasn't really created to be an XHTML
validator in the first place.

On the other hand, the TidyLib was *there*, I could actually use
it without too much hassle, and the result has certainly served
its purpose: I run our files through it, it flags things as
errors or warnings, I fix those, our XHTML files become neater
(and the real validators agree with that).

This is a net win for us any which way I look at it.

Tidy is built on HTML and isn't a good basis for an XML tool.
What's its behaviour depending on the content-type returned ?
I'm not sure I understand your question. Tidy (and Tidybot) work
on local files, not on HTML pages retrieved from a server, so
there is no content-type "returned" as I understand the phrase.

Also I would never claim that Tidybot was an XML tool -- I see it
more as a kind of 'lint' for XHTML files. Nothing more, nothing
less.

Does it correctly handle XHTML _as_XML_ ?

I think specifying the "input-xml:yes" to the underlying TidyLib
takes care of that, yes, but perhaps you can give me a specific
example of a situation that might not be handled correctly?

--
Leo Breebaart <le*@kronto.org>

[Andy Dingley]

On 23 Jun 2005 16:42:14 GMT, Leo Breebaart <le*@kronto.org> wrote:

Does it correctly handle XHTML _as_XML_ ?

I think specifying the "input-xml:yes" to the underlying TidyLib
takes care of that, yes, but perhaps you can give me a specific
example of a situation that might not be handled correctly?

Namespacing wasn't supported last time I looked. As this is one of the
few reasons for going XML over HTML, that's significant IMHO.

[Leo Breebaart]

Andy Dingley <di*****@codesmiths.com> writes:

On 23 Jun 2005 16:42:14 GMT, Leo Breebaart <le*@kronto.org> wrote:

Does it correctly handle XHTML _as_XML_ ?

I think specifying the "input-xml:yes" to the underlying
TidyLib takes care of that, yes, but perhaps you can give me a
specific example of a situation that might not be handled
correctly?

Namespacing wasn't supported last time I looked. As this is one
of the few reasons for going XML over HTML, that's significant
IMHO.

If TidyLib does not support XML namespaces, than obviously
Tidybot won't support it either.

I can't shake the feeling that you're finding fault with Tidybot
for it not being the utility you feel it ought to be, rather than
for any deficiency in what it actually is.

I've already described the scenario in which I find Tidybot
helpful to have around, and I am really not trying to make any
claims beyond that.

--
Leo Breebaart <le*@lspace.org>

[Andy Dingley]

On 25 Jun 2005 10:21:58 GMT, Leo Breebaart <le*@lspace.org> wrote:

I can't shake the feeling that you're finding fault with Tidybot
for it not being the utility you feel it ought to be,

Like an XHTML syntax checker ?

[Leo Breebaart]

Andy Dingley <di*****@codesmiths.com> writes:

On 25 Jun 2005 10:21:58 GMT, Leo Breebaart <le*@lspace.org>
wrote:

I can't shake the feeling that you're finding fault with
Tidybot for it not being the utility you feel it ought to be,

Like an XHTML syntax checker ?

Can you perhaps give me an example of an XHTML document that you
think contains one or more syntax errors which Tidy/Tidybot will
not catch?

I think we've already established to everyone's satisfaction that
I'm not the world's foremost authority on XHTML/XML matters. I am
perfectly willing to be educated (or be told where to go to
educate myself, even), but your snarky one-liners aren't exactly
the most constructive criticism I can think of.

--
Leo Breebaart <le*@lspace.org>

[Leif K-Brooks]

Leo Breebaart wrote:

Can you perhaps give me an example of an XHTML document that you
think contains one or more syntax errors which Tidy/Tidybot will
not catch?

No, but here's a good document which it will find bogus errors in:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<h:html xmlns:h="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<h:head>
<h:title>Example</h:title>
</h:head>
<h:body>
<h:h1>Example</h:h1>
<h:p>Some text.</h:p>
</h:body>
</h:html>

[Leo Breebaart]

Leif K-Brooks <eu*****@ecritters.biz> writes:

Leo Breebaart wrote:

Can you perhaps give me an example of an XHTML document that you
think contains one or more syntax errors which Tidy/Tidybot will
not catch?
No, but here's a good document which it will find bogus errors in:

Thanks, having an actual example to work with is very helpful.

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<h:html xmlns:h="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<h:head>
<h:title>Example</h:title>
</h:head>
<h:body>
<h:h1>Example</h:h1>
<h:p>Some text.</h:p>
</h:body>
</h:html>

Tidy (and therefore Tidybot) finds zero errors in this file if
you specify the "input-xml:true" flag.

Conversely, both the W3C and the WDG validators choke on this
example file just as badly as Tidy/Tidybot in normal mode does.

If I change the body in the above snippet to

<h:h1>Example
<h:p>Some text.</h:p>
</h:h1>

then Tidy in XML mode will still not complain, but in XHTML mode
the equivalent

<h1>Example
<p>Some text.</p>
</h1>

would yield a "Warning: missing </h1> before <p>".

Is that the issue you guys are getting at? XML mode only checking
for well-formedness, and not actually doing any validating?

--
Leo Breebaart <le*@lspace.org>

[Andy Dingley]

On 25 Jun 2005 16:24:33 GMT, Leo Breebaart <le*@lspace.org> wrote:

Can you perhaps give me an example of an XHTML document that you
think contains one or more syntax errors which Tidy/Tidybot will
not catch?

Not offhand. I'm not thinking of catching errors that are hard to catch,
so much as valid XHTML documents that are incorrectly flagged as invalid
- because they use XML features such as namespacing that aren't part of
HTML.

[Henri Sivonen]

In article <3i************@individual.net>,
Leo Breebaart <le*@lspace.org> wrote:

Conversely, both the W3C and the WDG validators choke on this
example file just as badly as Tidy/Tidybot in normal mode does.

Those validators use an SGML parser--not an XML parser. Furthermore,
they are DTD-validators and DTDs do not support namespaces, which makes
DTD-validation inappropriate for languages that are layered on top
Namespaces in XML. RELAX NG validation is better suited for namespaced
XML.

Online DTD-validation using an XML parser (Xerces from the popup):
http://valet.webthing.com/page/
Online RELAX NG validation:
http://hsivonen.iki.fi/validator/

--
Henri Sivonen
hs******@iki.fi
http://hsivonen.iki.fi/
Mozilla Web Author FAQ: http://mozilla.org/docs/web-developer/faq.html