Re: substitution-groups.testSet and DTDs

On 21 Jun 2010, at 03:40 , Michael Kay wrote:

> I noticed a significant delay while running the new substitution- 
> groups.testSet through my test driver. On analysis, this is caused  
> by the reference to  
> contained in ./ancillary/xhtml-for-annotation.dtd.
> The problem of fetching DTDs from the W3C site is of course well  
> known, and the circumventions well-documented. But people  
> implementing XSD test drivers have got better things to do than  
> writing catalog files to reduce W3C web traffic. I wonder if it's  
> really necessary for the testSet file to use a DTD, and if so,  
> whether it's really necessary for it to be remote?

Mea culpa.

My usual XML editor is Emacs in psgml mode, which requires a
DTD rather than any other form of schema, so I have constructed
DTDs for the mixtures of XSTS + HTML and XSD + HTML I use in
the tests I construct; otherwise I suppose my test cases would
have even more syntax errors than they already have semantic

The right solution, I think, is to use catalogs to avoid the
remote fetch (or at least a local HTTP cache which remembers
from one moment to the next that you just read this same resource
a few seconds ago and it has an expiration date still some way
into the future).  But of course, this cuts both ways:  it means
you can avoid the remote fetch without any action on my part,
by adding appropriate entries in your local catalogs and using
them.  But I can equally well use catalogs to tell Emacs that
every document I edit with root element 'testSet' or 'xsd:schema'
should use this or that particular DTD, and this would make life
easier for other users of the test suite.

OK, done and checked in.  None of the documents in wgMeta or
wgData should now have an overt doctype declaration.  Please let
me know if I overlooked any; thanks.


* C. M. Sperberg-McQueen, Black Mesa Technologies LLC

Received on Monday, 21 June 2010 14:08:41 UTC