W3C home > Mailing lists > Public > public-xml-id@w3.org > April 2005

Re: Test Suite catalog

From: Norman Walsh <Norman.Walsh@Sun.COM>
Date: Fri, 08 Apr 2005 13:57:07 -0400
To: Elliotte Rusty Harold <elharo@metalab.unc.edu>
Cc: public-xml-id@w3.org
Message-id: <873bu1i230.fsf@nwalsh.com>
/ Elliotte Rusty Harold <elharo@metalab.unc.edu> was heard to say:
| Possibly we could simplify this some since xml:id is much simpler than XSLT and
| the test suite is much smaller.

Yes a catalog was definitely needed. On the whole I'd slapped it all
together pretty quickly (and poorly).

I've added a catalog now: 


I also constructed an XSLT (1.0) stylesheet that can drive the tests
(if the xml:id implementation in question can inject the IDs before
running the processor, of course).


| The main difference from the OASIS format is the use of id elements to specify
| the expected output rather than output files. Each scenario would contain a list
| of zero or more id elements. The value of each such element is the value that
| should be reported by a conformant processor. For instance the above test case
| requires that the processor normalize the white space in the original attribute.
| Order matters here. Processors would be required to generate the exact values
| shown in the exact order.
| This makes writing a test driver fairly straight-forward. Load the catalog file.
| Iterate through all the tests. For each input file, list all the ID-type
| attributes in the order they appear and compare to the id elements in the
| scenario element.
| There are only eleven test cases currently, so it shouldn't be too hard to put
| this together. I'll try make a first pass on the problem this weekend, and then
| we can edit it or play with the markup as people think appropriate.

Thanks. I also added a "result" element, used by the XSLT harness.

| Here are some things I'm not sure of yet:
| 1. Should we assume people will install the files on their local file systems or
| should we just point to the test cases with absolute URLs to the W3C site?

Does it matter?

| 2. Do we list only xml:id values or all ID-type attribute values?

I think I've listed all ID values.

| 3. The test-case element should have a features attribute to identify optional
| features required to pass a particular test case that not all implementations
| may support. For instance, some of the test suites require schema support or XML
| 1.1 support, which not all processors will supply. Reporters should list these
| tests as skipped or "not applicable" rather than failed. What is the list of
| such features?

Only 'schema' and 'xml11' occur to me.

| 4. Would the working group like to define a standard output format for
| submitting test results? I think the XInclude working group has a nice format we
| could borrow if so.

I've coded up a format in my runtests stylesheet, please tell me what
you think.

You can see the results for my xmlidfilter at


| Comments? Questions? Concerns?

Thank you!

                                        Be seeing you,

Norman.Walsh@Sun.COM / XML Standards Architect / Sun Microsystems, Inc.
NOTICE: This email message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution is prohibited.
If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.

Received on Friday, 8 April 2005 17:57:20 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:53:49 UTC