On Tue, Nov 30, 2010 at 2:18 PM, Dave Pawson
On Tue, 30 Nov 2010 12:04:00 +0000
Philip Fearon <pgfearo(_at_)googlemail(_dot_)com> wrote:
With the above, there are possibly 2 areas where it would be useful to
have some kind of community standard:
(1) the XML for expressing the test input environment
? Shared with any test environment?
Yes, to be usable on any test environment with the capabilities I
mention. XPath should be in there somewhere to allow for parameters
that aren't just xs:string types and to provide the context node. Its
possible that a subset of XProc (mentioned by @vojtech) would be a
What's being tested (versions ....)
When its being tested, who by?
Test program (version... parts ...)
software in use (xslt engine, java vsn, OS)
This data should be in the summary report, most of it extracted
automatically, I don't think it all needs to be specified in the input
environment. The output summary also confirms the input parameters
etc. that were ultimately used by the test environment, which should
of course match those specified in the input environment XML.
Expected output file.
(2) the XML summarising the output
Common to any testing?
No, just to XSLT testing/
reference to test definition(s)
Test count run
Tests passed, failed, not run.
Templates [matched/named] not used.
Input elements not matched (??? If applicable)
All included in the output summary. Unmatched templates appear as errors
XML comparison of expected/actual from each template.... Possible?
Not sure. How to encapsulate depth? XMLdiff definitely needed.
Agree XMLdiff would be invaluable for regression testing, but this
isn't the only kind of test.
I've previously appealed for views on a common format for the XML
output summary but this wasn't met with enthusiasm at the time.
Define what's wanted first Phil? Defining XML wrappers ins't magic.
W3C has a markup IIRC?
The requirement is that all output data not normally accessible to
XSLT-based frameworks is collated into a single XML resource, that can
So, I'm afraid that, as yet, my company provides no framework that you
can run from the command line, this is all managed from a GUI. Its
kind of like an IDE, but with multi-column file lists intended for
reviewing the large set of files that comprise a test suite, instead
of the more conventional tabbed pages, which require you to first open
a file to view it. - Tabbed pages are used, but only to let you switch
between associated input, output and the XSLT.
I'm almost convinced you need more than XSLT to test XSLT properly.
Well, I guess that depends on how much data the XSLT processor gives you.
I can't see the GUI coming until testing has been defined?
Perhaps not for unit testing where you just produce a text-only test
report. But there are end to end tests or acceptance tests, where
HTML, Office documents etc are being produced where a GUI becomes
pretty essential. Not everything can or should be automated and this
is where a GUI fits in. A command line tool could I guess produce an
interactive HTML report, but if you're going to end up using a GUI
(browser) to review test results you might as well start with one.
I think you'll probably need help from Mike to use Saxon for testing.
The Saxon API and documentation is pretty extensive, but yes, there
I'm presuming you can't tweak the tested XSLT to test the XSLT... if
you see what I mean. That chappie and his cat effect?
I think there's a lot of potential in what you say, there's already a
capability to use XSLT on the test XSLT to produce the sets of XPath
used by the GUI for presenting and navigating through specific input
or output. One of the many benefits of using a declarative language.
No, not easy.
XSLT XSL-FO FAQ.
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>