[xmlroff] What is the useful default 'testing' setup? [Was: Nearly ready for xmlroff 0.5.2]

[Tony Graham]

On Thu, Aug 09 2007 08:12:31 +0100, elvstone at gmail.com wrote:
> 2007/8/8, Tony Graham <Tony.Graham at menteithconsulting.com>:
...
>> I would appreciate it if someone would verify that the current code
>> works for them.
>
> The current code works for me on an up-to-date Kubuntu Feisty system.
> These are my test results (no DocBook tests):
> http://franzkafe.se/xmlroff-testing/

What is the useful default setup for the 'testing' module?

I was already thinking yesterday about making a separate
'testsuites.xml' file and separate test results file for the DocBook
tests, and today it seems that when people run the tests to check
xmlroff (and thank you all who have done that), they run just the
xmlroff test suite and not the DocBook testdocs.

Is it less confusing if the default tests and test results are just for
the xmlroff testsuite?

(There's already multiple test results files, and there will soon be
more since it's now easier to reuse test configurations, so having a
minimal default isn't such a problem.)

Also, when you run the tests for the first time, the tests that are
listed as having differences are just the tests that produced output.

Using the '--disable-reference' option will disable the comparison to a
reference and hence disable the listing of differences, but would it be
more useful if, for example, it was possible to 'preload' the reference
version with the sample outputs provided in the testsuite?

Regards,


Tony.