Dave Syer wrote:
> I already took that approach with the patch I submitted for DOXIA-169. The
> test actually goes beyond the one for the APT version. Also, the APT one
> uses AptSink to accept output and make assertions about it. There is no
> ConfluenceSink so I used TextSink - which I think is a better approach
> anyway because I wouldn't want the test case to tightly couple the *Parser
> and *Sink.
Absolutely. I think in the apt case, the apt sink is only used to test
some special features, like macros, but in general, testing a parser
should not depend on a particular sink, and vice versa (see eg
DOXIA-100, DOXIA-101 for cases we have fixed already).
>
> I can't write the whole test suite in one go, and I'm not sure why that
> would help (all the tests would fail to start with), but we can do it bit by
> bit if you like, one feature at a time.
I mainly meant that it would help me to get familiar with the confluence
format. It would also be good to have a standard test document for each
parser, since then we can compare the parsing output with any arbitrary
test sink, eg the TextSink.
I don't see why all the tests would fail first, because the
AbstractParserTest by itself doesn't assert anything (apart some basic
well-formedness), it only parses the document so as long as parsing is
fine, you won't break anything.
-Lukas
>
>
> Lukas Theussl-3 wrote:
>
>>Hi Dave,
>>
>>I am a currently active doxia committer but I'm not really familiar with
>>the confluence module. If you submit some patches I will review them,
>>what would help me most as a start would be a complete confluence test
>>model test.confluence, to replace the current one in
>>src/test/resources/. It should produce the same text output as the
>>corresponding test files test.apt and test.xml in the apt and xdoc
>>modules.
>>
>>Cheers,
>>-Lukas
>>
>>Dave Syer wrote:
>>
>>>Is anyone actively involved in developing the Confluence module right
>>>now?
>>>Several issues have been raised this week (some by me), but no-one seems
>>>to
>>>be reviewing them, or working on them. Some are really trivial.
>>
>>
>