geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bruce Snyder <>
Subject Re: [Testing] Testing
Date Tue, 12 Aug 2003 16:00:52 GMT
This one time, at band camp, Alex Blewitt said:

AB>I think you'll end up with a situation where a spec-test developer will 
AB>just put in the section-number, since that's essentially what 
AB>highlights it. Also, if the section is prefixed with a document name 
AB>(J2EE_1.4-4.2.7) then you can pretty much get rid of everything else.
AB>I agree that having everything self-contained is good, but I can't see 
AB>developers caring about typing in chapter-name and section-name each 
AB>time they need a test. It should be relatively easy to put the sections 
AB>in a properties file listing each of them and just do a lookup (using 
AB>J2EE_1.4-4.2.7 as a key in a properties file)
AB>  :
AB>J2EE_1.4-4.2.7=J2EE 1.4: Transaction Management (Transactional JDBC 
AB>Technology Support)
AB>  :
AB>J2EE_1.4-6.17=J2EE 1.4: Java Management Extensions
AB>  :

Another very good point which has made me rethink the idea completely. How
about a cross reference which contains all the metadata? 

AB>> I agree that there needs to be a measuring stick of sorts for 
AB>> determining
AB>> how much further we need to go. But the idea is that the inclusion of 
AB>> the
AB>> section-number adds the ability to view a report and easily see gaps in
AB>> the section numbers. I'll be honest that I don't care to volunteer for
AB>> going through each spec and making a list of the tests. Maybe someone
AB>> else is interested in performing this task.
AB>I know the feeling :-)
AB>Anyway, instead of having to type chapter-title-section-name each time, 
AB>I've munged the J2EE 1.4 ToC and prepared the attached properties file. 
AB>Should save developers having to type in the lot each time.
AB>The Doclet can then weave the section number in and lookup entries in 
AB>the keyset. Should also be able to do sufficient sorting so that it can 
AB>order things by chapter and title anyway, since they're encoded in the 
AB>key itself.

Right idea, but what about the rest of the metadata? I don't think that
parsing a TOC and a metadata file is the right approach. I vote for
housing it all together in an XML file ala JMSCTS (
as suggested by Tim Anderson in the next message in this thread.

perl -e 'print unpack("u30","<0G)U8V4\@4VYY9&5R\"F9E<G)E=\$\!F<FEI+F-O;0\`\`");'

View raw message