[Mono-list] HTML Mockup for Class Lib Todo's

Nick Drochak ndrochak@gol.com
Tue, 12 Feb 2002 16:35:22 +0900

| Thinking about the todo generator - it must be possible to organise the
| directory structure and naming conventions of the tests so that we can
| automagically determine whether or not there are at least _some_
| tests for a
| given class. Could we go (or have we already gone) further and
| arrange that
| the tests deposit some kind of standardised pass/fail mark file so that we
| can figure out whether the implementation has passed the tests? Is this
| similar to what Duncan's up to?

I'm not really sure what Duncan is up to, but I know he was building
something using CorCompare as a starting point, and was looking into getting
at the [MonoTODO] attributes.

The convention for the tests is that each class's test is in a namespace
which is the same as the namespace for the class plus "MonoTests." tacked on
the front.  So for instance, the test class for Hashtable is "HashtableTest"
and it is in "MonoTests.System.Collections" namespace.

It should be pretty straight forward to look in the assembly that houses the
NUnit tests, e.g. corlib_test.dll and see what classes have tests.
Something that would be cool is to automagically figure out which
methods/properties have tests too.  Again, a naming convention would
probably be enough, but I'm not sure if we want to go that far.

| Also, who or what is the definitive source of module maintainer
| information?

The current source for maintainer info is the mono/status/class.xml.  We had
a discussion a month or two ago about putting attributes in the sources to
mark the maintainer and/or author(s).  It was decided that extra attributes
are a "bad thing" and we'll just use [MonoTODO].  The MonoTODO's are removed
as a piece is completed, so in the shipping assembly, there won't be any
"extra" attributes to slow things down, cause problems, etc.

In any case, putting the maintainer on the status page for a class is
probably not necessary right now, IMHO.

Let's both have a look at what JBarn has done in doctools and see where we
go from there.

In the meantime, do you have any comments about the XML structure I sent

Nick D.