Charlie Poole
2008-06-30 16:06:54 UTC
Hi All,
In short, I'm setting aside the distinction I was trying to make between
"static" and "dynamic" data sources and parameterized tests for 2.5. All
parameterized tests will be "static" - which means they will work like
the current examples we have: TestCase, RowTest, IterativeTest, etc.
Longer story...
I had the idea that "static" parameters would be determined when the test
is loaded, "dynamic" when the test is run. This turns out to be extremely
complicated to implement without (even more) changes to the underlying
design of NUnit.
More important, I can't find any good uses for "dynamic". Apparently,
other people couldn't find a reason to have two types of data either,
as my tests of mbunit, xunit and junit show.
Note that some test case providers may be "dynamic" in a different
sense: they may not return the same data every time they are called.
That's somewhat different from what I had been talking about and
NUnit will have to deal with it correctly.
So, my plan is to clean up the current model and publish another
alpha at the end of the week or on the weekend. This will include
parameterized fixtures and methods, including generics, and should
be close to feature complete for normal (non-theory) data-driven
tests.
Charlie
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
In short, I'm setting aside the distinction I was trying to make between
"static" and "dynamic" data sources and parameterized tests for 2.5. All
parameterized tests will be "static" - which means they will work like
the current examples we have: TestCase, RowTest, IterativeTest, etc.
Longer story...
I had the idea that "static" parameters would be determined when the test
is loaded, "dynamic" when the test is run. This turns out to be extremely
complicated to implement without (even more) changes to the underlying
design of NUnit.
More important, I can't find any good uses for "dynamic". Apparently,
other people couldn't find a reason to have two types of data either,
as my tests of mbunit, xunit and junit show.
Note that some test case providers may be "dynamic" in a different
sense: they may not return the same data every time they are called.
That's somewhat different from what I had been talking about and
NUnit will have to deal with it correctly.
So, my plan is to clean up the current model and publish another
alpha at the end of the week or on the weekend. This will include
parameterized fixtures and methods, including generics, and should
be close to feature complete for normal (non-theory) data-driven
tests.
Charlie
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php