I'm working on a project where I'm doing various file manipulations. I'm not familiar with the idioms are common or good for this situation, so I'd appreciate advice and suggestions.
Question 1: A number of checks apply to a set of files. So I'm wondering whether to have something like:
def test_SomeTest(...): ... self.AssertAllFilesExist(fileList)
or
def test_SomeTest(...): ... [ self.AssertFileExists(f) for f in fileList ]
Personally I avoid using list comprehensions with side-effects; if I don't actually want the list I write it out:
for f in fileList:
self.AssertFileExists(f)
Question 2: A more complicated example: My object defines its own conditions that can be queried, e.g.
def test_SomeTest(...): myObject.DoSomething() self._assert(myObject.isFileTypeQ(someFile))
but the parameter someFile is often an expression that should be encapsulated into the assertion. This gives me two possibilities: I can define a stand-alone query that takes myObject as a parameter:
def AssertIsTypeQ(theObjectBeingTested, rawFileInfo): someComputedFile = DoStuff(rawFileInfo) return theObjectBeingTested.isFileTypeQ(someComputedFile)
I would write this method to actually do the assertion: def AssertIsTypeQ(theObjectBeingTested, rawFileInfo): someComputedFile = DoStuff(rawFileInfo) self.assert_(theObjectBeingTested.isFileTypeQ(someComputedFile))
and then call it from within various tests as:
self.AssertIsFileTypeQ(myObject, someFile)I think of it as extending the standard assertions with domain-specific ones.
I often write methods called checkSomething() that do a bit of work and check the result. So I might actually call the above method something like checkDoStuffIsFileTypeQ().
So, the convention I use is - write assertSomething() primitives that just check a condition - write checkSomething() methods that do some work and check the result - build the actual tests using the above
Kent -- http://mail.python.org/mailman/listinfo/python-list
