ENH: Add numbers to the three itkPyBufferMemoryLeakTest print messages

[N-Dekker]

Add numbers (1, 2, 3) to the print messages in itkPyBufferMemoryLeakTest,
to ease analysis of the output from a test failure.

[N-Dekker]

@dzenanz Could the itkPyBufferMemoryLeakTest failures possibly be introduced by pull request #2554 "BUG: add support for long long pixel types in PyBuffer", merged by cd36c64 ?

[dzenanz]

Incremental build didn't work for me, I needed to manually rebuild something to avoid a clean build.

While it is possible that PR did it, I see no reason why it would do that. This same test used to be flaky on Mac before, if I remember correctly.

[N-Dekker]

While it is possible that PR did it, I see no reason why it would do that.

I don't know either. Would it matter that these types from itkPyBuffer and itkImageDuplicator have become out of sync, with PR #2554 ?

ITK/Modules/Bridge/NumPy/wrapping/itkPyBuffer.wrap

Line 5 in cd36c64

UNIQUE(types "${WRAP_ITK_SCALAR};UC;D;US;UI;UL;ULL;SC;SS;SI;SL;SLL;F")

ITK/Modules/Core/Common/wrapping/itkImageDuplicator.wrap

Lines 11 to 13 in cd36c64

	# Making sure that all types wrapped in PyBuffer are also wrapped for ImageDuplicator
	# as this filter is used in the Python NumPy bridge.
	UNIQUE(types "${WRAP_ITK_SCALAR};UC;D;US;UI;UL;SC;SS;SI;SL;F")

[dzenanz]

Even if it does not fix the pyBuffer memory leak, ImageDuplicator should be wrapped with LL types. -> Done in #2584.