I'm trying to imagine what would be a good way to test for the successful processing (reading or writing) of a Windows Media Format file.

The WMReader interface has to parse the file correctly.  I believe it would be impossible to test for this, without having a natural WMF (.wma, .wmv, or .asf) file to read.  There are two ways of getting one: include one in the test suite, or create one as part of the test.  The first option has certain disadvantages, such as A) the file size and B) inability to tweak the contents, or easily make a variety of similar files, should the need arise.

Also, the WMWriter has to write the file correctly.  I see only two ways of testing this: compare the output byte-for-byte with a file generated on a Windows PC, or attempt to read the generated file with a known good wmf implementation.  The first way would assume the output is always the same byte-for-byte, which it may not be.  For the second, we would need a conforming implementation to use.

It seems to me, libav would be quite useful, as it can already decode and encode many if not most of the WMF formats.  Do you think it's a valid assumption to make, that libav's codecs are a conforming implementation of the WMF formats?

Jefferson Carpenter