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