The testdata files in this directory start with a header, terminated by a blank line,
and then contain the test input. By convention, the files are named Kind.tN
where Kind describes the test (the license name for a single license), and
N is a sequence number.

The header starts with any number of comment lines beginning with #,
which are ignored.

After that optional line comes the expected Coverage result, in the form:

	90.5%
	BSD 98.9% 1234,4567 [URL]

That is, the first line is the Percent field and the subsequent lines are
the list of Match entries. Each Match contains the license Name, Percent,
Start, and End offsets. As a special case, the End offset can be written as "$"
if it extends to the end of the file. If IsURL is true, the line ends with the
literal field "URL". Otherwise that field is omitted.

After that stanza comes an optional additional expected Coverage result,
for use with the Scan function. It looks the same but starts with a line
that says "scan":

	scan
	100%
	BSD 100% 0,$

If the Scan stanza is omitted, the Scan function is expected to find nothing.

As a special case, if the Cover and Scan results are the same:

	100%
	BSD 100% 0,$
	scan
	100%
	BSD 100% 0,$

this can be shortened to

	cover+scan
	100%
	BSD 100% 0,$