rcorder: add some primitive tests

[kevans]

Summary

Test basic PROVIDES/REQUIRE functionality, as well as various keyword
operations. Also document, via a test, the current behavior when a
REQUIRE is missing.

[brooks]

I've not worked through every case, but over all looks good.

[ngie]

Correct me if I'm wrong, but doesn't this rely on the fts order returned by the filesystem? If so, it seems like it would be better to explicitly list which files should be included by appending the files to a variable, then passing that variable to rcorder.

[kevans]

In D50559#1155255, @ngie wrote:

Correct me if I'm wrong, but doesn't this rely on the fts order returned by the filesystem? If so, it seems like it would be better to explicitly list which files should be included by appending the files to a variable, then passing that variable to rcorder.

The shell glob is expected to reliably sort by the current collation; I went with all lowercase filenames to avoid any surprises for weird locales, but we could probably export LANG=C to remove all doubt

[ngie]

In D50559#1155260, @kevans wrote:

In D50559#1155255, @ngie wrote:

Correct me if I'm wrong, but doesn't this rely on the fts order returned by the filesystem? If so, it seems like it would be better to explicitly list which files should be included by appending the files to a variable, then passing that variable to rcorder.

The shell glob is expected to reliably sort by the current collation; I went with all lowercase filenames to avoid any surprises for weird locales, but we could probably export LANG=C to remove all doubt

👍