Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
With all the new functionality, the complexity of the serd-pipe command-line
interface is starting to push the limits of available flags. So, instead of
grafting on further options to control a model, this commit adds a new tool,
serd-sort, which acts somewhat like a stripped-down serd-pipe that stores
statements in a model in memory.
This keeps the complexity (including the user-facing complexity) of any one
tool down, since other more focused tools can be used for streaming tasks in a
pipeline.
In other words, abandon Swissarmyknifeism, take a page from the Unix
philosophy, and try to expose the model functionality to the command-line in a
dedicated focused tool. The model implementation is tested by using this tool
to run a subset of the usual test suites, and a special suite to test statement
sorting.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This separates the command-line tool code from the library implementation.
|
|
This is slightly less excruciatingly slow on Windows, sometimes, maybe.
|
|
|
|
Everything covered by the old hairy runner script is now covered by more
focused test suites, so the "thru" pass can be eliminated without losing
significant coverage.
|
|
|
|
|
|
The earlier "test" was just hitting the code without actually checking the
output. This new suite is a set of pretty-printed documents which serd must
reproduce exactly to pass. This should make it easy to add cases in the
future, since each case is just a document, as it should look.
|