Age | Commit message (Collapse) | Author |
|
This way there is only one place where we read chunks and where the
read/write wrapper is used.
This also allows to remove LoadedChunk as it is not used anymore.
|
|
to try to fix tests on windows
|
|
|
|
|
|
|
|
|
|
added some file path management function for prfixes and trailing slash
|
|
|
|
|
|
|
|
|
|
so that tests can be cached and to make sure it is independant of
the rest of the code
also move tests in testdata as this folder is ignored by go test by default
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
this way a partial chunk may have less superfeatures than
a complete one
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Still missing a real test...
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
If a chunk matched, directly skip to the next chunk instead of
calculating every single byte by byte checksum in it.
|
|
|
|
|
|
|
|
|
|
- also add a test for store and load
- precise channel direction in function args
|
|
first working chunking logic
|