On 02/07/2019 11:16 pm, Hannah Pan wrote: > Add self-tests for the lzo-rle algorithm. > > Signed-off-by: Hannah Pan <[email protected]> > --- > crypto/testmgr.c | 10 ++++++ > crypto/testmgr.h | 80 ++++++++++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 90 insertions(+) Hi Hannah,
Thanks for adding these tests. It looks like these tests don't include any runs of zeros in the input. In this case, most of the new RLE-related code-paths in lzo-rle aren't invoked and it behaves exactly the same as standard lzo. To get better coverage, I would recommend adding some input data which includes some runs of at least 4 zeros (ideally, including some edge cases like all-zero, input starts / ends with a run of zeros, empty input, etc). Regards Dave
