GitLab

for simple counters. Switch default plots to use report objects.

Updata tests: stdio - cherry pick some value to verify, report - reflect switch from r.report to r.data.

Update jupyter notebooks to reflect refactorings (moving plots to experimental, seperating namespaces)

Prepare seperation of raw data and summary namespaces, add provenance and consistency structures for report algebra.

Refactoring to move aggregations to experimental for the time being. Convieniently reenable by calling darshan.enable_experimental() after import darshan.

Have cffi backend cache module index and name_records along with log handle as they are often used internally.

reported by Francois Trahay

various tests (pytest, coverage, lint/flake8) and make targets for easy
releases.

for MPIIO, PNETCDF, and HDF5 modules.

the struct alignment is important here, as we are going to read this
struct once to get the fixed values and a second time to fill in the
extra histogram buckets.

this requires a hacked up version of mdhim until there's a real api for
getting this information.