Pauli Virtanen

This however fails (run in `astropy-benchmarks` directory): ``` from benchmarks.io_ascii.table import TableSuite from asv.extern import asizeof f = TableSuite() f.setup() obj = f.mem_table_outputter() asizeof.asizeof(obj) ```

I guess the main point here would be to find a fix for `asizeof`, but it's bundled from some other project, and I've so far never actually read its code.

I'd prefer to split to subgroups only if there are more than 5 (or something) benchmarks. The number of benchmarks per class apparently varies between projects, and e.g. the scipy...

I think both use cases can be motivated. Can probably add a new switches for that (`--steps-after-skip` or some such).

Why not use `try: expect:` in your `setup` instead? I'd like to avoid adding new API methods unless there's very clear need for them.

I see. That the setup/teardown routines are not paired is a design problem. If the fixture-type behavior is wanted, how about instead a `finalize_teardown` hook, which runs last also when...

Can you run the precise command line it shows i.e. the last `/Users/dstansby/github/pfsspy/.asv/env/063e9b2d74f371a5aa0c00791c9eb7cb/bin/python /Users/dstansby/miniconda3/lib/python3.7/site-packages/asv/benchmark.py discover /Users/dstansby/github/pfsspy/benchmarks /var/folders/6y/j8hmrrxs1vz3bmtvzvgln5zh0000gn/T/tmp2g1coiwj/result.json` Does that get stuck? Does it get stuck when the command is run...

No reply, closing.

Can you run benchmark suites from other projects / the minimal quickstart one? Do they hang? If yes, try commenting out parts from your benchmark suite to find out what...

asv run "`git tag -l | awk '{ print $1 "^!" }'`"is the incantation, but as seen above, not so easy to recall...