Huon Wilson

In https://github.com/stellargraph/stellargraph, we're benchmarking/tracking memory-use of particular objects using a technique hacked on top of Python's built-in [`tracemalloc`](https://docs.python.org/3/library/tracemalloc.html) of creating a clock-like API for memory use: https://github.com/stellargraph/stellargraph/blob/15114f2258a93bc92f28882eea766639f3201ad1/tests/test_utils/alloc.py Example test using...

@AngryUbuntuNerd you may be interested in an expansion of https://github.com/ionelmc/pytest-benchmark/issues/28#issuecomment-561960809 above: since that comment was written, we've added benchmarking of peak memory too. New references: - code: https://github.com/stellargraph/stellargraph/blob/1e6120fcdbbedd3eb58e8fecc0eabc6999101ee6/tests/test_utils/alloc.py (see the...

Yeah, you're right that tracemalloc works really well for this: both solutions are using it. The only difference between them is the return value of the "timer" and how it...

The "modern" phrasing of `set_defaults` seems to be `__defaults__`, e.g. ```python __defaults__({ (pex_binary, pex_binaries): dict(complete_platforms=["path/to:platform"]) }) ``` (With or without `extend`.) I think this could potentially be customised on CI...

This was deprecated in 2.15 (#16415) and has been removed in 2.16.

That's the global `--process-cleanup`, while this issue seems to be specific to `run`'s old `--cleanup`, which was removed in https://github.com/pantsbuild/pants/pull/17647? (And, confusingly for this discussion, deprecated in the same PR...

Do you know of any preexisting data for 1 and 2 that cargo can reuse?

We're seeing the same thing with `existing: true` for a Cognito user pool trigger (https://www.serverless.com/framework/docs/providers/aws/events/cognito-user-pool#using-existing-pools): > Received response status [FAILED] from custom resource. Message returned: User: arn:aws:sts::ACCOUNT_ID:assumed-role/SERVICE-STAGE-IamRoleCustomResourcesLambdaExe-XYZ/SERVICE-STAGE-custom-resource-existing-cup is not authorized...

Brainstorming some potential ways to improve this: 1. use the "GET API" somehow (no idea how, yet) 2. reduce the number of requests by aggregating more blobs, particularly in the...

I've asked support and they said: - no way to increase rate limit - should ask at `actions/cache` for more info: https://github.com/actions/cache/issues/1280