| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2025-07-27 | Remove unnecessary comments and logic from get_deepest_frame | bd | |
| 2025-07-27 | Fix bug where duplicate frames would get added (current tasks) | bd | |
| 2025-07-27 | Use a more typical strategy for profiling current tasks | bd | |
| 2025-07-27 | minor modifications to deepest_frame debug printing | bd | |
| 2025-07-27 | Partial solution to tracing all lines (coroutines calling methods) | bd | |
| 2025-07-26 | Profile the current task as well | bd | |
| 2025-07-22 | Add imperfect logic to include async generators in profiled frames | bd | |
| 2025-07-22 | Mark internal aergia methods as private | bd | |
| 2025-07-22 | Clean up some comments in Aergia | bd | |
| 2025-07-21 | Update comments reflecting on asyncio.gather and task groups | bd | |
| 2025-07-20 | Aergia.get_deepest_traceable_frame refactor, tests | bd | |
| Generators and list comprehension gives correct but less ideal results, matching yappi. Can this be improved in the future? | |||
| 2025-07-19 | Relicensed as Apache2.0 | bd | |
| 2025-07-19 | Fix a bug where inactive loops would be indexed (read desc) | bd | |
| Cleaned up Aergia.walk_back_until_loop to only get running event loops. Previously, this manifested as a bug. The asyncio.Process class would fork off a thread, and the event loop would be found in that thread's local variables, even though that thread was not executing that event loop. This would result in get_event_loops containing duplicate values, and affected event loops would be assigned duplicate time | |||
| 2025-07-19 | Reformat repository to allow for new unit tests | bd | |
 |
index : aergia.git | |
| A sampling profiler for suspended coroutines, loosely based off of SCALENE | bd |
| summaryrefslogtreecommitdiff |