Filprofiler

Features

- Report memory stats when out-of-memory event is detected. ([114](https://github.com/pythonspeed/filprofiler/issues/114))


Bugfixes

- Correctly handle bad data from cgroups about memory limits, fixing erroneous out-of-memory caused by Docker. ([113](https://github.com/pythonspeed/filprofiler/issues/113))

Features

- Out-of-memory detection should work in many more cases than before. ([96](https://github.com/pythonspeed/filprofiler/issues/96))

Features

- Fil now supports Python 3.9. ([83](https://github.com/pythonspeed/filprofiler/issues/83))

Bugfixes

- Fil no longer uses a vast amount of memory to generate the SVG report. ([102](https://github.com/pythonspeed/filprofiler/issues/102))

Bugfixes

- Fixed bug that would cause crashes when thread-local storage destructors allocated or freed memory. Thanks to winash12 for reporting the issue. ([99](https://github.com/pythonspeed/filprofiler/issues/99))

Features

- Allocations in C threads are now considered allocations by the Python code that launched the thread, to help give some sense of where they came from. ([72](https://github.com/pythonspeed/filprofiler/issues/72))
- It's now possible to run Fil by doing `python -m filprofiler` in addition to running it as `fil-profile`. ([82](https://github.com/pythonspeed/filprofiler/issues/82))
- Small performance improvements reducing overhead of malloc()/free() tracking. ([88](https://github.com/pythonspeed/filprofiler/issues/88) and [#95](https://github.com/pythonspeed/filprofiler/issues/95))


Bugfixes

- When running in Jupyter, NumPy/BLOSC/etc. thread pools are only limited to one thread when actually running a Fil profile. This means Fil's Jupyter kernel is even closer to running the way a normal Python 3 kernel would. ([72](https://github.com/pythonspeed/filprofiler/issues/72))