Installer memray avec Homebrew, apk, dnf, Nix, pacman

Liens du projet

• Page d'accueil
• Dépôt
• Docs amont
• Page du gestionnaire de paquets

Commandes et alias

• memray
• memray3.14

Historique du projet

Memray's README and documentation present it as a profiler for allocation tracing rather than sampling alone. It records function call stacks for allocations, can include native C/C++ frames, works with Python threads and native threads, and can be used both as a CLI and as a Python library.

The package entered public Python packaging with a 1.0.0 release on PyPI in April 2022. Its release stream continued through the 1.x line, with GitHub release notes and PyPI metadata showing active maintenance into 2026.

Historique d'adoption

Memray is distributed through Python packaging and system package managers. The upstream README links PyPI, PyPI download statistics, conda-forge, GitHub Actions for wheels, and documentation; the batch input lists apk, Homebrew, Fedora, Nix, and Arch packages. The README also points to `pytest-memray`, which brought Memray into test-suite workflows as well as ad hoc command-line profiling.

Modes d'utilisation

The common CLI workflow is `python3 -m memray run -o output.bin my_script.py`, followed by a reporter such as `python3 -m memray flamegraph output.bin`. Users can also run `memray run my_script.py`, use `memray run -m my_module`, enable native tracking with `--native`, attach to processes in supported scenarios, or use the API for finer-grained profiling.

Memray is limited to Linux and macOS. Because it includes a C extension and native tooling, source builds may require dependencies such as libdebuginfod, libunwind, and liblz4, while normal users are steered toward binary wheels from PyPI.

Pourquoi les passionnés de paquets s'y intéressent

Memray is notable because Python memory profiling often crosses the Python/native boundary. Packaging it means shipping compiled extensions, platform-specific wheels, and native library integration while still exposing a friendly CLI. It is also a useful package-index signal because it bridges Python developers, C-extension debugging, flamegraph tooling, and pytest integration.

Chronologie

• 2022-04-09: memray 1.0.0 uploaded to PyPI.
• 2022: README documents Linux and macOS support, PyPI installation, native allocation tracking, and multiple reporters.
• 2026-04-08: v1.19.3 GitHub release published.

Related projects

• Related projects include pytest-memray for test integration, PyPI and conda-forge as distribution channels, and native debugging/profiling dependencies such as libunwind, libdebuginfod, and liblz4.

Sources

• https://bloomberg.github.io/memray/
• https://github.com/bloomberg/memray#readme
• https://github.com/bloomberg/memray/releases.atom
• https://pypi.org/pypi/memray/json
• source_facts.package-manager-url