While reading the hashlib code, I found some issues in the error branches where the EVP_MD_ctx is not freed upon failure or when we call py_digest_name with a NULL EVP_MD *.
@gpshead Should I consider this as a security issue? (some places might be a security issue since we are leaking some EVP_MD context objects but others are just leaking un-initialized contexts).
Activity
gh-127667: fix memory leaks in `hashlib` (#127668)
pythongh-127667: fix memory leaks in `hashlib` (python#127668)
pythongh-127667: fix memory leaks in `hashlib` (python#127668)
[3.12] gh-127667: fix memory leaks in `hashlib` (GH-127668) (#130783)
[3.13] gh-127667: fix memory leaks in `hashlib` (GH-127668) (#130784)
gh-127667: refactor and improve `_hashopenssl.c` error branches (#131145
pythongh-127667: refactor and improve `_hashopenssl.c` error branches (…
pythongh-127667: refactor and improve `_hashopenssl.c` error branches (…
pythongh-127667: refactor and improve `_hashopenssl.c` error branches (…
[3.12] gh-127667: refactor and improve `_hashopenssl.c` error branches (
[3.13] gh-127667: refactor and improve `_hashopenssl.c` error branches (