Repository navigation
C accelerator for Base32 character encoding #146192
Description
Activity
The Base 32 codec is the only one which is implemented in pure Python. This not only makes it slower and increases memory consumption (the same issue as was with the Base 85 codec, see #101178), but it makes difficult to add new features, code for which can be shared between other codecs implemented in C. So, rewriting it in C will help with maintenance.
- added a commit that references this issue
on Mar 20, 2026 - added a commit that references this issue
on Mar 23, 2026
Proposal:
The
base64module contains implementations of encoders/decoders for several character encodings.Among these are the base32 and base32hex encodings mentioned in RFC 4648.
These are currently implemented in pure Python. Implementing them in C will greatly improve performance (~100x) and memory usage.
I have written a draft PR that accomplishes this, and will be submitting it shortly.
Has this already been discussed elsewhere?
No response given
Links to previous discussion of this feature:
Suggested by @serhiy-storchaka in #102753.
Linked PRs