Skip to content

v0.1.2

Latest
Latest

Choose a tag to compare

View all tags
@ferinazumaDEV ferinazumaDEV released this 14 Sep 05:39
v0.1.2
3c99640
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Built from 3c99640c161839243df166ac22883269f72791fc by run 34810448566, never from a working tree.
Provenance is attested; verify with:

gh attestation verify <file> --repo ferinazumaDEV/politeclient

SHA-256

ef8da49eee3a557eb972355a89a7b48d6105be71303b4c9cb2e34144fafa3c8b  politeclient-0.1.2-py3-none-any.whl
ea14dcf4283918e7499a5ab04e69c10e0639a3e84f0dad0e11f81edd751365d9  politeclient-0.1.2.tar.gz

sbom.cyclonedx.json describes the package and its runtime dependencies as installed from this wheel.

How this release was assembled. The tag and this release were created by hand before the workflow ran, so the run's own "GitHub Release" step failed with "a release with the same tag name already exists" after PyPI had already accepted the upload. Per RELEASING.md the run was not repeated; the assets above are the run's dist artefact — the same bytes PyPI holds — and were attached on 2026-09-14 after checking SHA256SUMS, the attestations, the PyPI digests and the Integrity API provenance by hand.

Changelog

Deprecated

  • Python below 3.11 is compatibility, not support. The floor becomes 3.11 in 0.2.0 (3.10 reaches
    end of life in October 2026). Nothing changes within 0.1.x; CI keeps running on every version the
    package still declares.

Fixed

  • RetryPolicy.backoff_for no longer raises OverflowError past attempt ~1024. 2 ** attempt
    is an exact int in Python and stops fitting a float there, so a client configured with a large
    max_retries crashed on the attempt it should simply have slept max_backoff for. It now does.
  • parse_retry_after follows the delay-seconds grammar of RFC 9110 (ASCII digits only).
    int() also accepted +5, -5 and digits from other scripts (٣, １２); those now return
    None, so the computed backoff applies instead. A value with more digits than a float can hold
    raised OverflowError; it now means "never", which compute_delay clamps to max_backoff.
  • TokenBucket.acquire rejects zero, negative and NaN token counts with ValueError. A negative
    amount refilled the bucket past its capacity (acquire(-10) on a bucket of 2 left 12 tokens: a
    limiter you could bypass), zero was a silent no-op, and NaN never returned because it compares
    false with everything.
  • TokenBucket.acquire no longer spins when the refill lands a few ULPs short of the request.
    With a deterministic clock (what the clock/sleep parameters exist for) the sub-ULP sleep was
    absorbed and the loop never advanced; with the real clock it busy-waited until the monotonic clock
    ticked. A deficit within floating-point noise now counts as satisfied.
  • RateLimit, TokenBucket and RetryPolicy reject non-finite configuration at construction.
    rate=inf overflowed in capacity, rate=nan slipped past rate <= 0 and blew up in math.ceil,
    burst=inf built a bucket that never limits, and max_backoff=nan would have slept NaN. Rate-limit
    cases raise RateLimitConfigError; retry-policy cases raise ValueError.

All five were found by tests/test_adversarial.py, a new suite that states the invariants a caller
relies on (a bounded, finite backoff; a bucket whose tokens only go down when acquired; the RFC
grammar) and feeds the code the inputs a real client meets at the edges. Every test in it failed
against the previous code before the fix was written.

Archived at Software Heritage

Resolved on 2026-09-14 against the archive, after a Save Code Now visit that followed this tag.

Assets 6
Loading