Sandboxes / #1082

perf: Add fast path for parsing simple versions (digits and dots only)

pypa/packaging#1082 · merged 2026-02-11 · base 59f183fe90 → head ee13ba2c1c · ran on Sandboxes

REVIEW

A valid input family runs 2.91× the instructions of the old code.

Known regressions in this PR

✓oversized-int-exception · fixed by #1155 · verified

Counterexamples · 1 class: 1 intended

Exception behaviour changed

known regression: oversized-int-exception

inputVersion("100000000000000000000000000000000000000000…000000.0.1" (4,305 chars))

baseraises ValueError

headraises InvalidVersion

Packaging.version.InvalidVersion is a subclass of builtins.ValueError.

Intendedvotes: 3 intended

The PR adds a fast‑path for versions that contain only digits and dots. In that path it explicitly catches a `ValueError` from `int()` conversion and re‑raises it as `packaging.version.InvalidVersion`. This makes the exception type more specific and matches the library’s own error hierarchy, so the new behavior is deliberate.

if _SIMPLE_VERSION_INDICATORS.issuperset(version):

self._release = tuple(map(int, version.split('.')))

Found by: hostile mutation · also caught by: Mutations of the PR's own test strings

ran onToken Factory Sandboxes · image cdd80561-4ef0-4ca1-b645-b13306e326c6

No regression test: adjudicated intended: no test pins the old behaviour.

Performance · Version

Worst family with mostly valid inputs: 2.91× the instructions of base (Leading zeros in first segment).

R4.1
R4
R4.2
R4.3

Head/base cachegrind instructions per call, by input size (log-log). Each line is one input family Nemotron proposed; points above 1× are where the "faster" code does more work.

Family	Inputs	Ratio range	Slope base → head	Verdict
`R4.1`	Trailing zerossuspected slowdown · expand:R4	1.05× – 2.36×	0.60 → 0.71	slower
`R4`	Leading zeros in first segmentsuspected slowdown · archive	1.05× – 2.91×	0.56 → 0.70	slower
`R4.2`	Mixed leading/trailingsuspected slowdown · expand:R4	1.04× – 2.66×	0.65 → 0.75	slower
`R4.3`	Increasing common prefixsuspected slowdown · expand:R4	1.03× – 2.59×	0.64 → 0.74	slower

Receipt

Target	Backend	Image	Run	Audit	Adjudication
`packaging.version:Version`	Token Factory Sandboxes	`cdd80561-4ef0-4ca1-b645-b13306e326c6`	`bt-1082-4a405a`	$0.17	$0.018

Nemotron cost on Nebius Token Factory. Backend: where the untrusted code ran. Token Factory Sandboxes (a microVM branched from a kept image, cited by its UUID). Every image is content-addressed (repo at each commit + harness), so every card re-runs from the image and seed 0. Raw data: audit, adjudication.