CPython drops its experimental JIT compiler from the main branch after a withdrawn proposal
Python's Steering Council found the plans for maintenance and performance too vague; work can continue outside the main code.
CPython's experimental just-in-time compiler, the reference implementation of Python, is leaving the main branch of the code. The Steering Council had paused work on it in June and demanded an accepted PEP covering maintenance, compatibility and measurable goals, with a six-month deadline, LWN reports. The PEP 836 proposal was withdrawn on 19 October; co-author Savannah Ostrowski said the team lacks the time to carry it forward.
According to InfoWorld, the council considered the proposal too optimistic about performance: the measured gains were small and mainly benefited programs with long, tight loops. Commitments on support and on working with debugging and profiling tools were also missing. LWN writes that the council feared the JIT would get in the way of other projects.
The JIT appeared in Python 3.13 as an experiment; the current situation stays in place for 3.15. Those who want a faster Python still have the adaptive interpreter from 3.11, and outside CPython, PyPy and Meta's CinderX offer a JIT, InfoWorld notes. The work can continue elsewhere, but must meet the same strict requirements to return.