Python 3.9.1 Lands as Developers Get Comfortable With the New PEG Parser

Python 3.9.1 went out yesterday, December 7th, and if you blinked you probably missed it. That’s the point of a bugfix release — no new syntax, no new modules, just the usual mid-cycle cleanup after 3.9.0 shipped back on October 5th. But it’s a good excuse to check in on how that October release is actually landing with people writing code day to day.

The headline change in 3.9 was swapping out the old LL(1) parser for a new PEG-based one. That’s not something most of us think about while writing code, but it’s the kind of foundational change that unblocks future language features without the old grammar getting in the way. CPython had been carrying that LL(1) parser for essentially its entire existence, so this was overdue plumbing work rather than a flashy feature.

The stuff people actually notice day to day:

• Dictionary merge operators. | and |= for merging dicts. d1 | d2 instead of reaching for {**d1, **d2} or a manual loop. It’s a small thing but it reads a lot cleaner once you get used to it.
• Built-in generic types for hints. You can now write list[int] or dict[str, float] directly instead of importing List and Dict from typing. If you’ve been typing your codebase, this cuts out a chunk of import boilerplate.
• New string methods. removeprefix() and removesuffix() finally give you a proper way to strip a known prefix or suffix without the str.startswith() plus slicing dance everyone’s been hand-rolling for years.
• zoneinfo. A standard library module for IANA timezone support, meaning you don’t automatically need pytz just to know what time it is in another region. Given how much pain timezone handling has caused in basically every language, this is a welcome bit of stdlib maturity.

None of this is revolutionary on its own, but it’s the kind of release that quietly makes day-to-day code a bit shorter and a bit more honest about intent. The generic type hints in particular feel like Python catching up to how people were already trying to use its type system — the typing module workarounds always felt like a tax you paid for wanting better tooling.

If you’re still on 3.8 or earlier, there’s no urgency here — 3.9.1 is a stability release, not a reason to rush an upgrade mid-project. But if you were waiting for the dust to settle on 3.9 before adopting it, this is roughly the signal: the new parser hasn’t caused any notable churn since October, and the point release is exactly the kind of unglamorous, expected maintenance you want to see. Worth scheduling the upgrade for early next year if you haven’t already.