feat(cymbal): set posthog-python wire-order cutoff at 7.31.0

[cat-ph]

Problem

The eighth and final SDK flip of the wire-order standardization (PostHog/sdk-specs#11): posthog-python ships $exception_list in canonical order — caught/outermost exception first, root cause last — in 7.31.0 (PostHog/posthog-python#728, latest release posthog-v7.30.1 + that PR's lone minor changeset; no other changesets pending on main).

Python is unique in the table: its fix reverses the exception list, not frames (python frames were always bottom-up). It's also the mildest flip by construction — single-exception events (the vast majority) are unchanged by the reversal and hash identically either way; only chained-exception events change on the wire.

Changes

• posthog-python gets canonical_since: Some(7.31.0); everything below (or unparseable) keeps having its exception list flipped.
• New gate test asserting the list-order semantics specifically: below-cutoff versions get the list reversed with frames untouched, at/above passes through, and legacy reconstruction (re-flipping the list) ignores the cutoff.

Merge sequencing

1. Re-verify posthog-python's latest posthog-v* release is still 7.30.x immediately before merging (other packages in the repo — e.g. the openfeature provider — release independently and don't claim this version stream).
2. Merge this, wait for cymbal autodeploy.
3. Merge fix(errors): emit $exception_list in canonical order (caught-first, root cause last) posthog-python#728 and release as 7.31.0; no other posthog package minor in between.

Watch et-wire-order-launch as with the previous seven. Highest event volume of the train (~2.9M/day) but the smallest behavioral delta: chained exceptions only, no frame changes, no resolution reshaping — reconstruction is byte-exact.

How did you test this code?

Automated only: new python_cutoff_gates_exception_list_normalization_by_version unit test (16 normalization tests green); clippy -D warnings and fmt clean. Eighth instance of the reviewed cutoff pattern. The companion flip branch was rebased onto posthog-python main (58 commits of drift), verified the reversal-removal is intact, with 156 tests + ruff green.

Automatic notifications

• Publish to changelog?
• Alert Sales and Marketing teams?

Docs update

Rollout process documented in the sdk-specs change.

🤖 Agent context

Autonomy: Human-driven (agent-assisted)

Claude Code authored this under my direction as the final step of the flip rollout, following the sequencing contract established across #69837/#70879/#71365/#72350/#72962/#73319. Target version derived from the latest posthog-v release plus #728's pending changeset.

[greptile-apps]

Prompt To Fix All With AI

Fix the following 1 code review issue. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 1
rust/cymbal/src/modes/processing/normalization.rs:359-371
**Test uses reversed fixture orientation**

The post-cutoff and reconstruction assertions reuse `make_list()`, which places `RootCause` before `Wrapper`, even though canonical post-flip order places the outermost exception first. This tests the inverse flow and does not verify that canonical Python input remains unchanged before reconstruction restores legacy root-cause-first order.

_{Reviews (1): Last reviewed commit: "fix(cymbal): move posthog-python cutoff ..." | Re-trigger Greptile}

[deployment-status-posthog]

Deploy status

Environment	Status	Deployed At	Workflow
dev	✅ Deployed	2026-07-27 18:50 UTC	Run
prod-us	✅ Deployed	2026-07-27 19:17 UTC	Run
prod-eu	✅ Deployed	2026-07-27 19:18 UTC	Run