[timothyanderson096-ocdealcheck]

Summary of Changes

• Allow BitTiming.iterate_from_sample_point() and BitTiming.from_sample_point() to use controller-specific tseg1_max, tseg2_max, and brp_max limits.
• Preserve the classical CAN bitrate range and strict standard-limit validation, while allowing the bounded solver to use extended controller limits.
• Retain and enforce controller constraints in both clock-recreation paths, including the prescaler shortcut.
• Add regression, boundary, documentation, and towncrier coverage.

At 160 MHz / 250 kbit/s / 87.5%, the default segment limits now allow brp=40, tseg1=13, tseg2=2. Explicit extended limits of 256/128/512 allow brp=4, tseg1=139, tseg2=20. A brp_max=28 control rejects the configuration when no candidate fits the default segment limits.

Clock recreation also respects smaller controller limits: an 8 MHz / 500 kbit/s timing with brp_max=1 cannot be recreated at 16 MHz by returning brp=2; the same change is accepted when brp_max=2. Strict construction with tseg1=17 still raises independently of the stored fallback limits.

Related Issues / Pull Requests

• Closes BitTiming.from_sample_point rejects valid timing solutions due to hardcoded register limits #2083

Type of Change

• Bug fix
• New feature
• Documentation update
• Refactoring
• Other (please describe):

Checklist

• I have followed the contribution guide.
• I have added or updated tests as appropriate.
• I have added or updated documentation as appropriate.
• I have added a news fragment for towncrier.
• Applicable local checks pass as listed below.
• Hosted checks on the latest revision have completed successfully.

Testing

Revision 33a212a86f94682b9ca740913e7163134f5e4efc, locally on Python 3.14.10:

• Timing and timing-utility tests: 36 passed.
• New regression against the preceding revision: 1 failed, 2 passed; all three pass after the fix.
• Configured tox -e lint: Black, Ruff and Pylint pass. Black ran with one worker because this environment restricts multiprocessing.
• tox -e type: all configured mypy targets, Python 3.11–3.15, pass.
• tox -e docs: Sphinx HTML with warnings as errors succeeds; 39 doctests pass.
• Cross-platform full-suite, documentation, packaging and static checks are running in hosted CI.

Additional Notes

No physical CAN controller was tested locally. The change was prepared with Codex assistance; the local results above were run against the published source tree.

Timothy Anderson — Spotted Ridge Engineering / OC Labs · GDN (Global Decision Network)
timothyanderson096@gmail.com

[mergify]

Tick the box to add this pull request to the merge queue (same as @mergifyio queue).

• Queue this pull request

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8865a0a935

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[zariiii9003]

How can this ever fail if you set self._tseg1_max = max(16, tseg1) in the init method? Am i missing something?

[timothyanderson096-ocdealcheck]

tseg1_max in this check is the _validate() argument (default 16), rather than self._tseg1_max. The stored attributes retain limits for clock recreation; the sample-point factory replaces those fallback values with the caller's explicit controller limits. For example, constructing with tseg1=17, strict=True still raises ValueError, even though the stored fallback maximum is 17. I've added a regression for that case and clarified the comment.

While checking this, I found a separate gap in the clock-recreation shortcut: a timing created at 8 MHz / 500 kbit/s with brp_max=1 could be recreated at 16 MHz with brp=2. In 33a212a, the shortcut validates against the retained controller limits before returning, then falls back to the bounded solver when necessary. The new regression rejects that case and accepts the brp_max=2 boundary control.

Local validation on Python 3.14.10: 36 timing/util tests passed; configured lint and all five mypy targets (3.11–3.15) passed; Sphinx HTML and 39 doctests passed. Black used one worker for this environment. The three intentional internal-access annotations follow the existing convention in this file. Hosted CI is running on the new revision.

Timothy Anderson — Spotted Ridge Engineering / OC Labs · GDN (Global Decision Network)
timothyanderson096@gmail.com