docs: correct the install page — wrong failure mode, wrong cut release, wrong support range - #576
Merged
Merged
Conversation
The install page promised behaviour that did not happen. Verified on real 3.9/3.10/3.11 venvs against PyPI before the fix: `pip install autogalaxy` did not fail, it backtracked to 2026.7.29.1 and installed a months-old, JAX-less stack silently. Three corrections: - "will fail with a no matching distribution error" — it did not. It now stops with an explanation, via the 2026.7.29.1.post1 tombstone release. - "dropped support ... in release 2026.4.5.3" — wrong release. That floor was reverted on 2026-04-30 and 2026.5.1.4 ... 2026.7.29.1 all shipped >=3.9 again. The real cut is 2026.7.29.2. - "pre-2026.4.5.3 releases on PyPI have been yanked" — they were not. 396 of 421 autolens releases are live, including every release a sub-3.12 pip was landing on. The section's whole premise was a yank that never happened. Also states the one remaining gap plainly: --only-binary=:all: skips sdists, so it steps past the tombstone to the old wheel. Issue: PyAutoLabs/PyAutoHands#238 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Python 3.14 was promoted on 2026-07-31: it is in the classifiers, and PyAutoHands python_matrix.yml runs it as a REQUIRED unit + smoke leg alongside 3.12 and 3.13. The installation overview still advertised "3.12 - 3.13", understating support and pointing 3.14 users away from a version that is tested every night. Reworded to "Python 3.12 or later", which matches `requires-python = ">=3.12"` and does not need editing again at the next promotion. Checked and deliberately not changed: PyAutoCTI's "Python 3.12 or 3.13" is accurate — its classifiers stop at 3.13, so it genuinely was not promoted. The workspace AGENTS.md files saying CI gates on "3.12 and 3.13" are accurate too: PyAutoHeart's reusable smoke-tests workflow defaults to `["3.12", "3.13"]` and none of them override it. Issue: PyAutoLabs/PyAutoHands#238 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Collaborator
Author
|
Added a second, related correction to the same install docs: Checked and deliberately not changed: PyAutoCTI's "Python 3.12 or 3.13" is accurate (its classifiers stop at 3.13 — it genuinely was not promoted), and the workspace |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Phase 2 of PyAutoLabs/PyAutoHands#238. The tombstone releases are live on PyPI (PyAutoHands #240, #242); this makes the install page describe what actually happens.
Three false claims, not one
1. The failure mode. The page said
pip install autogalaxyon 3.9/3.10/3.11 "will fail with a no matching distribution error". It did not. Verified on real venvs with pip 26.2.1, before the fix:No JAX, no warning, months out of date. Worse than the documented hard failure, because nothing told the user which version they were running.
2. The cut release. "We dropped support ... in release
2026.4.5.3" is the wrong release. That floor was reverted on 2026-04-30, and2026.5.1.4…2026.7.29.1all shipped>=3.9again. The real cut is2026.7.29.2.3. The yank. "Pre-
2026.4.5.3releases on PyPI have been yanked, so they will not install" — they were not. 396 of 421autolensreleases are live; the 25 yanked ones are a scattering from the autoconf-rename period. The section's entire premise was a yank that never happened, which is why the resolver fallback it claimed was blocked was in fact wide open.What it says now
The real cut release, why raising a floor does not retract a back catalogue, and the
2026.7.29.1.post1tombstone that now refuses the install. It also states the one remaining gap rather than hiding it:--only-binary=:all:skips source distributions, steps past the tombstone, and installs the old wheel silently.Pinned historical installs are unaffected and still resolve on old Pythons.
API Changes
None. Documentation only.
🤖 Generated with Claude Code