Skip to content

Port: a rename's two names, grounded on the compare the port reads (#138) - #353

Merged
don-petry merged 3 commits into
mainfrom
adapters/rename-grounded
Oct 27, 2026
Merged

don-petry merged 3 commits into
mainfrom
adapters/rename-grounded

Conversation

@don-petry

@don-petry don-petry commented Oct 27, 2026 •

Copy link
Copy Markdown
Contributor

Closes #138.

The port #138 asks for was built with issue 284: adapters/pull.py's changed_paths reads every file the head changes from the compare of the base branch's head with it, gives a rename or a copy both names as {path, previous}, and sets complete.changed_paths only when the list holds exactly the pull request's count, each once. Decision 0020 records its rules. What was missing was the grounding: trials/P-43.json recorded a rename through the pull request's own list of files, which the port does not read, and the test held only that the fixture had one.

Part What changed
The grounding trials/P-43.json gains exchange 29, read after the trial, read-only, through the Maintainer's gh login, captured outside the repo and redacted by tools.redact_trial before it was read: GET /repos/combsmith/software-factory-canary/compare/{base}...{head} of the canary's pull request 48 (closed at the trial's end). Nothing was written to the canary
The test test_a_renamed_file_is_read_with_its_old_name_from_a_fixture now drives pull.changed_paths with that compare: three files, the rename trial-notes/p-43-renamed.md with previous trial-notes/p-1.md, whole; and the pull request's own list names the same three the same way
The record Decision 0072: the grounding, and one source of a rename's two names: the port gives {path, previous}, the Kernel matches both (0020), and the push refusal of decision 0070 takes the same two names as two lists
"Paginated in full" A compare lists at most 300 files and pages its commits, not its files; a change of more is read short by its count alone and deferred, as 0020 and issue 284 rule. 0072 says the box is met as "never read whole unless it holds every file"

Verify

  • make verify on an export of the first commit, run alone: every check but the context gate's record gap passes. It runs again once 0071 is on main and merged in.
  • One mutant, the rename's old name dropped, killed by hand; CI's shards are the gate.

Gates

🤖 Generated with Claude Code

)

trials/P-43.json gains one exchange, read after the trial and read-only through the Maintainer's gh login: the
compare of the canary's pull request 48, its base commit against its head, which adapters/pull.py's
changed_paths reads the files by. The port reads it whole, the rename with both its names, as the pull request's
own list of files names them; the test now drives the port with it, where it held only that the fixture had a
rename. Decision 0072 records the grounding, and that one {path, previous} shape carries a rename's names for
the port, the Kernel's rule of decision 0020 and the push refusal of decision 0070. A mutant for the old name.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@don-petry
don-petry requested a review from a team as a code owner October 27, 2026 16:45
@coderabbitai

coderabbitai Bot commented Oct 27, 2026 •

Copy link
Copy Markdown

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 6 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

Run configuration
Commits

Reviewing files that changed from the base of the PR and between 3f3895b and eddd2a5.

Files selected for processing: 5
  • docs/decisions/0072-a-renames-two-names-are-read-from-the-compare-the-port-reads-and-one-shape-carries-them.md
  • sim/grounding.py
  • tests/test_adapters.py
  • tools/mutants.py
  • trials/P-43.json
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 27, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-27T17:39:27.016461Z eddd2a5 New commits
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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

…mpare's limit unverified (#138)

The Claude review's read: decision 0070's push_refusal takes a push's paths and their previous paths as two
lists, not {path, previous} records, and is not yet on main; the 300-file limit is as GitHub documents it, which
no trial records. The Kernel's matching of either name gets its row.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 662043fd47

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".

…de by side, and 0072 naming 0070 by its file

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@sonarqubecloud

Copy link
Copy Markdown

@don-petry
don-petry added this pull request to the merge queue Oct 27, 2026
Merged via the queue into main with commit 78c0bb1 Oct 27, 2026
20 checks passed
@don-petry
don-petry deleted the adapters/rename-grounded branch October 27, 2026 18:13
don-petry added a commit that referenced this pull request Oct 27, 2026
…est's record 0073

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
YasirBinMudassir pushed a commit to YasirBinMudassir/software-factory that referenced this pull request Nov 7, 2026
…ld is unreadable, and five rules are held (combsmith#383)

* Argon's owed reads of three merged PRs: a list pattern no path may hold is unreadable, and five rules are held

Argon (gemini-4-argon-high) read the merged combsmith#348, combsmith#353, combsmith#355 and combsmith#379 against their parents once its quota
came back. Each finding was checked by hand on main. One was false: protected.parse has no AttributeError
path, and 620 malformed lists all end in Unreadable. combsmith#353 had none.

- push_refusal refuses a list pattern with a backslash or a control character as push_list_unreadable: no
  path it is matched against may hold one, so it protected nothing. The hooks schema already keeps such a
  pattern out of config/hooks.json; two counter-examples now hold that, so the Kernel's readable_pattern
  never meets one from main's list and is unchanged.
- adapters/rules.py: an answer with no 200 whose body is the file's, a size that is no integer, and a
  repo named "." are each held by a case of their own.
- adapters/applied.py: no current content, missing or empty, supersedes nothing even where a later Run's
  record names none the same way.
- tools/check_workflows.py: the clause that refused an empty with: or env: on a step that takes none is
  dropped, as that step has the key as one of its own, refused as "has exactly"; a test holds that each
  step's given keys are its allowed ones. A step of text after the last step is refused.

No record: Boundaries rows on 0070, 0075 and 0080. Eight mutants, two re-pointed and one changed, each
killed by hand.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Hooks: a list path with a control character or a backslash makes config/hooks.json unreadable (combsmith#383)

Muse's read of combsmith#383. The push refusal turned such a pattern away, but the Kernel's own readable_pattern
still read it, so a list holding one would protect less instead of deferring. The Kernel is read from the
Release's own config/hooks.json, which CI's schema check keeps clean, so the route is not reached today.
Even so, protected.parse is the one reader both the Kernel's configuration and a push's list go through,
and it already refuses ".." and "**". It now refuses a path with a control character or a backslash as
Unreadable too, and the Kernel stays as it is.

The Claude review and the security pass approved, with no finding this PR made. Their one minor, a quoted
empty `with: ""` beside a step's keys, was on main before this PR and fails closed, since GitHub would not
load such a file.

One test and one mutant, killed by hand; the 0070 row names the test.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Port: the paths a pull request changes, with both names of a rename

1 participant