[myleshorton]

Summary

Apple's CloudKit ticket-delivery endpoint has now failed xcrun stapler staple with NSURLErrorDomain -1005 after notarization was already accepted twice in four days — the v10.0.4-beta.3 release run (Oct 25, run 40008674087) and the Oct 28 nightly dispatch (run 40765793487). Both outages outlasted retry-transient.sh's default three attempts, which only wait ~90 seconds in total, so the whole macOS build failed and each release needed a manual gh run rerun --failed.

• scripts/ci/retry-transient.sh gains a -b BACKOFF flag: the wait before attempt N+1 is N*BACKOFF seconds (default 30, the previously hardcoded value, so the gomobile call site is unchanged).
• The Makefile's STAPLER_RETRY becomes -n 6 -b 60: six attempts waiting 60/120/180/240/300s between them — 15 minutes of patience for a staple/validate that is idempotent and already past notarization, instead of failing the platform build.
• Leading-zero values are rejected for both -n and -b (a round-1 gate catch): $((...)) arithmetic reads 010 as octal 8 and chokes on 08, so -b 08 would have broken the retry loop instead of waiting 8 seconds.

Testing

• bash -n clean; exercised the script directly: success passthrough, retry loop with -n 2 -b 1 (1s elapsed, correct messages), non-matching -m failure exits immediately with the command's code, and 0/08/010/x/empty are rejected for both flags with exit 2.
• make -n notarize-darwin confirms both staple and validate expand to retry-transient.sh -n 6 -b 60 -- xcrun stapler ....
• Not covered: a real CloudKit outage can't be reproduced on demand; worst-case added wall-clock on a persistent staple failure is 15 minutes, well inside the macOS job's timeout.

Risk

• CI-only. Riskiest line: the shared option parsing in retry-transient.sh — the other call site (gomobile bind) passes neither -n nor -b and keeps the exact previous behavior (3 attempts, 30s base).

How to revert

Revert the commit; no state to undo.

Pre-PR review (local Codex gate)

• Rounds: 2 — final: round=2 range=origin/main...HEAD CODEX GATE: verdict=approve critical=0 important=0 minor=0 preexisting=0 reviewed=5
• Reviewed at: 73edcbbdb
• Exit: approve (high reasoning effort; the gate read both call sites, the workflows that invoke the Makefile targets, and the option-parsing edge cases)
• Step back: not needed
• Fixed: scripts/ci/retry-transient.sh — -b validation accepted leading-zero values that $((...)) reads as octal (-b 08 would error, -b 010 would wait 8s); both -n and -b now reject leading zeros.
• Dismissed: none
• Pre-existing (not introduced here, worth filing): none
• Proposed AGENTS.md additions: none

🤖 Generated with Claude Code

Summary by CodeRabbit

• Improvements
  • Automated retries for temporary stapling and validation failures now allow up to six attempts, with a 60-second interval between attempts. Retry intervals can also be configured for other automated tasks. Invalid retry counts and intervals, including values with leading zeros, are now rejected to help prevent unexpected retry behavior.

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: df44c50c-cca0-420c-8365-37964c40b789

Commits

Reviewing files that changed from the base of the PR and between 7d32f92 and 73edcbb.

Files selected for processing: 2

• Makefile
• scripts/ci/retry-transient.sh

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 7 remain after this review.

Walkthrough

The retry script now supports a configurable backoff and validates retry and backoff values. The Makefile configures the stapling retry command with six attempts and a 60-second backoff.

Changes: CI Retry Configuration

Layer / File(s)	Summary
Add configurable retry backoff scripts/ci/retry-transient.sh	The script adds -b BACKOFF, defaults it to 30 seconds, and waits attempt * backoff seconds before each retry. It rejects invalid values for -n and -b, including numbers with leading zeros.
Configure stapling retries Makefile	STAPLER_RETRY passes -n 6 -b 60 to the retry script.

Priority: Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Suggested reviewers: atavism

Merge Risk: Minimal · up to 73edc

The longer stapling retry window introduces no identified merge-blocking risk. Merge after normal checks pass.

Pre-merge checks: 5

Check name	Status	Explanation
Description Check		Check skipped - CodeRabbit’s high-level summary is enabled.
Title check		The title clearly and concisely describes the main change: extending retry time for CloudKit-related stapling outages.
Docstring Coverage		No eligible functions found in the changed files. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentat…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

No eligible functions found in the changed files. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentation-only edits and deletions within surviving functions are included; unchanged diff context and fully deleted functions are excluded. Analyzed 0 functions across 1 files. (1 skipped: 1 unsupported.)

Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

[Copilot]

🟢 Approved

No active review issues remain for the stapling retry changes.

No open issues 🎉

What changed in this PR

This PR makes macOS release builds more resilient to CloudKit outages after notarization by extending retries for stapling and validation.

Changes:

• Adds configurable retry backoff with -b, preserving the existing 30-second default.
• Configures stapler operations for six attempts with 15 minutes of total waiting per operation.
• Rejects leading-zero values for attempts and backoff to avoid Bash octal arithmetic issues.

File	Description
scripts/​ci/​retry-transient.sh	Adds configurable backoff, validates numeric options, and updates usage documentation.
Makefile	Sets stapler retries to six attempts with a 60-second base backoff.

_{Review effort: Balanced}