Repository navigation
chore: neutralize downstream-consumer references in tracked files - #157
Merged
Merged
Conversation
This repository is public. Tracked files referenced two private consumer
product names — the current one and its pre-rename predecessor — 40
occurrences across 12 files, including cross-repo citations to private
issue trackers. Those citations are unresolvable for any external reader
and disclose more than a public package should about its consumers.
Replacements are deliberately neutral ("a downstream consumer", "the
consumer's embedder"), not a substitution of one product name for the
other: naming the current product in a public repo is the same leak
wearing newer branding.
Cross-repo citations to private trackers are replaced with substantive
descriptions of what changed, since an issue number in a private repo
conveys nothing to a reader who cannot open it.
Code changes are limited to test-local identifiers: the WAL-concurrency
regression test and its print tag are renamed to describe the behaviour
they exercise (searchWriteStallRegressionTest, [StallRegression]),
matching the existing [PerfTest] precedent. No API, type, or behaviour
changes. Technical content of the affected ADRs is untouched — only the
names of who reported what.
Verified: a case-insensitive search for either product name and the
associated org name returns nothing across all tracked files; swift
build clean; SQLiteStorage WAL Concurrency suite 2/2 passing with the
renamed test.
Not addressed here: 16 commit messages in the existing history contain
the predecessor name. Rewriting them means a force-push on a public
repo, invalidating every clone, fork, and pinned SHA — including the
revision consumers currently depend on. That needs an explicit decision,
not a side effect of this cleanup.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
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
1 participant
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.
Why
This repository is public. Tracked files referenced two private consumer product names — the current one and its pre-rename predecessor — 40 occurrences across 12 files, including cross-repo citations to private issue trackers. Those citations resolve to nothing for an external reader and disclose more about the package's consumers than a public repo should.
What changed
Replacements are neutral — "a downstream consumer", "the consumer's embedder" — rather than substituting one product name for the other. Renaming the older name to the current one would be the same disclosure in newer branding.
Cross-repo citations to private trackers are replaced with substantive descriptions of what actually changed, since a private issue number tells an external reader nothing.
Code changes are limited to test-local identifiers. The WAL-concurrency regression test and its
printtag are renamed to describe the behaviour they exercise —searchWriteStallRegressionTestand[StallRegression]— matching the existing[PerfTest]precedent. No API, type, or behaviour changes. Technical content of the affected ADRs is untouched; only the names of who reported what.Verification
swift build— clean.swift test --filter SQLiteStorageConcurrency— 2/2 passed:Write completes while slow FTS read is in flight (WAL liveness)Bulk write loop is not stalled by a concurrent slow FTS scan—[StallRegression] tIsolated=0.0043s tConcurrentWrites=0.0041s ceiling=0.0343sNot addressed here — needs a decision
16 commit messages in the existing history contain the predecessor name. Rewriting them requires a force-push on a public repo, invalidating every clone, fork, and pinned commit SHA — including the revision consumers currently depend on. That is a deliberate call, not something to fold into a text cleanup.
No recurrence guard is included. A grep-based CI check would have to contain the forbidden strings in order to detect them, reintroducing into the tree exactly what it protects against. Resolving that properly (encoded pattern, CI secret, or partial match) is worth its own change.
Housekeeping
An earlier branch for this work,
chore/purge-legacy-consumer-name(PR #156, now closed), carries a commit message that itself contains the names. Repository rules block both force-push and deletion on it, so it needs an admin action to remove.🤖 Generated with Claude Code