Repository navigation
fix(init): finish rebrand on all user-facing print output - #3
Merged
Merged
Conversation
PR #1 caught the load-bearing hook-rewrite emit prefix, but `init.rs` itself still printed dozens of `RTK …` strings in success/status/help output. Empirically surfaced by a fresh Framework install — `contextcrawler init -g` reported: RTK hook registered (global). RTK.md: /home/.../CONTEXTCRAWLER.md (label/path mismatch) CLAUDE.md: @RTK.md reference added (the actual ref written to CLAUDE.md is now @CONTEXTCRAWLER.md via RTK_MD_REF) Scope of this PR: * Replace user-visible `RTK …` print/eprintln literals with `ContextCrawler …` (status messages, uninstall messages, codex config, agent hooks for cline / windsurf / kilocode / antigravity). * Replace hardcoded `RTK.md` / `@RTK.md` strings with interpolation of the existing `RTK_MD` / `RTK_MD_REF` constants (which already evaluate to `CONTEXTCRAWLER.md` / `@CONTEXTCRAWLER.md`). This keeps the label and the path consistent and centralizes future renames. * Updated the `init -g` usage help text to match. * `verify` self-check messages and Codex configuration listing both go through the same constants now. Out of scope (intentionally untouched): * Doc comments and source-level `RTK_MD` / `RTK_BLOCK_START` etc. identifier names — kept for upstream-rebase friction. * `LEGACY_RTK_MD` / `LEGACY_RTK_MD_REF` references — these specifically point at the legacy filename and must NOT be renamed; they exist so the install path can clean up post-upstream-rtk installs. * `RTK_INSTRUCTIONS` legacy claude-md block content — already replaced by `RTK_SLIM` for the active install path, but the legacy constant is preserved as-is for the `--claude-md` migration code. * gain.rs / cc_economics.rs strings — already addressed in PR #1. Verified: `cargo test` → 1784 passed / 0 failed / 6 ignored. Independence from PR #1 / PR #2: branched off develop, touches only init.rs print statements. No conflicts expected with either pending PR. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
3 of 4 tasks
This was referenced May 14, 2026
Merged
noogalabs
pushed a commit
to noogalabs/contextcrawler
that referenced
this pull request
Jun 4, 2026
…e tests Addresses Codex round-2 review findings on the exec_capture hardening branch: Truncation contract (Codex finding thehoff#3): - Add truncated_stdout / truncated_stderr bools to CaptureResult. Silent truncation was a real API ambiguity — callers parsing JSON on a capped prefix would treat parse failures as real errors. - New drain_with_cap helper reads at most `max` bytes then peeks one byte to detect whether the stream had more pending. Detection is exact: a stream that ends precisely at the cap reports truncated=false (no false positives at exact-EOF boundary). - exec_capture_with_limits propagates both flags through. Edge cases (Codex finding thehoff#6): - truncated flag set/clear at boundary (3 tests including exact-EOF no-false-positive) - very-short deadline race (1ms) — either Ok or Err is acceptable, no panic/hang/orphan - cap-of-0 pathological config — child still runs, output empty, flag set if child wrote anything - stderr-only flood completes without stdout-drain deadlock (validates the threaded-drain design holds when one stream is dead) - kill-on-already-exited-child does not surface as error (back-to-back short calls confirm no leftover state from unused timeout-error path) - drain_with_cap helper unit tests for the truncation peek logic Suite: 1848 → 1856 passing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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.
Summary
PR #1 caught the load-bearing hook-rewrite emit prefix, but `init.rs` itself still printed dozens of `RTK …` strings in success / status / help / uninstall output. Empirically surfaced by a fresh Framework install — `contextcrawler init -g` reported:
```
RTK hook registered (global).
RTK.md: /home/.../CONTEXTCRAWLER.md ← label says RTK.md, path is CONTEXTCRAWLER.md
CLAUDE.md: @RTK.md reference added ← actual ref written is @CONTEXTCRAWLER.md
```
What changed
Out of scope (intentionally untouched)
Test plan
Independence
Branched off `develop`, touches only init.rs print statements. No conflicts expected with PR #1 or PR #2.
🤖 Generated with Claude Code