docs: fix two unfollowable instructions in fork docs (stale SDK pin, phantom runtime.ts bump) - #23
Merged
Merged
Conversation
…ervation The "MCP SDK 1.22.0 inline-stdio regression" entry's Status line still says "pinned the SDK to ~1.21.2 until upstream ships a fix". That pin has not existed for nine SDK releases: package.json now carries `@modelcontextprotocol/sdk: ^1.30.0` (1.30.0 installed, 1.30.0 is latest on npm). The pin was lifted in fefe54e ("chore: bump sdk to 1.22.0 and adapt inline stdio test"), which also touched the inline stdio harness the repro depends on. A reader triaging a `typeName` failure today would follow this entry to a remediation that is not in place and a version floor the repo left long ago. Deliberately NOT claiming the regression is fixed — nobody has re-run the repro. The three live possibilities (fixed upstream / worked around in the test harness / no longer exercised) are named as open, and the entry now tells the reader to re-verify before relying on it. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Release step 1 said "Update version in package.json and src/runtime.ts",
but src/runtime.ts holds no version literal to update. It only imports
and re-exports MCPORTER_VERSION from ./version.js:
import { MCPORTER_VERSION } from './version.js';
export { MCPORTER_VERSION } from './version.js';
and src/version.ts derives that value at runtime by require-ing
package.json (falling back to $MCPORTER_VERSION, then '0.0.0-dev'). The
version has exactly one source of truth — package.json — and it
propagates on its own.
The instruction is unfollowable as written: a releaser either wastes
time hunting for a literal that is not there, or worse, invents one and
reintroduces the drift version.ts exists to prevent.
Co-Authored-By: Claude Fable 5 <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.
Two docs-reality corrections found by a 10x tech-debt scan. Docs-only; no code, no behavior change, no dependency change.
1.
docs/known-issues.md— SDK entry pins a version nine releases staleThe "MCP SDK 1.22.0 inline-stdio regression" entry's Status line still reads "pinned the SDK to
~1.21.2until upstream ships a fix."That pin has not been in effect for nine SDK releases:
~1.21.2package.json@modelcontextprotocol/sdk: ^1.30.0The pin was lifted in
fefe54e— "chore: bump sdk to 1.22.0 and adapt inline stdio test" — which also modifiedtests/cli-generate-cli.integration.test.ts, the very inline-stdio harness the entry's repro depends on.Today a reader triaging a
Cannot read properties of undefined (reading 'typeName')failure follows this entry to a remediation that isn't in place and a version floor the repo left long ago.This PR deliberately does not claim the regression is fixed. Nobody has re-run the repro. Three possibilities remain live — fixed upstream, worked around when the harness was adapted, or simply no longer exercised — and the entry now names them as open and tells the reader to re-verify before relying on it. Marking it "fixed" would trade one false statement for another.
2.
docs/RELEASE.md— step 1 names a file with no version to updateStep 1 said "Update version in package.json and src/runtime.ts."
src/runtime.tsholds no version literal. It only re-exports:and
src/version.tsderives the value at runtime byrequire-ingpackage.json(falling back to$MCPORTER_VERSION, then'0.0.0-dev'). The version has exactly one source of truth and propagates on its own.The instruction is unfollowable as written: a releaser either wastes time hunting for a literal that isn't there, or — worse — invents one and reintroduces exactly the drift
version.tsexists to prevent. Dropping the phantom half.Notes for the reviewer
docs/RELEASE.mdis heavily fork-delta (upstream's version is a different document), but theknown-issues.mdStatus line is byte-identical toupstream/main— it is stale there too. Expect a conflict on the next upstream sync, and consider sending the correction upstream.pnpm docs:sitebuilds clean (exit 0). This repo does have PR CI (.github/workflows/ci.ymlonpull_request→main), so checks will run normally.🤖 Generated with Claude Code