fix(mcp): don't mask a resolver failure with an un-wired instance#156
Merged
Conversation
resolveOrConstruct caught any resolver throw and fell back to new Ctor(), so a real DI misconfiguration (a registered token whose construction fails) was silently replaced by a fresh un-wired instance, contradicting the package's loud-resolver philosophy. Add an optional McpResolver.has(token) hook. When present (createResolver now implements it), only owned tokens go through resolve() and a genuine construction failure propagates. Resolvers without has() keep the legacy fallback, so the change is backward compatible. Closes #149
5 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #149.
resolveOrConstructcaught any resolver throw and fell back tonew Ctor(). So a real DI misconfiguration — a token the container does own, whose construction throws because a constructor dependency is missing — was silently replaced by a fresh, un-wired instance, masking the bug. This contradicts the loud-resolver philosophyresolveHandleDeps/createResolverenforce elsewhere.The interface has no way to tell "no binding" from "threw", so I added an optional
McpResolver.has?(token)hook:createResolvernow does) get a precise path: only owned tokens go throughresolve(), and a genuine construction failure propagates loudly.resolvemiss falls back to a plain constructor), so this is backward compatible.Tests: new
runtime/handle-deps.test.ts— propagation on an owned-but-failing token, skip-resolve for un-owned tokens, and the two legacy fallback paths.pnpm typecheck+pnpm testgreen (111, +6). Minor changeset (additivehas?+createResolver().has).