Stop re-proposing the two upgrades that have nowhere to land - #136
Merged
Conversation
Both of these reopen every week, fail the same way, and cost a review each time. Neither is fixable in this repository, so the honest move is to say so in the config rather than let the queue relitigate it. GitHub.Copilot.SDK — ignored outright, not by update-type. SquadAgent.cs uses CopilotClient through Microsoft.Agents.AI.GitHub.Copilot, and MAF 1.11.0-rc1 declares `GitHub.Copilot.SDK [1.0.0, )`. The open range means restore accepts a newer SDK and the compiler then rejects it — CS0012, "defined in an assembly that is not referenced ... Version=1.0.0.0", on net8.0, net9.0 and net10.0 alike. Note the break arrived on 1.0.3 -> 1.0.11, a *patch*, which is why a semver-major ignore would not have caught it. Lift this with MAF: 1.19.0+ declares `[1.0.5, )`. typescript — majors only. typescript-eslint still peer-requires `typescript >=4.8.4 <6.1.0` at 8.69.0, its newest release, so TS 7 has no supported lint toolchain. npm "resolves" the conflict by nesting duplicate @typescript-eslint trees, which means lint:eslint would run a parser against a compiler it does not support — failing quietly rather than loudly, which is worse. Minor and patch TypeScript updates still flow. Both comments name the condition that should retire them, so whoever finds this later can tell whether it still holds.
Contributor
🛫 PR Readiness Check
PR Scope: 🔧 Infrastructure
|
| Status | Check | Details |
|---|---|---|
| ✅ | Single commit | 1 commit — clean history |
| ✅ | Not in draft | Ready for review |
| ✅ | Branch up to date | Up to date with dev |
| ❌ | Copilot review | No Copilot review yet — it may still be processing |
| ✅ | Changeset present | No source files changed — changeset not required |
| ✅ | Scope clean | No .squad/ or docs/proposals/ files |
| ✅ | No merge conflicts | No merge conflicts |
| ✅ | Copilot threads resolved | No Copilot review threads |
| ✅ | CI passing | All checks passing |
Files Changed (1 file, +27 −0)
| File | +/− |
|---|---|
.github/dependabot.yml |
+27 −0 |
Total: +27 −0
This check runs automatically on every push. Fix any ❌ items and push again.
See CONTRIBUTING.md and PR Requirements for details.
Contributor
🟢 Impact Analysis — PR #136Risk tier: 🟢 LOW 📊 Summary
🎯 Risk Factors
📦 Modules Affectedci-workflows (1 file)
This report is generated automatically for every PR. See #733 for details. |
This was referenced Sep 1, 2026
DeDuva
added a commit
that referenced
this pull request
Sep 1, 2026
#136 ignored GitHub.Copilot.SDK outright, on the reasoning that it was not independently upgradable while Microsoft.Agents.AI.GitHub.Copilot sat at 1.11.0-rc1. That comment ended with a condition — "lift this together with MAF, not before: 1.19.0+ declares [1.0.5, )" — and #137 met it about an hour later, moving MAF to 1.20.0 and the SDK to 1.0.5 in one PR, green on both .NET jobs across net8.0, net9.0 and net10.0. Leaving the ignore in place would now freeze the SDK at 1.0.5 indefinitely, including for security releases, behind a comment telling the next reader to remove it because a condition that is already true. So the ignore goes, and the constraint is expressed where it actually belongs: GitHub.Copilot.SDK joins both Microsoft groups. It is not a Microsoft.* package, so without that it is proposed alone — and alone is the shape that breaks, since SquadAgent.cs reaches CopilotClient through MAF's re-export and the two must compile as a pair. This is a nudge rather than a guarantee: grouping bundles whatever updates exist in a run, so a solo SDK bump can still appear. That is an acceptable residual risk because the failure mode is loud rather than silent — #124 turned both .NET jobs red on CS0012, which is exactly what a review gate is for. Trading a permanent freeze for a well-caught failure is the better side of that deal.
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.
Companion to #135. That one fixed updates Dependabot proposed but could not lock; this one covers updates it proposes that cannot build or cannot lint, and will keep proposing weekly until told otherwise.
Both PRs stay open — this only stops the re-proposal treadmill.
GitHub.Copilot.SDK— ignored outright (#124)Not by update-type, and that distinction matters: the break arrived on 1.0.3 → 1.0.11, a patch. A
semver-majorignore would have sailed straight past it.SquadAgent.csusesCopilotClientviaMicrosoft.Agents.AI.GitHub.Copilot, pinned at1.11.0-rc1, which declares:An open lower bound, so NuGet restore accepts any newer SDK — and the compiler then refuses it:
on
net8.0,net9.0andnet10.0alike, on both ubuntu and windows.The SDK is only independently upgradable once MAF moves with it: 1.19.0+ declares
[1.0.5, )and is built against the newer line. The directPackageReferenceis itself flagged in the csproj as a workaround pending microsoft/agent-framework#6457, so one upgrade will likely retire both.typescript— majors only (#33, #26)typescript-eslintpeer-requirestypescript >=4.8.4 <6.1.0, and that cap is still there at 8.69.0, its newest release. TS 7 has no supported lint toolchain.What makes this worse than a plain failure: npm does not refuse the install, it nests duplicate
@typescript-eslinttrees to satisfy the impossible peer graph.npm run lint:eslintwould then run typescript-estree against a compiler it does not claim to support — degrading quietly instead of failing loudly.Minor and patch TypeScript updates are unaffected and still flow.
Why not just close the PRs
Closing them clears the list until Dependabot's next run re-creates both, unchanged and still broken. The ignore is the part that actually holds. Each entry carries the specific condition that should retire it — a MAF version range, a peer range — so a future reader can check whether it still applies instead of guessing.