summaryrefslogtreecommitdiffhomepage
path: root/packages/core/tests
diff options
context:
space:
mode:
authorAdam Malczewski <[email protected]>2026-05-30 23:15:18 +0900
committerAdam Malczewski <[email protected]>2026-05-30 23:15:18 +0900
commit4e636511ae748d606d8871f5068a2bd18b386bd0 (patch)
treea5e0726d71d9d88d09d938ea2318a61e36ade68f /packages/core/tests
parent624b808da0f2f8bbad8a4fbbcca3f82f24ecfc47 (diff)
downloaddispatch-4e636511ae748d606d8871f5068a2bd18b386bd0.tar.gz
dispatch-4e636511ae748d606d8871f5068a2bd18b386bd0.zip
chore(notes): collect loose root docs into notes/; add reconcile edge-cases note
Move all loose root-level .md files (plans, reports, gemini reviews, incident notes) into a single notes/ directory, and update the doc-reference breadcrumbs in code comments/test labels to the notes/ path. Add notes/queue-interrupt-reconcile-edge-cases.md: documents why the queue/interrupt/turn-sealed reconcile path keeps surfacing edge cases (a catalog of the four review-pass bugs, the no-loss/no-duplicate invariants, the recommended membership-based reconcile refactor, and interleaving-test guidance).
Diffstat (limited to 'packages/core/tests')
-rw-r--r--packages/core/tests/agent/agent.test.ts6
-rw-r--r--packages/core/tests/llm/provider.test.ts2
2 files changed, 4 insertions, 4 deletions
diff --git a/packages/core/tests/agent/agent.test.ts b/packages/core/tests/agent/agent.test.ts
index 6c2b452..a12ea07 100644
--- a/packages/core/tests/agent/agent.test.ts
+++ b/packages/core/tests/agent/agent.test.ts
@@ -1038,7 +1038,7 @@ describe("Agent", () => {
expect(rc).toBeUndefined();
});
- // ─── Prompt-caching: tool-result grouping & breakpoints (claude-report.md) ──
+ // ─── Prompt-caching: tool-result grouping & breakpoints (notes/claude-report.md) ──
it("groups a turn's tool results into a SINGLE role:'tool' message (Root Cause 2)", async () => {
// The agent batches three distinct read_file calls in one step. The
@@ -1158,7 +1158,7 @@ describe("Agent", () => {
}
});
- // ─── Tool-call dedup (tool-runner-duplication-incident.md) ─────────────────
+ // ─── Tool-call dedup (notes/tool-runner-duplication-incident.md) ─────────────────
it("deduplicates byte-identical tool calls within a single batch", async () => {
// Claude can degenerate and emit the same tool call (same name + args)
@@ -1265,7 +1265,7 @@ describe("Agent", () => {
// A 3-step tool turn. The messages for steps 0 and 1 must serialize
// identically in the step-2 request and the step-3 request — that
// byte-stability is what lets Anthropic's rolling prompt cache extend
- // instead of re-writing the whole prefix every step (cache-miss-report.md).
+ // instead of re-writing the whole prefix every step (notes/cache-miss-report.md).
// Uses the openai-compatible provider so no cacheControl markers (which
// intentionally move each step) obscure the content comparison.
let n = 0;
diff --git a/packages/core/tests/llm/provider.test.ts b/packages/core/tests/llm/provider.test.ts
index 9e6b2ad..12b6350 100644
--- a/packages/core/tests/llm/provider.test.ts
+++ b/packages/core/tests/llm/provider.test.ts
@@ -179,7 +179,7 @@ describe("createClaudeOAuthProvider", () => {
}
});
- it("sends the anthropic-beta header so prompt-caching is honored (claude-report.md Root Cause 1)", () => {
+ it("sends the anthropic-beta header so prompt-caching is honored (notes/claude-report.md Root Cause 1)", () => {
// Without `anthropic-beta: ...,prompt-caching-scope-2026-01-05,...` the
// Anthropic API silently ignores every `cache_control` marker we attach
// to messages, producing a 0% cache hit rate. `@ai-sdk/anthropic` does