docs(bf-1j21w): document verification that assert_stderr_contains does not exist on ExtractionResult

The method exists only on TestExecutionResult in tests/encryption_fixtures.rs,
not on ExtractionResult as the bead description specifies.

ExtractionResult has no stderr field and no impl block with assertion methods.
It is a data structure (JSON-serializable output of PDF extraction), not a
command execution result wrapper.

This bead cannot be closed as the specified method does not exist on the
specified type.

notes/bf-1j21w.md

@@ -0,0 +1,102 @@
+# Verification Note: bf-1j21w - assert_stderr_contains on ExtractionResult
+
+## Task
+Verify that the `assert_stderr_contains` method on `ExtractionResult` is implemented and functional.
+
+## Findings
+
+### Method Does NOT Exist on ExtractionResult
+
+After a thorough search of the codebase:
+- `ExtractionResult` is defined in `/home/coding/pdftract/crates/pdftract-core/src/extract.rs` (line 237)
+- `ExtractionResult` has **NO impl block** with any assertion methods
+- The only associated function is `result_to_json()` (line 1483), which is a standalone function, not a method
+
+### assert_stderr_contains Exists Elsewhere
+
+The `assert_stderr_contains` method **DOES** exist, but on a different type:
+- **Type:** `TestExecutionResult`
+- **Location:** `/home/coding/pdftract/tests/encryption_fixtures.rs` (line 243)
+- **Purpose:** Test helper for CLI command execution results
+
+### Code Comparison
+
+**TestExecutionResult::assert_stderr_contains** (exists):
+```rust
+impl TestExecutionResult {
+    pub fn assert_stderr_contains(&self, text: &str) -> &Self {
+        let stderr = self.stderr();
+        assert!(
+            stderr.contains(text),
+            "Expected stderr to contain '{}', got: {}",
+            text,
+            stderr
+        );
+        self
+    }
+}
+```
+
+**ExtractionResult** (no such method):
+```rust
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct ExtractionResult {
+    pub fingerprint: String,
+    pub pages: Vec<PageResult>,
+    pub metadata: ExtractionMetadata,
+    pub signatures: Vec<SignatureJson>,
+    pub form_fields: Vec<FormFieldJson>,
+    pub links: Vec<LinkJson>,
+    pub attachments: Vec<AttachmentJson>,
+    pub threads: Vec<ThreadJson>,
+    pub javascript_actions: Vec<JavascriptActionJson>,
+}
+```
+
+## Conclusion
+
+The `assert_stderr_contains` method does **NOT** exist on `ExtractionResult`. The task asks to verify a method that has not been implemented.
+
+### Possible Explanations
+
+1. **Typo in task description:** The task may have meant `TestExecutionResult` instead of `ExtractionResult`
+2. **Missing implementation:** The method was planned but never implemented on `ExtractionResult`
+3. **Wrong bead:** This may be testing the wrong type
+
+### Why It Doesn't Make Sense for ExtractionResult
+
+Looking at the structure:
+- `ExtractionResult` is the **output** of PDF extraction (JSON serializable data)
+- It has **no stderr field** - it's not a command execution result
+- It contains: pages, metadata, signatures, form_fields, links, attachments, threads, javascript_actions
+- stderr/stdout are properties of **process execution**, not extraction results
+
+The `TestExecutionResult` type (which DOES have `assert_stderr_contains`) wraps `std::process::Output` and is used for testing CLI commands that produce stderr.
+
+## Recommendation
+
+**DO NOT CLOSE** - This bead describes verification work for a method that doesn't exist. One of:
+
+1. Update task to verify `TestExecutionResult::assert_stderr_contains` instead
+2. Implement the method if it's actually needed (though it doesn't make semantic sense)
+3. Mark bead as superseded/invalid if the task was based on incorrect assumptions
+
+## Verification Commands Run
+
+```bash
+# Search for assert_stderr_contains in codebase
+grep -rn "assert_stderr_contains" /home/coding/pdftract --include="*.rs"
+# Results: Only found in tests/encryption_fixtures.rs on TestExecutionResult
+
+# Search for ExtractionResult impl blocks
+grep -n "impl ExtractionResult" /home/coding/pdftract/crates/pdftract-core/src/extract.rs
+# Results: None found
+
+# Search for any assert_stderr on ExtractionResult
+grep -rn "assert_stderr" /home/coding/pdftract/crates --include="*.rs"
+# Results: None found
+```
+
+## Status
+
+**FAIL** - Method does not exist on the specified type.

notes/bf-224fc.md

@@ -0,0 +1,51 @@
+# bf-224fc: Add basic test function skeleton to forms_integration.rs
+
+## Status: COMPLETE
+
+## Summary
+
+The file `tests/forms_integration.rs` already contains multiple test functions that meet all acceptance criteria.
+
+## Acceptance Criteria Verification
+
+### ✅ File contains at least one `#[test]` function
+The file contains 3 test functions:
+1. `test_discover_pdf_fixtures` (line 42) - Tests PDF fixture discovery
+2. `test_cli_extract_json_on_fixtures` (line 127) - Tests CLI invocation with bounded timeout
+3. `test_forms_extraction` (line 233) - Tests forms extraction and JSON serialization
+
+### ✅ Test function compiles without syntax errors
+Verified with `cargo test --test forms_integration --no-run` - compilation succeeds.
+
+### ✅ Test function does something non-empty
+All three test functions contain:
+- Assertions (e.g., `assert!(bin.exists(), ...)`)
+- Logic for discovering fixtures
+- CLI invocation with timeout protection
+- JSON serialization verification
+- Print statements for debugging
+
+## Additional Notes
+
+The test file includes:
+- `discover_pdf_fixtures()` helper function for recursive PDF discovery
+- `pdftract_bin()` helper to locate the compiled binary
+- `wait_with_timeout()` helper to prevent hanging tests (critical for test hygiene)
+- Graceful handling when no fixtures exist (tests return early rather than fail)
+
+## Test Hygiene Considerations
+
+The tests implement the timeout protection pattern recommended in CLAUDE.md:
+- Uses `wait_with_timeout()` with bounded waits (10 seconds per fixture)
+- Falls back to release binary if debug binary not found
+- Returns early when fixtures directory is empty
+- Does not fail the entire test suite if fixtures are missing
+
+This prevents the exact issue described in the test hygiene section where hung tests can stall the entire marathon.
+
+## Verification
+
+- File path: `tests/forms_integration.rs`
+- Compilation: PASS (cargo check --tests)
+- Test function count: 3
+- Code review: All tests are non-empty and functional