Merge pull request #195 from firedancer-io/cmoyes/protocolfix

Add support for new FCNG protocol changes, friendlier error messages

Author: cmoyes-jump

commands.md

@@ -166,7 +166,7 @@ $ solana-conformance debug-mismatches [OPTIONS]
 * `-l, --section-limit INTEGER`: Limit number of fixture per section  [default: 0]
 * `--use-ng`: Use fuzz NG CLI (fuzz list/download repro) instead of API scraping  [default: True]
 * `-d, --debug-mode`: Enables debug mode, which spawns a single child process for easier debugging
-* `--all-artifacts`: Download all artifacts per repro (default: only latest)
+* `--all-artifacts`: (Deprecated, all artifacts are now always downloaded)
 * `--help`: Show this message and exit.
 
 ## `solana-conformance decode-protobufs`
@@ -268,7 +268,7 @@ $ solana-conformance download-fixtures [OPTIONS]
 * `-p, --num-processes INTEGER`: Number of parallel download processes  [default: 4]
 * `--use-ng`: Use fuzz NG CLI (fuzz list/download repro) instead of API scraping  [default: True]
 * `--interactive / --no-interactive`: Prompt for authentication if needed  [default: interactive]
-* `--all-artifacts`: Download all artifacts per repro (default: only latest)
+* `--all-artifacts`: (Deprecated, all artifacts are now always downloaded)
 * `--help`: Show this message and exit.
 
 ## `solana-conformance exec-fixtures`

src/test_suite/fuzzcorp_api_client.py

@@ -32,19 +32,60 @@ class LineageRepro:
     created_at: datetime
     count: int
     all_verified: bool
+    # Optional extra metadata available on newer APIs
+    drivers: Optional[List[str]] = None
+    config_indices: Optional[List[int]] = None
 
     @classmethod
     def from_dict(cls, data: Dict[str, Any]) -> "LineageRepro":
-        created_at_str = data["CreatedAt"]
-        # Handle both RFC3339 formats
-        if created_at_str.endswith("Z"):
-            created_at_str = created_at_str[:-1] + "+00:00"
+        # Hash is required in all known schemas
+        hash_value = data.get("Hash") or data.get("hash")
+        if not hash_value:
+            raise KeyError("Hash")
+
+        # CreatedAt is present in both schemas, normalize if available
+        created_at_str = data.get("CreatedAt") or data.get("created_at")
+        if created_at_str:
+            # Handle both RFC3339 formats
+            if created_at_str.endswith("Z"):
+                created_at_str = created_at_str[:-1] + "+00:00"
+            created_at = datetime.fromisoformat(created_at_str)
+        else:
+            # Fallback to a stable value so sorting still works
+            created_at = datetime.min
+
+        # Count:
+        # - prefer explicit Count from legacy schema
+        # - otherwise derive from ConfigIndices (one count per config index)
+        # - final fallback is a single repro
+        if "Count" in data:
+            count = int(data["Count"])
+        else:
+            cfg_indices = data.get("ConfigIndices") or data.get("config_indices") or []
+            if isinstance(cfg_indices, list):
+                count = max(len(cfg_indices), 1) if cfg_indices else 1
+            else:
+                # Unexpected shape: treat as a single repro
+                count = 1
+
+        # AllVerified:
+        # - legacy AllVerified flag if present
+        # - otherwise fall back to the newer Verified flag
+        if "AllVerified" in data:
+            all_verified = bool(data["AllVerified"])
+        else:
+            all_verified = bool(data.get("Verified") or data.get("verified") or False)
+
+        drivers = data.get("Drivers") or data.get("drivers")
+        cfg_indices_val = data.get("ConfigIndices") or data.get("config_indices")
 
         return cls(
-            hash=data["Hash"],
-            created_at=datetime.fromisoformat(created_at_str),
-            count=data["Count"],
-            all_verified=data["AllVerified"],
+            hash=hash_value,
+            created_at=created_at,
+            count=count,
+            all_verified=all_verified,
+            drivers=drivers,
+            config_indices=cfg_indices_val,
         )
 
 
@@ -56,14 +97,23 @@ class ReproIndexResponse:
     @classmethod
     def from_dict(cls, data: Dict[str, Any]) -> "ReproIndexResponse":
         lineages = {}
-        lineages_data = data.get("Lineages") or {}
+        # Support both legacy "Lineages" and potential future "lineages"
+        lineages_data = data.get("Lineages") or data.get("lineages") or {}
         for lineage_name, repros in lineages_data.items():
             lineages[lineage_name] = [LineageRepro.from_dict(r) for r in (repros or [])]
 
-        return cls(
-            bundle_id=data.get("BundleID", "00000000-0000-0000-0000-000000000000"),
-            lineages=lineages,
-        )
+        # Bundle ID:
+        # - legacy compat: top-level "BundleID"
+        # - current NG API: "Bundle" object with "id" field
+        bundle_id = data.get("BundleID")
+        if not bundle_id:
+            bundle = data.get("Bundle") or data.get("bundle")
+            if isinstance(bundle, dict):
+                bundle_id = bundle.get("id") or bundle.get("ID")
+        if not bundle_id:
+            bundle_id = data.get("bundle_id", "00000000-0000-0000-0000-000000000000")
+
+        return cls(bundle_id=bundle_id, lineages=lineages)
 
 
 @dataclass
@@ -78,14 +128,40 @@ class ReproMetadata:
 
     @classmethod
     def from_dict(cls, data: Dict[str, Any]) -> "ReproMetadata":
+        if "repros" in data and isinstance(data["repros"], dict):
+            inner = data["repros"]
+        else:
+            inner = data
+
+        # Core identifiers
+        hash_val = inner.get("hash") or inner.get("Hash")
+        if not hash_val:
+            raise KeyError("hash")
+
+        bundle_val = inner.get("bundle") or inner.get("Bundle")
+        lineage_val = inner.get("lineage") or inner.get("Lineage")
+        asset_val = inner.get("asset") or inner.get("Asset")
+
+        # Artifact hashes:
+        artifact_hashes_val: List[str]
+        raw_hashes = inner.get("artifact_hashes")
+        if raw_hashes:
+            artifact_hashes_val = list(raw_hashes)
+        else:
+            single_hash = inner.get("artifact_hash") or inner.get("ArtifactHash")
+            if single_hash:
+                artifact_hashes_val = [single_hash]
+            else:
+                artifact_hashes_val = []
+
         return cls(
-            hash=data["hash"],
-            bundle=data["bundle"],
-            lineage=data["lineage"],
-            asset=data.get("asset"),
-            artifact_hashes=data.get("artifact_hashes") or [],
-            summary=data.get("summary") or "",
-            flaky=data.get("flaky", False),
+            hash=hash_val,
+            bundle=str(bundle_val) if bundle_val is not None else "",
+            lineage=lineage_val or "",
+            asset=asset_val,
+            artifact_hashes=artifact_hashes_val,
+            summary=inner.get("summary") or "",
+            flaky=bool(inner.get("flaky", False)),
         )
 
 

src/test_suite/multiprocessing_utils.py

@@ -684,30 +684,24 @@ def download_and_process(source):
                     project=config.get_project(),
                 )
 
+        # Require at least one artifact hash. If none are present, there is
+        # nothing we can download or convert into fixtures...
         if not repro_metadata.artifact_hashes:
             return {
                 "success": False,
                 "repro": f"{section_name}/{crash_hash}",
                 "message": "Failed to process: no artifacts found",
             }
 
-        # Determine which artifacts to download
-        download_all = getattr(globals, "download_all_artifacts", False)
-        if download_all:
-            artifacts_to_download = repro_metadata.artifact_hashes
-            print(
-                f"  [{section_name}/{crash_hash[:8]}] Downloading all {len(artifacts_to_download)} artifact(s)",
-                file=sys.stderr,
-                flush=True,
-            )
-        else:
-            # Take only the first artifact hash
-            artifacts_to_download = [repro_metadata.artifact_hashes[0]]
-            print(
-                f"  [{section_name}/{crash_hash[:8]}] Using first artifact: {artifacts_to_download[0][:16]}",
-                file=sys.stderr,
-                flush=True,
-            )
+        # Determine which artifacts to download: Always process all available
+        # artifacts for this repro. Any duplicates will be naturally deduped
+        # later at the fixture level by hash.
+        artifacts_to_download = repro_metadata.artifact_hashes
+        print(
+            f"  [{section_name}/{crash_hash[:8]}] Downloading {len(artifacts_to_download)} artifact(s)",
+            file=sys.stderr,
+            flush=True,
+        )
 
         # Download and extract artifacts
         fix_count = 0
@@ -772,11 +766,7 @@ def download_and_process(source):
 
         # Always return success if we processed artifacts (even if no new fixtures extracted)
         # Not extracting new fixtures just means they already exist or artifacts don't contain .fix files
-        artifact_msg = (
-            f"{len(artifacts_to_download)} artifact(s)"
-            if len(artifacts_to_download) > 1
-            else "latest artifact"
-        )
+        artifact_msg = f"{len(artifacts_to_download)} artifact(s)"
         return {
             "success": True,
             "repro": f"{section_name}/{crash_hash}",