ci: run the launcher-only test (accept launcher cache-key worker flags)

We run the launcher transport, so opt into VGI_REQUIRE_LAUNCHER_TRANSPORT=1 to
run launcher/options_smoke.test (as `make test_launcher` does). That test
appends extra argv (--describe, --threaded, ...) to the launch: LOCATION to
vary the launcher cache key; the example worker rejected unknown args and
exited, so the ATTACH produced 0 schemas. Accept the vgi-python fixture's
launcher/log flags (--describe/--no-describe/--threaded/--quiet/-q/--debug/
--log-level) as no-ops in Main.runWorker. options_smoke.test now passes;
suite is 172 pass / 12 skip.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Author: rustyconover

CLAUDE.md

@@ -265,9 +265,16 @@ without a C++ build:
   `ci/run-integration.sh` stages the preprocessed tree, sets the four
   `VGI_*_WORKER` vars (`launch:` + the three `ci/wrappers/` catalog wrappers —
   the committed, path-relative replacements for the old `/tmp/vgi-worker-*`),
-  warms the extension cache once, and tallies pass/skip/fail. Green on Linux CI
-  (**171 pass / 13 skip**) and macOS (**172/13** — one more because the
-  `launch:` osx run includes a test the Linux exclusion drops). **Excludes**
+  warms the extension cache once, then runs the whole suite in a **single
+  unittest invocation** (like `make test_launcher`) so the CI log streams the
+  native sqllogictest report (per-test progress + `All tests passed (… N
+  assertions in M test cases)` — ~8700 assertions across the ~185 files), not a
+  rolled-up count. Green on both lanes: **172 pass / 12 skip**. Sets
+  `VGI_REQUIRE_LAUNCHER_TRANSPORT=1` (we *are* the launcher transport) so
+  `launcher/options_smoke.test` runs — which required `Main.runWorker` to accept
+  the launcher cache-key / fixture flags (`--describe`/`--no-describe`/
+  `--threaded`/`--quiet`/`--debug`/`--log-level`) as no-ops instead of exiting
+  on unknown argv. **Excludes**
   `bool_in_union.test` in addition to `nested_type_combinations` — CI surfaced
   that it characterizes a **pre-existing platform-dependent union-bool bug**:
   the worker reads uninitialized memory for boolean union variants after row 1

ci/run-integration.sh

@@ -61,6 +61,9 @@ case "$TRANSPORT" in
     export VGI_VERSIONED_WORKER="launch:${HERE}/wrappers/vgi-worker-versioned"
     export VGI_VERSIONED_TABLES_WORKER="launch:${HERE}/wrappers/vgi-worker-versioned-tables"
     export VGI_ATTACH_OPTIONS_WORKER="launch:${HERE}/wrappers/vgi-worker-attach-options"
+    # We are the launcher transport, so opt into the launcher-only tests
+    # (launcher/options_smoke.test) — matches vgi's `make test_launcher`.
+    export VGI_REQUIRE_LAUNCHER_TRANSPORT=1
     ;;
   http)
     # Boot the example worker as an HTTP server on an ephemeral port; it prints

vgi-example-worker/src/main/java/farm/query/vgi/example/Main.java

@@ -1049,6 +1049,14 @@ private static void runWorker(Worker w, String[] args) {
                 case "--unix" -> unixSocket = args[++i];
                 case "--idle-timeout" ->
                         idleTimeoutMs = (long) (Double.parseDouble(args[++i]) * 1000.0);
+                // Launcher cache-key / fixture-parity flags. The vgi-python
+                // fixture worker implements these (quiet/debug logging,
+                // description pages, threading); here they only need to be
+                // accepted, so a launch: LOCATION that appends them to vary the
+                // launcher cache key (launcher/options_smoke.test) starts the
+                // worker instead of being rejected.
+                case "--describe", "--no-describe", "--threaded", "--quiet", "-q", "--debug" -> { }
+                case "--log-level" -> i++; // consumes its value
                 default -> { System.err.println("unknown arg: " + args[i]); System.exit(2); }
             }
         }