gateway: Cache rinfo for main_thread_only

Fix a bug triggered by pytest-cov when it tries to call _rinfo
after the gateway is already busy with a remote_exec call:

pytest-dev/pytest-xdist#1070 (comment)

Author: zmedico

src/execnet/gateway.py

@@ -31,6 +31,11 @@ def __init__(self, io: IO, spec: XSpec) -> None:
         super().__init__(io=io, id=spec.id, _startcount=1)
         self.spec = spec
         self._initreceive()
+        if self.execmodel.backend == "main_thread_only":
+            # This will cause a deadlock if there is already a running
+            # remote_exec, so cache the result before there are any
+            # other remote_exec calls.
+            self._rinfo()
 
     @property
     def remoteaddress(self) -> str:

testing/test_gateway.py

@@ -645,3 +645,35 @@ def test_main_thread_only_concurrent_remote_exec_deadlock(
             ch.close()
         gw.exit()
         gw.join()
+
+
+def test_concurrent_rinfo(
+    execmodel: gateway_base.ExecModel, makegateway: Callable[[str], Gateway]
+) -> None:
+    gw = makegateway(f"execmodel={execmodel.backend}//popen")
+
+    try:
+        ch = gw.remote_exec(
+            """
+                import os, time
+                time.sleep(1.5)
+                channel.send(os.getpid())
+        """
+        )
+        # For main_thread_only _rinfo will deadlock unless the gateway
+        # has cached the result earlier, but only if the above sleep
+        # exceeds the 1 second grace period in the WorkerGateway
+        # _local_schedulexec method.
+        if execmodel.backend == "main_thread_only":
+            assert hasattr(gw, "_cache_rinfo")
+        rinfo = gw._rinfo()
+        try:
+            res = ch.receive()
+        finally:
+            ch.close()
+            ch.waitclose()
+
+        assert res == rinfo.pid
+    finally:
+        gw.exit()
+        gw.join()