s3-torture: Use static printer for smbd spooler test

diff --git a/source4/torture/rpc/spoolss.c b/source4/torture/rpc/spoolss.c
index a724aeeae16791e35eeb30bc304b5db0c41f884a..fd065fc6183a6a03884c3aa30a9a69b364ca802a 100644 (file)
--- a/source4/torture/rpc/spoolss.c
+++ b/source4/torture/rpc/spoolss.c
@@ -54,6 +54,7 @@
 #define TORTURE_DRIVER_TIMESTAMPS      "torture_driver_timestamps"
 #define TORTURE_DRIVER_DELETER         "torture_driver_deleter"
 #define TORTURE_DRIVER_DELETERIN       "torture_driver_deleterin"
+#define TORTURE_PRINTER_STATIC1                "print1"
 
 #define TOP_LEVEL_PRINT_KEY "SOFTWARE\\Microsoft\\Windows NT\\CurrentVersion\\Print"
 #define TOP_LEVEL_PRINT_PRINTERS_KEY TOP_LEVEL_PRINT_KEY "\\Printers"
@@ -7747,7 +7748,14 @@ static bool test_print_test_smbd(struct torture_context *tctx,
        struct cli_credentials *credentials = cmdline_credentials;
        struct smbcli_options options;
        TALLOC_CTX *mem_ctx = talloc_new(tctx);
-       const char *share = t->info2.printername;
+       /*
+        * Do not test against the dynamically added printers, printing via
+        * smbd means that a different spoolss process may handle the
+        * OpenPrinter request to the one that handled the AddPrinter request.
+        * This currently leads to an ugly race condition where one process
+        * sees the new printer and one doesn't.
+        */
+       const char *share = TORTURE_PRINTER_STATIC1;
 
        torture_comment(tctx, "Testing smbd job spooling\n");
        lpcfg_smbcli_options(tctx->lp_ctx, &options);