Update test expectations to match reality.

tests/language/language.status

Index: tests/language/language.status
diff --git a/tests/language/language.status b/tests/language/language.status
index c625f80837671216e45abb49b5724d3565362aa1..facf8c5e530f2b05089ae41e3ec159dda9862592 100644
--- a/tests/language/language.status
+++ b/tests/language/language.status
@@ -205,8 +205,12 @@ GCTest: Skip # Issue 1487
 
 
 [ $component == frogsh ]
-IntrinsifiedMethodsTest: Fail # Issue 1470
 ArithmeticTest: Fail # Issue 1470
+AssignStaticTypeTest/01: Fail # Unrecognized option: --no-enable-type-checks.
+AssignStaticTypeTest/02: Fail # Unrecognized option: --no-enable-type-checks.
+AssignStaticTypeTest/03: Fail # Unrecognized option: --no-enable-type-checks.
+AssignStaticTypeTest/04: Fail # Unrecognized option: --no-enable-type-checks.

[ngeoffray, 2012/03/09 08:47:22]

Don't they also fail on frog?

[kasperl, 2012/03/09 08:53:27]

I don't think so. I think on frog the parameters are passed in a different order
so the problematic option comes after the --ignore-unrecognized-options flag.

+IntrinsifiedMethodsTest: Fail # Issue 1470
 
 # This set of four test components (frog, frogsh, frogium, and webdriver) all
 # consume frog output, so if frog is generating the incorrect code, chances are