Sema: Don't coerce subexpression of 'try' to rvalue

The change in 8ab8b2e, was too
broad. We want to coerce the subexpression of `try!` to an rvalue,
but not the subexpression of a `try`.

If the subexpression of a `try` becomes an rvalue even though the
type of the parent expression is an lvalue, we can end up with
infinite recursion in coerceToType(), as demonstrated by the
test case.

Fixes #85034.

Author: slavapestov

lib/Sema/CSApply.cpp

@@ -3752,7 +3752,11 @@ namespace {
       return expr;
     }
 
-    Expr *visitAnyTryExpr(AnyTryExpr *expr) {
+    Expr *visitTryExpr(TryExpr *expr) {
+      return simplifyExprType(expr);
+    }
+
+    Expr *visitForceTryExpr(ForceTryExpr *expr) {
       auto *subExpr = expr->getSubExpr();
       auto type = simplifyType(cs.getType(subExpr));
 

validation-test/compiler_crashers_fixed/issue-85034.swift

@@ -0,0 +1,10 @@
+// RUN: %target-typecheck-verify-swift
+
+class C {
+  var x = 0
+}
+
+do {
+  let x = C()
+  let _ = (0, try x.x)  // expected-warning {{no calls to throwing functions occur within 'try' expression}}
+}