[patch] Overload resolution test case

[sami wagiaalla <swagiaal at redhat dot com>]

After consulting the C++ spec, I have created a test case for untested 
and unimplemented overload resolution scenarios. The test case has been 
marked with the appropriate kfails.

Test case for unimplemented overload cases.

2010-10-05  Sami Wagiaalla  <swagiaal@redhat.com>

	* gdb.cp/oranking.exp: New test.
	* gdb.cp/oranking.cc: New test program.

diff --git a/gdb/testsuite/gdb.cp/oranking.cc b/gdb/testsuite/gdb.cp/oranking.cc
new file mode 100644
index 0000000..69e0bed
--- /dev/null
+++ b/gdb/testsuite/gdb.cp/oranking.cc
@@ -0,0 +1,203 @@
+
+/* 1. A standard covnersion sequence is better than a user-defined sequence
+      which is better than an elipses conversion sequence.  */
+
+class A{};
+class B: public A {public: operator int (){ return 1;}};
+
+// standard vs user-defined
+int foo0  (int) { return 10; }
+int foo1  (int) { return 11; } // B -> int   : user defined
+int foo1  (A)   { return 12; } // B -> A     : standard
+int test1 () {
+  B b;
+  return foo1(b); // 12
+}
+
+// user-defined vs ellipsis
+int foo2 (int) { return 13;} // B -> int   : user defined
+int foo2 (...) { return 14;} // B -> ...   : ellipsis
+int test2(){
+  B b;
+  return foo2(b); // 13
+}
+
+/* 2. Standard Conversion squence S1 is better than standard Conversion
+      S2 if:  */
+
+//      - S1 has a better rank than S2
+//        see overload.exp for more comprehensive testing of this.
+int foo3 (double) { return 21; } // float->double is 'promotion rank'
+int foo3 (int)    { return 22; } // float->int is 'conversion rank'
+int test3(){
+  return foo3 (1.0f); // 21
+}
+
+//      - S1 and S2 are both 'qualification conversions' but S1 cv-qualification
+//        is a subset of S2 cv-qualification.
+int foo4 (const volatile int*) { return 23; }
+int foo4 (      volatile int*) { return 24; }
+int test4 () {
+  volatile int a = 5;
+  return foo4(&a); // 24
+}
+
+//      - S1 and S2 have the same rank but:
+//        - S2 is a conversion of pointer or memeber-pointer to bool
+int foo5 (bool)  { return 25; }
+int foo5 (void*) { return 26; }
+int test5 () {
+  char *a;
+  return foo5(a); // 26
+}
+
+//        - Class B publicly extends class A and S1 is a conversion of
+//          B* to A* and S2 is a conversion B* to void*
+int foo6 (void*) { return 27; }
+int foo6 (A*)    { return 28; }
+int test6 () {
+  B *bp;
+  return foo6(bp); // 28
+}
+
+//        - Class C publicly extends Class B which publicly extends
+//          class A and S1 is a conversion of C* to B* and S2 is a
+//          conversion C* to A*.
+class C: public B {};
+int foo7 (A*)    { return 29; }
+int foo7 (B*) { return 210; }
+int test7 () {
+  C *cp;
+  return foo7(cp); // 210
+}
+
+//        - Same as above but for references.
+int foo8 (A&) { return 211; }
+int foo8 (B&) { return 212; }
+int test8 () {
+  C c;
+  return foo8(c); // 212
+}
+
+//        - Same as above but passing by copy.
+int foo9 (A) { return 213; }
+int foo9 (B) { return 214; }
+int test9 () {
+  C c;
+  return foo9(c); // 212
+}
+
+//        - S1 is a conversion of A::* to B::* and S2 is a conversion of
+//          A::* to C::8.
+int foo10 (void (C::*)()) { return 215; }
+int foo10 (void (B::*)()) { return 216; }
+int test10 () {
+  void (A::*amp)();
+  return foo10(amp); // 216
+}
+
+////      - S1 is a subsequence of S2
+// FIXME: example needed.
+//int foo5 (const int*) { return 23; }
+//int foo5 (      int*) { return 24; }
+//int test5 () {
+//  const char a = 5;
+//  return foo5(&a); // 24
+
+/* 3. User defined conversion U1 is better than user defined Conversion U2,
+      if U1 and U2 are using the same conversion function but U1 has a better
+      second standard conversion sequence than U2.  */
+class D {public: operator short(){ return 0;}};
+int foo11 (float) { return 31; }
+int foo11 (int)   { return 32; }
+int test11 () {
+  D d;
+  return foo11(d); // 32
+}
+
+/* 4. Function Level Ranking.
+      All else being equal some functions are preferred by overload resolution.
+      Function F1 is better than function F2 if: */
+//    - F1 is a non-template function and F2 is a template function
+template<class T> int foo12(T)   { return 41; }
+                  int foo12(int) { return 42; }
+int test12 (){
+  return foo12(1); //42
+}
+
+//    - F1 is a more specialized template instance
+template<class T> int foo13(T)  { return 43; }
+template<class T> int foo13(T*) { return 44; }
+int test13 (){
+  char *c;
+  return foo13(c); // 44
+}
+
+//    - The context is user defined conversion and F1 has
+//      a better return type than F2
+class E {
+  public:
+    operator double () {return 45; }
+    operator int    () {return 46; }
+};
+int foo14 (int a) {return a;}
+int test14 (){
+  E e;
+  return foo14(e); // 46
+}
+
+int main() {
+  B b;
+  foo0(b);
+  foo1(b);
+  test1();
+
+  foo2(b);
+  test2();
+
+  foo3(1.0f);
+  test3();
+
+  volatile int a;
+  foo4(&a);
+  test4();
+
+  char *c;
+  foo5(c);
+  test5();
+
+  B *bp;
+  foo6(bp);
+  test6();
+
+  C *cp;
+  foo7(cp);
+  test7();
+
+  C co;
+  foo8(co);
+  test8();
+
+  foo9(co);
+  test9();
+
+  void (A::*amp)();
+  foo10(amp);
+  test10();
+
+  D d;
+  foo11(d);
+  test11();
+
+  foo12(1);
+  test12();
+
+  foo13(c);
+  test13();
+
+  E e;
+  foo14(e);
+  test14();
+
+  return 0; // end of main
+}
diff --git a/gdb/testsuite/gdb.cp/oranking.exp b/gdb/testsuite/gdb.cp/oranking.exp
new file mode 100644
index 0000000..f5cfbb7
--- /dev/null
+++ b/gdb/testsuite/gdb.cp/oranking.exp
@@ -0,0 +1,96 @@
+# Copyright 2008 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+set testfile oranking
+set srcfile ${testfile}.cc
+if { [prepare_for_testing ${testfile}.exp ${testfile} ${srcfile} {debug c++}] } {
+     return -1
+}
+
+############################################
+
+if ![runto_main] then {
+    perror "couldn't run to breakpoint main"
+    continue
+}
+
+gdb_breakpoint [gdb_get_line_number "end of main"]
+gdb_continue_to_breakpoint "end of main"
+
+# The 'test*' functions are to prove our understanding
+# of the overload resolution performed by the compiler
+# So, they should always pass, and the returned value
+# should match the corresponding call to 'foo*'
+
+setup_kfail "gdb/12096" *-*-*
+gdb_test "p foo0(b)" "10"
+
+gdb_test "p test1()" "12"
+gdb_test "p foo1(b)" "12"
+
+gdb_test "p test2()" "13"
+setup_kfail "gdb/12098" *-*-*
+gdb_test "p foo2(b)" "13"
+
+gdb_test "p test3()"    "21"
+gdb_test "p foo3(1.0f)" "21"
+
+gdb_test "p test4()"  "24"
+setup_kfail "gdb/12098" *-*-*
+gdb_test "p foo4(&a)" "24"
+
+gdb_test "p test5()" "26"
+setup_kfail "gdb/12098" *-*-*
+gdb_test "p foo5(c)" "26"
+
+gdb_test "p test6()"  "28"
+setup_kfail "gdb/10343" *-*-*
+gdb_test "p foo6(bp)" "28"
+
+gdb_test "p test7()"  "210"
+setup_kfail "gdb/10343" *-*-*
+gdb_test "p foo7(cp)" "210"
+
+gdb_test "p test8()"  "212"
+setup_kfail "gdb/10343" *-*-*
+gdb_test "p foo8(co)" "212"
+
+gdb_test "p test9()"  "214"
+setup_kfail "gdb/12098" *-*-*
+gdb_test "p foo9(co)" "214"
+
+gdb_test "p test10()"   "216"
+setup_kfail "gdb/12098" *-*-*
+gdb_test "p foo10(amp)" "216"
+
+gdb_test "p test11()"   "32"
+setup_kfail "gdb/12096" *-*-*
+gdb_test "p foo11(d)" "32"
+
+gdb_test "p test12()" "42"
+# this passes only because gdb does not yet
+# implement template function calling
+gdb_test "p foo12(1)" "42"
+
+gdb_test "p test13()" "44"
+setup_kfail "gdb/12098" *-*-*
+gdb_test "p foo13(c)" "44"
+
+gdb_test "p test14()" "46"
+setup_kfail "gdb/12096" *-*-*
+gdb_test "p foo14(e)" "46"
+
+
+