RFC: fix PR 7643

[Tom Tromey <tromey@redhat.com>]

This fixes PR exp/7643.

The bug is:

    (gdb) ptype &*"foo"
    type = char [4]

... where the correct answer is "char *".

The fix is to add a call coerce_array in the optimized UNOP_IND case in
evaluate_subexp_for_address.

I plan to check this in, but before doing so I thought I would ask to
see whether this messes anything up for non-C languages (specifically
Ada, I suppose).

I built and regtested this on x86-64 (compile farm).  So, if this is in
fact wrong for some other language, a test case would be handy.

Tom

2010-01-25  Tom Tromey  <tromey@redhat.com>

	PR exp/7643:
	* eval.c (evaluate_subexp_for_address) <UNOP_IND>: Call
	coerce_array on result.

2010-01-25  Tom Tromey  <tromey@redhat.com>

	PR exp/7643:
	* gdb.base/printcmds.exp (test_print_string_constants): Remove
	setup_kfail.

Index: eval.c
===================================================================
RCS file: /cvs/src/src/gdb/eval.c,v
retrieving revision 1.125
diff -u -r1.125 eval.c
--- eval.c	18 Jan 2010 20:54:33 -0000	1.125
+++ eval.c	25 Jan 2010 19:53:35 -0000
@@ -2629,7 +2629,7 @@
 	  goto default_case_after_eval;
 	}
 
-      return x;
+      return coerce_array (x);
 
     case UNOP_MEMVAL:
       (*pos) += 3;
Index: testsuite/gdb.base/printcmds.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/printcmds.exp,v
retrieving revision 1.28
diff -u -r1.28 printcmds.exp
--- testsuite/gdb.base/printcmds.exp	25 Jan 2010 19:31:24 -0000	1.28
+++ testsuite/gdb.base/printcmds.exp	25 Jan 2010 19:53:37 -0000
@@ -647,7 +647,6 @@
     gdb_test "p *\"foo\""		" = 102 'f'"
     gdb_test "ptype *\"foo\""		" = char"
     gdb_test "p &*\"foo\""		" = \"foo\""
-    setup_kfail "gdb/538" *-*-*
     gdb_test "ptype &*\"foo\""	"type = char \\*"
     gdb_test "p (char *)\"foo\""	" = \"foo\""
 }