[commit/Ada] Better handle new array pointer encoding

[Joel Brobecker <brobecker at adacore dot com>]

Hello,

Assuming we have a variable whose type is a pointer to a string
(in Ada we use "access" types for pointer to a given type):

   type String_Access is access String;
   String_P : String_Access := new String'("Hello");

If you print the value of that variable, GDB currently prints
the wrong type name:

    (gdb) p string_p
    $1 = (string) 0x8060010

The actual encoding is described in exp_dbug.ads:

   --  For each pointer to this unconstrained array type, the compiler will
   --  generate a typedef that points to the above "arr___XUP" fat-pointer
   --  type. As a consequence, when it comes to fat-pointer types:

   --    1. The type name is given by the typedef

   --    2. If the debugger is asked to output the type, the appropriate
   --       form is "access arr", except if the type name is "arr___XUP"
   --       for which it is the array definition.

The attached patch implements this suggestion.

2008-01-03  Joel Brobecker  <brobecker@adacore.com>

        * ada-lang.c (static_unwrap_type): Add forward declaration.
        (template_to_static_fixed_type): Fields of dynamic types sometimes
        also need to be unwrapped. Take this into account.
        (ada_to_fixed_type_1): Renamed from ada_to_fixed_type.
        (ada_to_fixed_type): New wrapper around ada_to_fixed_type_1.
        * ada-typeprint.c (ada_print_type): Get the typename from
        the original type, not the base type. 

Also attached is a small testcase that verifies the fix:

2008-01-03  Joel Brobecker  <brobecker@adacore.com>

        * gdb.ada/arrayptr/pck.ads, gdb.ada/arrayptr/pck.adb,
        gdb.ada/arrayptr/foo.adb: New files.
        * gdb.ada/arrayptr.exp: New testcase.

All tested on x86-linux, no regression.
Checked in.

-- 
Joel

Attachment: arrayptr.diff
Description: Text document

Attachment: arrayptr-tc.diff
Description: Text document