[FYI] Fix regression in variant part handling

[FYI] Fix regression in variant part handling

[Tom Tromey]

My series to change DWARF attribute handling to be type-safe
introduced a regression in gdb.ada/variant.exp.  handle_variant was
using as_unsigned on an attribute with DW_FORM_sdata.  This patch
changes it to use constant_value instead.

gdb/ChangeLog
2020-09-30  Tom Tromey  <tromey@adacore.com>

	* dwarf2/read.c (handle_variant): Use constant_value.
---
 gdb/ChangeLog     | 4 ++++
 gdb/dwarf2/read.c | 4 ++--
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index bcac94ed36a..c0a89ecbe1e 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -16178,7 +16178,7 @@ handle_variant (struct die_info *die, struct type *type,
   /* In a variant we want to get the discriminant and also add a
      field for our sole member child.  */
   struct attribute *discr = dwarf2_attr (die, DW_AT_discr_value, cu);
-  if (discr == nullptr)
+  if (discr == nullptr || !discr->form_is_constant ())
     {
       discr = dwarf2_attr (die, DW_AT_discr_list, cu);
       if (discr == nullptr || discr->as_block ()->size == 0)
@@ -16187,7 +16187,7 @@ handle_variant (struct die_info *die, struct type *type,
 	variant.discr_list_data = discr->as_block ();
     }
   else
-    variant.discriminant_value = discr->as_unsigned ();
+    variant.discriminant_value = discr->constant_value (0);
 
   for (die_info *variant_child = die->child;
        variant_child != NULL;
-- 
2.26.2