Anonymous struct ambiguity detection (PR c/4784)

[Joseph S. Myers]

On Sun, 9 May 2010, H.J. Lu wrote:

> On Sun, May 9, 2010 at 1:38 PM, Joseph S. Myers <joseph@codesourcery.com> wrote:
> > On Sun, 9 May 2010, H.J. Lu wrote:
> >
> >> This test failed on Linux/ia32.
> >
> > That's not a useful report.  How exactly did the diagnostics differ from
> > those the testcase expects?
> >
> 
> See:
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44051

Thanks.  I've applied this patch.  The test for a large structure was 
generating the identifier i386 among the identifiers i000 through i999.

Index: gcc.dg/anon-struct-9.c
===================================================================
--- gcc.dg/anon-struct-9.c	(revision 159206)
+++ gcc.dg/anon-struct-9.c	(working copy)
@@ -120,8 +120,8 @@
   S10(f)
   S10(g)
   S10(h)
-  S10(i)
   S10(j)
+  S10(k)
   struct
   {
     int a123; /* { dg-error "duplicate member" } */
Index: ChangeLog
===================================================================
--- ChangeLog	(revision 159206)
+++ ChangeLog	(working copy)
@@ -1,5 +1,10 @@
 2010-05-09  Joseph Myers  <joseph@codesourcery.com>
 
+	PR c/44051
+	* gcc.dg/anon-struct-9.c: Avoid generating identifier i386.
+
+2010-05-09  Joseph Myers  <joseph@codesourcery.com>
+
 	PR c/10676
 	* gcc.dg/anon-struct-10.c: New test.
 

-- 
Joseph S. Myers
joseph@codesourcery.com