This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Build/bootstrap failure on x86_64-linux


On Tuesday 20 July 2004 07:00, Mark Mitchell wrote:
> Andrew Pinski wrote:
> >This definetly looks like the wrong patch.
> >This is against EDG :).
>
> Indeed.  I hate it when my mail client isn't on the same machine as the
> file I'm trying to attach.
>
> Here it is again.

Your patch is Just Fine, but your first checkin was not (and you
did not notice when you reverted it, you must have something else
on your mind).

Look at http://gcc.gnu.org/ml/gcc-cvs/2004-07/msg01006.html:
The ChangeLog has entries for cp-tree.h, class.c, method.c, search.c.
The checking has paches for cp-tree.h, class.c, method.c!
The patch you sent me _does_ have a search.c hunk.

In fact, the diff between the patch you posted and the patch you
commited is:

+Index: search.c
+===================================================================
+RCS file: /cvs/gcc/gcc/gcc/cp/search.c,v
+retrieving revision 1.303
+diff -c -5 -p -r1.303 search.c
+*** search.c   19 Jul 2004 15:45:53 -0000      1.303
+--- search.c   20 Jul 2004 04:51:56 -0000
+*************** lookup_fnfields_1 (tree type, tree name)
+*** 1377,1388 ****
+           lazily_declare_fn (sfk_constructor, type);
+         if (CLASSTYPE_LAZY_COPY_CTOR (type))
+           lazily_declare_fn (sfk_copy_constructor, type);
+       }
+        else if (name == ansi_assopname(NOP_EXPR)
+!             && !TYPE_HAS_ASSIGN_REF (type)
+!             && !TYPE_FOR_JAVA (type))
+       lazily_declare_fn (sfk_assignment_operator, type);
+      }
+
+    method_vec = CLASSTYPE_METHOD_VEC (type);
+    if (!method_vec)
+--- 1377,1387 ----
+           lazily_declare_fn (sfk_constructor, type);
+         if (CLASSTYPE_LAZY_COPY_CTOR (type))
+           lazily_declare_fn (sfk_copy_constructor, type);
+       }
+        else if (name == ansi_assopname(NOP_EXPR)
+!             && CLASSTYPE_LAZY_ASSIGNMENT_OP (type))
+       lazily_declare_fn (sfk_assignment_operator, type);
+      }
+
+    method_vec = CLASSTYPE_METHOD_VEC (type);
+    if (!method_vec)

So it appears you just forgot to check in the search.c part of your
patch.  With the diffs2 patch you posted, the bootstrap works.

Gr.
Steven



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]