This is the mail archive of the gcc-patches@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]

Re: Aliasing patches


On Thu, Jun 01, 2000 at 12:38:53AM -0700, Richard Henderson wrote:
> I count twenty additional failures, not counting duplicates.  Some
> of the C++ ones are weird, and probably indicate miscompilation in
> the compiler itself.

This fixes all of the C regressions.  My build cycle started before
Mark checked in the one C++ fe change, so I don't know if all those
are fixed yet.  There is still one outstanding Fortran regression
at -O2.


r~


        * c-decl.c (init_decl_processing): Set lang_get_alias_set first thing.
        * cp/decl.c (init_decl_processing): Likewise.

Index: c-decl.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/c-decl.c,v
retrieving revision 1.114
diff -c -p -d -r1.114 c-decl.c
*** c-decl.c	2000/05/31 20:01:57	1.114
--- c-decl.c	2000/06/01 08:21:52
*************** init_decl_processing ()
*** 2896,2901 ****
--- 2896,2903 ----
    tree array_domain_type;
    tree t;
  
+   lang_get_alias_set = c_get_alias_set;
+ 
    current_function_decl = NULL;
    named_labels = NULL;
    current_binding_level = NULL_BINDING_LEVEL;
*************** init_decl_processing ()
*** 3136,3143 ****
    init_iterators ();
  
    incomplete_decl_finalize_hook = finish_incomplete_decl;
- 
-   lang_get_alias_set = c_get_alias_set;
  
    /* Record our roots.  */
  
--- 3138,3143 ----
Index: cp/decl.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/cp/decl.c,v
retrieving revision 1.620
diff -c -p -d -r1.620 decl.c
*** cp/decl.c	2000/05/31 19:27:11	1.620
--- cp/decl.c	2000/06/01 08:21:57
*************** init_decl_processing ()
*** 6251,6256 ****
--- 6251,6258 ----
    int wchar_type_size;
    tree array_domain_type;
  
+   lang_get_alias_set = c_get_alias_set;
+ 
    /* Check to see that the user did not specify an invalid combination
       of command-line options.  */
    if (flag_new_abi && !flag_vtable_thunks)
*************** init_decl_processing ()
*** 6640,6646 ****
    using_eh_for_cleanups ();
  
    print_error_function = lang_print_error_function;
-   lang_get_alias_set = &c_get_alias_set;
    valid_lang_attribute = cp_valid_lang_attribute;
  
    /* Maintain consistency.  Perhaps we should just complain if they
--- 6642,6647 ----

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