[Bug fortran/32088] [4.3 Regression] ICE (doesn't occur if given function standalone instead on internal)

["burnus at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

------- Comment #3 from burnus at gcc dot gnu dot org  2007-05-26 00:35 -------
Failure was introduced between 2007-04-04-r123491 and 2007-04-11-r123712

My guess would be that it is the following patch:
r123641 | pault | 2007-04-07 22:13:52 +0200 (Sa, 07 Apr 2007) | 14 lines

2007-04-07  Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/31293
        * symbol.c (gfc_check_function_type): New function. [...]

This function looks as follows:
+/* This function is called from parse.c(parse_progunit) to check the
+   type of the function is not implicitly typed in the host namespace
+   and to implicitly type the function result, if necessary.  */
+gfc_check_function_type (gfc_namespace *ns)
+  gfc_symbol *proc = ns->proc_name;
+  if (!proc->attr.contained || proc->result->attr.implicit_type)
+    return;
+  if (proc->result->ts.type == BT_UNKNOWN)

Here, proc->attr.contained = 1 and for some reason:
proc->result->attr.implicit_type = 0 (shouldn't this be 1? ts.type =
BT_UNKNOWN)

In any case, if I simply return in this function, the ICE is gone. The question
is now, why is proc->result->attr.implicit_type not set?


-- 

burnus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu dot
                   |                            |org, burnus at gcc dot gnu
                   |                            |dot org
           Keywords|                            |ice-on-valid-code
      Known to fail|                            |4.3.0
      Known to work|                            |4.2.0 4.1.3
   Target Milestone|---                         |4.3.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32088