[Bug fortran/60993] New: Trouble initializing double precision variable using boz literals

[fkrogh#gcc at mathalacarte dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60993

            Bug ID: 60993
           Summary: Trouble initializing double precision variable using
                    boz literals
           Product: gcc
           Version: 4.8.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: fkrogh#gcc at mathalacarte dot com

Created attachment 32701
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32701&action=edit
This code illustrates the problem above

The attached file gets this error during the compile

inf.f90:21:0: internal compiler error: in gfc_conv_array_initializer, at
fortran/trans-array.c:5446
    dinfn = transfer(infn, dinfn) ! Comment this to see failure above.

If you comment out the statement causing this problem you can see that a
transfer with a parameter array to a double works, but the same fails when
using an array literal.

I am just interested in getting plus and minus infinity set in parameters or
variables to use in my code.  Their use really simplifies some code.  I'd like
them defined for single, double, and quad precisions.