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

[Bug libfortran/34540] cshift, eoshift, kind=1 and kind=2 arguments...



------- Comment #16 from jvdelisle at gcc dot gnu dot org  2007-12-23 16:43 -------
The test case in comment #9 passes if I revert the code path required by the
original test case in pr33317.  Using gdb, I see no difference in the incoming
dim expression for either case.  This implies we need to approach this problem
differently.

The root of the problem is saving on the number of variations of the runtime
cshift and eoshift functions, gfortran converts the dim expression to a
convert_* function with dim as an argument.  The dim argument is optional and
when it is not present a null is provided.  The convert functions do not handle
a null and gfortran ends up segfaulting on memory access at run time.

Attempting to resolve the dim argument also results in a convert function
rather than a NULL constant.

I am unassigning myself for a while.


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |33317
         AssignedTo|jvdelisle at gcc dot gnu dot|unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34540


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