[Bug fortran/43605] [4.4/4.5 Regression] wrong results with ftell

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

------- Comment #2 from burnus at gcc dot gnu dot org  2010-03-31 15:07 -------
Janne, could you please have a look.

  * * *

The following looks very fishy:

  gfc_unit * u = find_unit (*unit);
  ret = (size_t) stell (u->s);


pt find_unit(*unit)->s
type = struct stream {
    ssize_t (*read)(struct stream *, void *, ssize_t);
    ssize_t (*write)(struct stream *, const void *, ssize_t);
    gfc_offset (*seek)(struct stream *, gfc_offset, int);
    gfc_offset (*tell)(struct stream *);
    int (*trunc)(struct stream *, gfc_offset);
    int (*flush)(struct stream *);
    int (*close)(struct stream *);
}

stell (stream * s)
{
  return s->tell (s);

 ****************

And in unix.c:

typedef struct
{
  stream st;
  (...)
  int fd;                       /* The POSIX file descriptor.  */


raw_tell (unix_stream * s)
{
  return lseek (s->fd, 0, SEEK_CUR);

Which only works because of eval casting:
buf_init (unix_stream * s)
  s->st.tell = (void *) buf_tell;

 * * *

I have no idea whether the pointer is initialized correctly, but I can imagine
several ways how it could break (e.g. also with LTO) and it makes debugging a
mess!


-- 

burnus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jb at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43605