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]
Other format: [Raw text]

Re: [patch, fortran] PR31201 Too large unit number generates wrong code


No, I do not disagree with you.  I was very bothered about changing the size of
unit.  I would appreciate your assistance on the front end.  I got as far as
determining if the value provided is int8.  Without the patch here, the
front-end is converting that to int4.  I will study your repeat patch and we can
email off list to work this.

Fine with me.


Also, take a look at what happens to exe_path in main.c.  Somewhere in there we
get unfreed memory.

Hum, I wrote that part and remember thinking: we never free exe_path, but then it's only allocated once. That was before everyone started to run valgrind :) Does the following fix it for you?

Index: runtime/main.c
===================================================================
--- runtime/main.c      (revision 123388)
+++ runtime/main.c      (working copy)
@@ -170,4 +170,5 @@
cleanup (void)
{
  close_units ();
+  free (exe_path);
}


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