[Russ Rew: netCDF bug in NcAtt::is_valid()]

Forgot to CC: support-netcdf when I sent this netCDF C++ interface bug fix
out ...

--Russ
------- Forwarded Message

To: address@hidden (Tom LeFebvre)
Subject: netCDF bug in NcAtt::is_valid()
Reply-to: address@hidden
Organization: UCAR Unidata Program
Date: Tue, 22 Feb 1994 11:04:50 -0700
From: Russ Rew <russ@buddy>

Hi Tom,

I just discovered a bug and a fix for it in netcdf.cc.  The problem is that
the is_valid() member function returns 0 for valid global attributes.  The
fix is a one-liner, for which I've appended a patch.

I still haven't gotten around to fixing the memory leak you reported that
Purify found, but it's on my list.

I intend to put out a patch soon with all the accumulated bug fixes, unless
I get really ambitious and make a release instead.

- --Russ

===================================================================
RCS file: RCS/netcdf.cc,v
retrieving revision 1.45
diff -c -r1.45 netcdf.cc
*** /tmp/T0a006ap       Tue Feb 22 10:58:41 1994
- --- netcdf.cc Tue Feb 22 10:58:07 1994
***************
*** 856,862 ****
  NcBool NcAtt::is_valid( void ) const
  {
      return the_file->is_valid() &&
!       the_variable->is_valid() &&
        ncattinq(the_file->id(), the_variable->id(), the_name, 0, 0) != ncBad;
  }
  
- --- 856,862 ----
  NcBool NcAtt::is_valid( void ) const
  {
      return the_file->is_valid() &&
!       (the_variable->id() == NC_GLOBAL || the_variable->is_valid()) &&
        ncattinq(the_file->id(), the_variable->id(), the_name, 0, 0) != ncBad;
  }

------- End of Forwarded Message