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] byte-wise access to masks

From: Brooks Moses <brooks dot moses at codesourcery dot com>
To: Thomas Koenig <tkoenig at alice-dsl dot net>
Cc: fortran at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
Date: Sun, 12 Aug 2007 20:21:53 -0700
Subject: Re: [patch, fortran] byte-wise access to masks
References: <1186866739.3442.7.camel@meiner.onlinehome.de> <46BFC50E.60501@codesourcery.com>

Brooks Moses wrote:

Thomas Koenig wrote:
Regression-tested on i686-pc-linux-gnu.  This absolutely needs
to be tested on a big-endian system before committing.  OK
once this has been done?
Here are the key results from "make check-fortran" on powerpc-apple-darwin8. The edit_real_1 and large_real_kind* failures are known ones on that platform. The zero_sized_1 failure, however, appears to be due to your patch; I tried again with your patch reverted, and it goes away.

Specifically, the following reduced testcase gives a "Bus error" on execution with your patch, but works fine without it:

program test
  integer :: tempn(1,5)
  tempn = 2
  if (size(pack(tempn(:,-4:-5),tempn(:,-4:-5)/=0, &
    (/1,3,4,5,1,0,7,9/))) /= 8) &
    call abort
end

- Brooks

References:
- [patch, fortran] byte-wise access to masks
  - From: Thomas Koenig
- Re: [patch, fortran] byte-wise access to masks
  - From: Brooks Moses

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