[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-18 Thread jb at gcc dot gnu dot org
--- Comment #1 from jb at gcc dot gnu dot org 2007-05-18 20:07 --- Seems to work fine on i686-pc-linux-gnu. I would have guessed it to be something related to padding, but it's weird that it's not seen in 32-bit mode. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31933

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-19 Thread jvdelisle at gcc dot gnu dot org
--- Comment #2 from jvdelisle at gcc dot gnu dot org 2007-05-19 22:39 --- On x86-64 using -m32 the size being passed to unformatted_write is 12. Its aligning to nearest 32-bit boundary. Without -m32 size comes in at 16. Both are long double which has different size on 32 vs 64 bit ma

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-20 Thread jb at gcc dot gnu dot org
--- Comment #3 from jb at gcc dot gnu dot org 2007-05-20 09:46 --- I see. With that in mind, I think the patch for PR31915 that went in as r124741 is incorrect (and I share some of the blame since I approved it). We should use the size argument the frontend provides to the unformatted_(r

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-20 Thread jb at gcc dot gnu dot org
--- Comment #4 from jb at gcc dot gnu dot org 2007-05-20 09:50 --- (In reply to comment #3) > I see. With that in mind, I think the patch for PR31915 that went in as > r124741 > is incorrect (and I share some of the blame since I approved it). We should > use > the size argument the fr

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-20 Thread jvdelisle at gcc dot gnu dot org
--- Comment #5 from jvdelisle at gcc dot gnu dot org 2007-05-20 16:55 --- Actually, there are two separate build subdirectories created under the arch directory in the objdir. So on systems like mine, two separate libraries are built. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-21 Thread jvdelisle at gcc dot gnu dot org
--- Comment #6 from jvdelisle at gcc dot gnu dot org 2007-05-22 05:01 --- The following simplistic patch solves this problem. Sometimes I think we get overzealous with definitions and fancy configury. We all know the size is 10 bytes and 12 works. This works: Index: size_from_kind.c

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-21 Thread jb at gcc dot gnu dot org
--- Comment #7 from jb at gcc dot gnu dot org 2007-05-22 06:52 --- (In reply to comment #6) > The following simplistic patch solves this problem. Sometimes I think we get > overzealous with definitions and fancy configury. We all know the size is 10 > bytes and 12 works. No. The non-p

[Bug libfortran/31933] Uninitialized memory when writing real(10) as unformatted

2007-05-25 Thread jvdelisle at gcc dot gnu dot org
--- Comment #8 from jvdelisle at gcc dot gnu dot org 2007-05-25 17:43 --- Closing. Added the following comment to transfer.c: /* Master function for unformatted writes. NOTE: For kind=10 the size is 16 bytes on 64 bit machines. The unused bytes are not initialized and never us