This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
RE: Porting gcc to HP Nonstop Ux
- From: Jim Wilson <wilson at tuliptree dot org>
- To: praveen <praveen dot khanna at wipro dot com>
- Cc: "'Neil Booth'" <neil at daikokuya dot co dot uk>, gcc-bugs at gcc dot gnu dot org
- Date: 06 May 2003 15:24:51 -0400
- Subject: RE: Porting gcc to HP Nonstop Ux
- References: <002c01c312e1$8051ec60$bb2a1fac@khannababysPC>
On Mon, 2003-05-05 at 04:37, praveen wrote:
> Please let me know your opinion on using the disassembler output for
> comparison on our platforms(I tested the same today on hash.o and it
> seemed to work. I have attached the same for your reference.)
This is a reasonable thing to do if the assembler is being difficult.
This won't find all possible problems though, since this is only looking
at the code. There could be differences in variable definitions or
debug info that won't be visible in the disassembly. Comparing .S files
is a better approach if at all possible.
You can force the intermediate assembly files to have the base file name
instead of a temp file name by using --save-temps. This will make
bootstraps take a bit longer (extra file I/O), and use more disk space,
but should allow the file comparisons to work. There is a comment about
this in the install.texi file. It mentions that this may be necessary
for some old Dec Unix machines; these are probably MIPS based machines
it is talking about.
The .file directive in the assembly output should be used by the
assembler for the file name it puts in the symbol table. Perhaps the
syntax is confusing the assembler you have. You could experiment with
the form of this assembler line. Try dropping the "1" from the line.
Try dropping the entire line. If you have access to an assembler
manual, it might have info about how to emit this line to get the result
you want.
Jim