This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: egcs-CVS19980627, mips-sgi-irix6.2 bootstrap problem ...
- To: Jim Wilson <wilson at cygnus dot com>
- Subject: Re: egcs-CVS19980627, mips-sgi-irix6.2 bootstrap problem ...
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 29 Jun 1998 21:58:49 -0600
- cc: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>, egcs-bugs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <199806292359.QAA00248@rtl.cygnus.com>you write:
> Apparently, the new SGI assembler requires that assembler files end with a .s.
> If there is no .s extension, then the assembler does the same thing the
> compiler would, it treats it as an object file that should be passed to the
> linker. The linker then gives an error because an assembler file is not
> a valid object file.
Damn "smart" assemblers.
> If it is necessary to drop the extension for security reasons, then we have
> a problem. It looks very hard to fix this in the irix6 port. Easiest solution
> would be a way to disable this gcc.c feature for targets that have
> inconvenient assemblers, if GNU as if not being used.
>
> This patch gets my irix6 build going again. Reversing the uses of suffix
> and p would probably be more elegant.
>
> Mon Jun 29 16:55:05 1998 Jim Wilson <wilson@cygnus.com>
>
> * gcc.c (do_spec_1, case 'g'): Delete redundant setting of
> delete_this_arg. Restore p at end of MKTEMP_EACH_FILE code.
This undoes many of the security fixes.
I think the way to handle this is to pass the desired suffix down
to our mkstemp and have it append the suffix to the template temporary
file.
I'll see what I can come up with. But the first order of business
tonight is to check in the merge.
jeff