Re: Re: [tree-ssa, cvs] spec failure: unrecognized spec option '<'

[Melvin Hadasht <melvin.hadasht@free.fr>]

Hi,

> I just checked on mainline cvs, and the same problem's
> there. However on 3.3 branch it looks fine. I'm
> guessing this was introduced when Neil rewrotes
> specs/options handling on mainline.

Ok, thanks for checking: on my pII 266, compilation is not instantaneous.

>P.S. I'm still puzzled with the error you get though
> (I see nothing)...

I searched google and the archive, and I did not found anyone having the 
same problem. So it seems I have a broken system.

I "echo"ed the offending command line (given in my last post), to do
env-var expansion and added the -v switch to the resulting gcc command,
and run it and I found this:

Reading specs from ./specs
Configured with: /usr/local/src/cvs/gcc-3.3.2/configure 
Thread model: posix
gcc driver version 3.3.2 executing gcc version 3.5-tree-ssa
gcc: spec failure: unrecognized spec option '<'
 /usr/local/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/cc1 -quiet -v -I. -I. -I/usr/local/src/cvs/gcc/gcc -I/usr/local/src/cvs/gcc/gcc/. -I/usr/local/src/cvs/gcc/gcc/../include -I/usr/local/src/cvs/gcc/gcc/../libbanshee/libcompat -I/usr/local/src/cvs/gcc/gcc/../libbanshee -I/usr/local/src/cvs/gcc/gcc/../libbanshee/points-to -DIN_GCC -DHAVE_CONFIG_H -DSTANDARD_STARTFILE_PREFIX="../../../" -DSTANDARD_EXEC_PREFIX="/usr/local/gcc-ssa/lib/gcc/" -DSTANDARD_LIBEXEC_PREFIX="/usr/local/gcc-ssa/libexec/gcc/" -DDEFAULT_TARGET_VERSION="3.5-tree-ssa" -DDEFAULT_TARGET_MACHINE="i686-pc-linux-gnu" -DSTANDARD_BINDIR_PREFIX="/usr/local/gcc-ssa/bin/" -DTOOLDIR_BASE_PREFIX="../../../../" /usr/local/src/cvs/gcc/gcc/fortran/gfortranspec.c mcpu=* -quiet -dumpbase gfortranspec.c -auxbase gfortranspec -O -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -version -fno-common -o .s
cc1: error: output filename specified twice
GNU C version 3.3.2 (i686-pc-linux-gnu)
	compiled by GNU C version 3.3.2.
GGC heuristics: --param ggc-min-expand=38 --param ggc-min-heapsize=15861

The "mcpu=*" gives cc1: error: bad value (*) for -mcpu= switch. Which I suppose 
is the result of the ./specs file being incorrect.
(Note that for this test I reconfigured without c++, thus here it is compiling gfortranspec.c 
and not g++spec.c)

Concerning the ./specs file, removing the "%<mcpu=*" is not sufficient, 
a closing brace "}" must also be adjusted.

I don't have any idea on how to find where the problem is, but I'm willing to do any tests. 
Should I file a bug report?

If I understand you well, there are two problems: 
i) the incorrect specs file (for which I would file a bug) which is seen also in other systems and 
ii) my system that issues the error while your does not.

Thanks.