Greg Schafer wrote:
> Ryan Oliver wrote:
>
>
>> LFS not affected in regards to the fact we can set any of
>> md_startfile_prefix{,_1} or startfile_prefix_spec in the specs file and
>> have it work because we DO use a standard specs file in the appropriate
>> place.
>>
>
> I'll restate in clear terms. You're modifying/creating files under /tools
> whilst inside the chroot. i.e. you have no clean separation.
>
>
rm /path/to/specs
mv /path/to/specs{,adjusted}
>> some unfair fud about the use of startfile_prefix_spec)
>>
>
> Fud? The *facts* haven't changed in years:
>
> - it's undocumented in the GCC literature
>
If you dont count the source code as documentation
> - some GCC devs have expressed a desire to get rid of it
>
And didn't
> - it's affected by the -specs= bug
>
and so is md_startfile_prefix, lets remove that too... by your
definition that spec is broken and shouldn't be used, even though it is
documented.
The question is is this a specs bug, or incorrect usage of the -specs
switch (you'd think it would have been fixed after all this time)
If you want I'll send you a patch to fix the "specs bug",
> - hardly anyone on the planet outside *LFS derived projects actually use
> it. In fact, its use in *LFS has actually _encouraged_ some folks to
> use it. Ughh.
>
I offered an alternative "documented spec" for those that find it
unpalatable.
> The only recent reference by a senior GCC dev I can find is here:
>
> http://gcc.gnu.org/ml/gcc-patches/2008-06/msg00842.html
>
> which confirms what we already know i.e. it was a prior attempt at solving
> multilib issues that should have been removed when MULTILIB_OSDIRNAMES was
> implemented.
>
But was left, presumably because it is the only way of allowing
non-sysrooted cross-compilers to override startfile prefix in the specs
file.
>
>
>> In your case wanting to keep using -specs you are probably better off
>> setting
>>
>
> Sorry dude. You can't just blow back in here years after being MIA and
> expect folks to listen to your idle speculation.
Idle speculation? I've been doing this a long time.
Read the gcc source. Test what I showed you.
> Folks will take notice
> when you have something concrete to offer i.e. well thought out, tested
> and published for all to see.
>
I encourage folks to test it and make up their minds for themselves.
Best Regards
[R]
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
