This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: libstdc++/7926

On Fri, Nov 01, 2002 at 07:41:29PM -0600, Benjamin Kosnik wrote:
> tested x86/linux
> Stops pretending that i386's atomicity.h ever existed, use generic. Arm,
> same. Use i486 for x86-64, and athlon. Remove extraneous bits.

Thanks for doing this.  One comment:

> !   i486 | i586 | i686 | i786)
> !     # add athlons here maybe
>       cpu_include_dir="cpu/i486"
>       ;;
> !   athlon* | i586 | i686 | i786 | x86_64)
>       cpu_include_dir="cpu/i486"
>       ;;

The comment about adding Athlons was a misleading mistake of mine that
shouldn't have stayed in the repository.  This case statement switches on
$target_cpu, which is canonicalized.  So Athlons have always been getting
this file:

fenric 17% ls -lF build/build-2002-11-01-trunk/athlon_mp-pc-linux-gnu/libstdc++-v3/include/athlon_mp-pc-linux-gnu/bits/atomicity.h
lrwxrwxrwx    1 pme      pme            60 Nov  1 16:52 build/build-2002-11-01-trunk/athlon_mp-pc-linux-gnu/libstdc++-v3/include/athlon_mp-pc-linux-gnu/bits/atomicity.h -> /home/pme/src/trunk/libstdc++-v3/config/cpu/i486/atomicity.h
fenric 18%

I've never done anything with x86-64, so I don't know whether they get
canonicalized to i[456]86 or not.

To avoid confusing future hackers, how about changing that line to just


with x86_64 if it needs to be there?


I would therefore like to posit that computing's central challenge, viz. "How
not to make a mess of it," has /not/ been met.
                                                 - Edsger Dijkstra, 1930-2002

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]