Re: libstdc++ wants clock_gettime without libpthread

[Jonathan Wakely <jwakely dot gcc at gmail dot com>]

On 19 October 2012 01:22, Roland McGrath  wrote:
>
> Every approach requires some amount of work in glibc, some perhaps a lot of
> work, and some require significant work in libstdc++ and elsewhere.  Even
> if infinite labor were instantaneously available, there is no approach that
> doesn't have some downside.

I suspected that would be the case. Thanks for the quick and detailed
answer.  Most of it involves ELF and/or linker features I'm unfamiliar
with but I'll try to grok it all!

I described an alternative in the last two paragraphs of
http://gcc.gnu.org/ml/libstdc++/2012-05/msg00085.html which (if it's
possible) would have the advantage that users only pay the performance
penalty of slower atomic ref-counts in single-threaded apps when they
actually need to use std::chrono::system_clock, rather than it being a
configure-time choice that affects everything linked to that build of
libstdc++, whether it uses the clocks or not.  I'll look further into
that option.

Thanks again.