[lsb-discuss] ISO standards for LSB standards after 3.1?

Dallman, John john.dallman at siemens.com
Mon Sep 30 09:18:05 UTC 2013

> There /is/ a task to uplift the level of C++ support to match relatively
> current libstdc++ usage.  Maybe that will be enough?

I can tell you the criterion: if one can use -std=c++11 and have everything
that is used from libstdc++ in using all the GCC-implemented C++11 features
controlled by LSB, then it's good enough.

It also needs to be possible to use -std=c++98 in a LSB build and have the
resulting binaries work on a C++11 run-time, and link against and run
with binaries built with C++11, to the extent that raw GCC permits this. I
believe that's "fully", but I can't demonstrate that: I'm going off what
I've been told by SUSE and Red Hat.

> Also, since 5.0 has been considerably slower getting to the finish line
> than anticipated, maybe the situation I refer to ("not ready") is no longer
> true?

I suspect not. As a development platform, I'd suggest using SLES11sp3 with the
add-on GCC 4.7.2. That's a pretty straight GCC 4.7.2 implementation.

I'd recommend *not* using the RHEL6 add-on for GCC 4.7.2, because that implements
the new libstdc++ functions weirdly: it statically links them to each and every
binary that uses them, so that said binaries can run on top of a GCC 4.4 libstdc++.


John Dallman
Siemens Industry Software Limited is a limited company registered in England and Wales.
Registered number: 3476850.
Registered office: Faraday House, Sir William Siemens Square, Frimley, Surrey, GU16 8QD.

More information about the lsb-discuss mailing list