On Wed, 26 Mar 2003, Marc Schwartz wrote:
[...]
> Confirmed Roger's finding on RH 8.0, otherwise it compiles fine.
That comes from
cvs log ./share/perl/massage-Examples.pl
...
revision 1.15
date: 2003/03/26 20:12:56;
Redefine T/F to something more useful than NULL.
...
which made it an error to get("F"), alhough that is a legitimate operation
(it is assuming the value is FALSE that is the problem). (get("T") would
have failed too.)
Perhaps we need to delay putting out betas for a day after feature freeze
to allow us to test the frozen version.
> As an aside, RH 9 to be called 'Shrike', (curiously skipping 8.1 due to
> binary incompatibility issues) will be available on March 31 to paid RHN
> subscribers. It will be available a week later to all folks.
Martyn Plummer has already mentioned this and that he will be testing it
before release.
> Folks doing development may want to be aware, if not already, of the
> change in the threading approach used in RH 9 which will go to the
> Native POSIX Threads Library. Some very brief information on the change
> and possible application related problems is available here:
>
> https://listman.redhat.com/pipermail/phoebe-list/2003-March/004919.html
>
> There is also a white paper here:
>
> http://people.redhat.com/drepper/nptl-design.pdf
>
> I raise this in case this might in any way affect anyone's development
> considerations.
I don't think anything in Linux R uses threads, and we have problems with
e.g. threaded ATLAS that suggest that threads are unsafe: see the R-admin
manual.
--
Brian D. Ripley, [EMAIL PROTECTED]
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595
______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
