[Rd] Minor issue building 2.3.0 under Red Hat Linux 9
Prof Brian Ripley
ripley at stats.ox.ac.uk
Thu Apr 27 22:34:53 CEST 2006
On Thu, 27 Apr 2006, Marc Schwartz wrote:
> On Thu, 2006-04-27 at 10:53 +0100, Prof Brian Ripley wrote:
>> On Thu, 27 Apr 2006, Dominick Samperi wrote:
>>
>>> Building R 2.3.0 fails under Red Hat 9.0 because
>>> ssize_t is not defined in src/modules/internet/sock.h.
>>> Inserting #include <sys/types.h> at the top of this
>>> file fixes the problem.
>>
>> That's interesting: POSIX says it must be defined in unistd.h, and that is
>> included before sock.h in all the C files AFAICS (and this is why all the
>> Unix-alike platforms tested before release do compile that module).
>>
>> It's a bug in RH9 (there seems to be no 9.0), and that is maintained by
>> Fedora Legacy -- do they have an update?
>
> Prior to RH 9 (ie. RH 8.0), there were fractional releases with updates.
> However, with 9, while updates were provided, the numbering scheme
> changed to integer releases only and this is still the case with FC.
>
> Fedora Legacy's raison d'etre is to provide security updates but not
> necessarily functional updates and/or general bug fixes. This seems to
> be the case for RH 9 if one reviews the list of updates that they have
> provided to date:
>
> http://www.fedoralegacy.org/updates/RH9/
>
> There is no guarantee from the FL folks as to how long they will support
> older releases beyond the roughly 1.5 years post EOL for a prior version
> of RH/FC. This is articulated in their "1-2-3 and out" policy and we are
> well beyond the 1.5 years for RH 9.
>
> Seeing Peter's response on SuSE 8.0, one of the things that might be in
> common here is the use of the 2.4 series kernels. Was there a change of
> some sort with the 2.6 kernels that would impact this situation?
I have no idea, but have modified the sock.h to have a belt-and-braces
approach.
I'll say once again how unfortunate it is that people do not report things
like this prior to release.
--
Brian D. Ripley, ripley at stats.ox.ac.uk
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
More information about the R-devel
mailing list