On 24 September 2005 at 22:02, Kurt Roeckx wrote:
| On Sat, Sep 24, 2005 at 01:53:56PM -0500, Dirk Eddelbuettel wrote:
| >
| > Could you please look by hand? This very much looks like a local hickup in
| > the buildd. If you look at the aggregation at
| > http://buildd.debian.org/build.php?&pkg=r-base
| > you see that 2.2.0.beta.20050922 built everywhere (but ia64 where the gnome
| > conf script it calls fell over).
| >
| > Not R's bug, and I intend to downgrade this -- unless you can convince that
| > there is something wrong the package. I have no reason to believe that there
| > is. The part failing above is also unchanged for many releases.
|
| The problem is that it's installing it in lib64 and then
| expects to find it in the lib dir. It should get
| installed in the lib dir (on atleast Debian) or various things
| will break.
|
| It seems to come from the configure script:
| ## We need to establish suitable defaults for a 64-bit OS
| libnn=lib
| case "${host_os}" in
| linux*)
| case "${host_cpu}" in
| x86_64|mips64|ppc64|sparc64|s390x)
| libnn=lib64
| ;;
| esac
|
| Atleast for Debian this is wrong.
Great work, Kurt R.! I will pass this along to Kurt H who handles R's
autoconf magic. Fixing this in configure is the right thing given the
dependency of all the other packging scripts on finding everything in
/usr/lib/R (as I just mentioned in my reply to Doug's reply to your bug
report).
Kurt H, Doug: Any reasons my we shouldn't force /usr/lib/R on amd64 as well?
[ As an aside, I think I wait for the 2.2.0 release to start mucking with the
nice options Brian has added to now also use /usr/share/R for the portable
stuff like docs and headers. ]
Cheers, Dirk
--
Statistics: The (futile) attempt to offer certainty about uncertainty.
-- Roger Koenker, 'Dictionary of Received Ideas of Statistics'
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
