On 07/ 3/10 02:28 PM, kcrisman wrote:
When R is built on Solaris, the configure option --without-ICU is added.
Otherwise R will not build.
The problem seems to be that R's configure script does not appear to have the
sense to disable the library if it can't find the library. One might reasonably
expect the configure script to build R without the library if it is able to do
so, rather than give an error message, which is what is happening on some
systems at least.
This seems reasonable - you may want to email the R list again about
this. It would also be worth knowing what one loses if you don't have
ICU.
On the other hand, this is about Sage being *built* on a machine with
the correct library, but then *run* on one which doesn't (which would
likely be the case for many downloaded binaries?), so there might be
no way around this.
- kcrisman
I've submitted this issue to the R bug list
https://bugs.r-project.org/bugzilla3/show_bug.cgi?id=14327
since it would seem most sensible to me that R actually checks for the library
before trying to use it. As you can see here, if one does not add the option to
disable the library and one does not have the library, the failure is noted at
link time
http://trac.sagemath.org/sage_trac/ticket/8285
I think Mariah's suggestion at
http://trac.sagemath.org/sage_trac/ticket/9396
to add '--with-ICU=no' as a configure option when SAGE_FAT_BINARY=yes, is very
sensible.
I don't know what R would gain by having libicu, but I know the ICU library is
pretty large at 14.7 MB, so I very much doubt Sage will add the ICU library for R.
Dave
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
