Regarding the static libraries.

 From LAPACK group:
"Because trying to support every shared library kind and method of
generation is far outside our scope.  Yes, we know of libtool.
We also have many users on AIX, which uses XCOFF and where
libtool falls flat.  CMake is an option, but it takes more work
than anyone can devote right now.

Plus, to build a shared, ATLAS-tuned version, you need the shared
objects in an ar archive.  Yet another variation. wheee...

Check the build scripts for GNU/Linux distributions for building
ELF-based shared libraries.  In particular, Debian GNU/Linux does
a good job.

Jason"

Also this package will not be integrated to any of the Solaris  
consolidations. It will be integrated to the "Developer Tools"  
consolidation. This is a special process be developed for integrating  
this type of tool to OpenSolaris. It is following the same process  
used to integrate SunStudio to OpenSolaris.

Bruce

On Nov 25, 2008, at 10:50 AM, James Carlson wrote:

> John Fischer writes:
>> This project proposes to integrate the LAPACK (Linear Algebra  
>> package) into
>> a Minor release of Solaris.  LAPACK provides routines for solving  
>> systems of
>> equations.  It is written in Fortran and requires no special  
>> compiler.  The
>> only issue that caused it to fail the check list and require a  
>> review is
>> the fact that it will only deliver static libraries.  This is how the
>> community delivers its components.
> [...]
>>      Will the project team work with the upstream community to  
>> resolve
>>      architectural issues of interest to Sun?
>>      [x ] Yes
>>      [ ] No - briefly explain
> [...]
>>      Are static versions of the libraries being delivered?
>>      [x ] Yes - ARC review required
>>      [ ] No
>>
>>      The package as is only generates static libraries. I would  
>> have to
>>      change the opensource package to make it generate dynamic.
>
> Given that avoiding static libraries _is_ an issue of architectural
> interest to Sun, could you explain how difficult it might be to
> generate normal dynamic libraries?  Is it significantly harder than at
> least changing "ar c" to "ld -r"?
>
> -- 
> James Carlson, Solaris Networking              <james.d.carlson at sun.com 
> >
> Sun Microsystems / 35 Network Drive        71.232W   Vox +1 781 442  
> 2084
> MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442  
> 1677


Reply via email to