blas error and crash when doing ldivide/rdivide of a scalar by a vector

[John W. Eaton]

On 15-Jan-2008, Jason Riedy wrote:

| And Ben Abbott writes:
| > Thanks for the confirmation. May we assume you'll make the needed
| > patches to the svn?
| 
| I'll get something done, but I have no idea when the next
| release will be.

Given that we can't count on a release date much less updates to
vendor libraries like vecLib, I guess we should consider replacing the
query call with an explicit formula for the workspace requirement.  It
would help a lot if someone could provide that as a patch for Octave.

Also, has the (correct) formula for the required workspace in *GELSD
always been the same, or has it changed from one release of LAPACK to
the next?

Thanks,

jwe