Mercurial > octave-nkf
view examples/myfevalf.f @ 14624:edf9ca8a92a8 stable
when redimensioning, always pad dim_vector objects with 1 (bug #33216)
* dim-vector.cc (dim_vector::redim): Always pad with 1.
* dim-vector.h (dim_vector::redim): Update comment.
* Array.cc (Array<T>::assign): Query dimensions for all zeros before
redimensioning.
* ov-struct.cc: New test.
author | John W. Eaton <jwe@octave.org> |
---|---|
date | Fri, 11 May 2012 12:33:13 -0400 |
parents | 4295d634797d |
children |
line wrap: on
line source
subroutine mexFunction (nlhs, plhs, nrhs, prhs) implicit none integer*4 nlhs, nrhs * The following will need to be integer*8 on 64-bit systems, otherwise * these variables won't be large enough to hold pointers... integer*4 plhs(*), prhs(*) integer*4 mxIsString, mxGetString, mxGetN, mexCallMATLAB integer*4 status, len character*100 str call mexPrintf ('Hello, World!') if (nrhs .lt. 1 .or. mxIsString (prhs(1)) .ne. 1) then call mexErrMsgTxt ('function name expected') endif len = mxGetN (prhs(1)) status = mxGetString (prhs(1), str, 100) call mexPrintf ('FORTRAN will call the interpreter now') status = mexCallMATLAB (nlhs, plhs, nrhs-1, prhs(2), str(1:len)) return end