Implement single precision rand, randn, rande, randg and randp generators (bug #34351, #36293) * oct-rand.cc (octave_rand:do_matrix): Remove method. (octave_rand::do_float_scalar, octave_rand::do_float_nd_array, octave_rand::do_float_vector): New methods. * oct-rand.h (octave_rand:do_matrix, octave_rand::matrix): Remove methods. (octave_rand::do_float_scalar, octave_rand::do_float_nd_array, octave_rand::do_float_vector, octave_rand::float_scalar, octave_rand::do_nd_array, octave_rand::float_vector): Declare new methods. * randgamma.c (oct_fill_float_randg, oct_float_randg): New functions. * randgamma.h (oct_fill_float_randg, oct_float_randg): Declare new functions. * randpoisson.c (oct_fill_float_randp, oct_float_randp): New functions. (poisson_cdf_lookup_float, poisson_rejection_float): New static functions. * randpoisson.h (oct_fill_float_randp, oct_float_randp): Declare new functions. * randmtzig.c (randu64) : Remove static function. (ALLBITS): Remove compile flag logic. (randu32): Change return type to float. (oct_float_randu, oct_float_randn, oct_float_rande, oct_fill_float_randu, oct_fill_float_randn, oct_fill_float_rande): New functions. (create_ziggurat_float_tables): New static function. * randmtzig.h (oct_float_randu, oct_float_randn, oct_float_rande, oct_fill_float_randu, oct_fill_float_randn, oct_fill_float_rande): Declare new functions. * rand.cc (do_rand): Add logic to parse "double" or "single" as final argument.

--- a/liboctave/randpoisson.h
+++ b/liboctave/randpoisson.h
@@ -32,6 +32,9 @@
 extern OCTAVE_API double oct_randp (double L);
 extern OCTAVE_API void oct_fill_randp (double L, octave_idx_type n, double *p);
 
+extern OCTAVE_API float oct_float_randp (float L);
+extern OCTAVE_API void oct_fill_float_randp (float L, octave_idx_type n, float *p);
+
 #ifdef  __cplusplus
 }
 #endif