It has mostly the same semantics as Matrix but can contain more general
things.  There might be some issues about different shapes of empties
though...  WIP on a fix for gnu-octave#1052.

Fixes gnu-octave#1211.

* inst/private/python_header.py: Add it.
* inst/private/python_ipc_native.m: Add it.

See gnu-octave#1194 for more information.

* inst/@sym/private/mat_rclist_access.m: Test it.

See gnu-octave#1194 for more information.

* inst/@sym/private/mat_rclist_asgn.m: Test it.

Closes gnu-octave#1213

* inst/@sym/{horzcat,vertcat}.m: Fix them.

* inst/@sym/vertcat.m: Implement it.

This is a prerequisite for making @sym/horzcat Array-compatible.

Partially fixes <gnu-octave#1215>.

* inst/@sym/transpose.m: Implement it.

* inst/@sym/horzcat.m: Simplify it.

* inst/@sym/private/mat_rclist_{access,asgn}.m: Remove them.

Closes gnu-octave#1216

The function is now compatible with non-Matrix 2D sym (e.g. Array,
TableForm). Also, it now outputs empty matrices in the same way as
upstream Octave (fixes gnu-octave#1218).

* inst/@sym/repmat.m: Re-implement it and add tests accordingly.

Closes gnu-octave#1219

Array linear indexing is not the same as Matrix linear indexing.
Fixes gnu-octave#1222.

This routine is supposed to return a logical Octave array.  Do linear
indexing on the flattened tolist output which will work on both Array
and Matrix.  Fixes gnu-octave#1221.

In fact it is failing elsewhere but the point of *this* test is to
change the orientation of a vector.

...with 'inst/private/python_header.py'.

Follow-up to 56bd719.

* inst/private/python_ipc_native.m: Disable it.

Closes gnu-octave#1223

uniop_bool_helper and friends: Array support

Fixes Issue gnu-octave#1226.  Related to Issue gnu-octave#1222.

Fixes gnu-octave#1228.

Closes gnu-octave#1227

This was broken with syntax errors since 2016...  Apply fix for Array
but leave a comment that this is unused and untested.  It is referred to
in isprime but commented out.

Follow-up to d3b1547.

* inst/@sym/vertcat.m: Use it.

* inst/@sym/private/mat_rclist_asgn.m: Use it.

Follow-up to 136184d.

* inst/@sym/transpose.m: Use it.

We did not run into issues in the past because many sympy functions
return a mutable matrix. However, 'transpose' used in
c9f6f0f returns an immutable
matrix. This change should avoid similar problems in the future.

Follow-up to c9f6f0f.

* inst/@sym/private/mat_replace.m: Make it mutable.

* inst/private/{python_header.py,python_ipc_native.m}: Generalise
function 'make_matrix_or_array', use sympy function 'flatten' in
'make_matrix_or_array'.

Closes gnu-octave#1231

Instead of doing it ourselves, prepare a list of lists and call the
centralized helper function.

More Array fixes

elementwise_op: call make_array_or_matrix

Generalises 'make_matrix_or_array' so that we can choose to represent
non-Matrix 2D sym to be something other than an Array in the
future. For example, we could use TableForm.

* inst/private/{python_header.py,python_ipc_native.m}: Rename function
'make_matrix_or_array' -> 'make_2d_sym', variable 'dbg_no_array' ->
'dbg_matrix_only' and adjust accordingly.
* inst/@sym/private/mat_rclist_{access,asgn}.m: Adjust accordingly.
* inst/@sym/vertcat.m: Adjust accordingly.

Add more 2D sym functions to abstract over the idea of 2D sym so that
we can choose to represent non-Matrix 2D sym to be something other
than an Array in the future. For example, we could use TableForm.

WIP: This is incomplete. More functions need to use these 2D sym
functions. Not sure if this is needed anymore as TableForm was shown
to be lacking various features such as indexing and taking
transpose. No other candidates are being considered at the moment.

* inst/private/{python_header.py,python_ipc_native.m}: Add new 2D sym
functions 'is_2d_sym', 'is_matrix', 'is_non_matrix_2d_sym',
'list_from_2d_sym' and 'shape_of_2d_sym'.
* inst/@sym/private/mat_rclist_{access,asgn}.m: Use them.
* inst/@sym/{transpose.m,vertcat.m}: Use them.