[bug #68300] blkdiag with arguments of type single returns result of type double

Rik <[email protected]> Sun, 3 May 2026 14:46:02 -0400 (EDT)

Newsgroups	gmane.comp.gnu.octave.bugs
Message-ID	<[email protected]>

Please use the bug tracker to post updates to a bug report.  The mailing list is intended as a read-only notification stream.  Info posted to this mailing list address won't appear in the tracker database where it is most useful.

Update of bug #68300 (group octave):

                  Status:                    None => Patch Reviewed

    _______________________________________________________

Follow-up Comment #5:

@Fernando: Thanks for noticing this.

@Dmitri: Interesting strategy to use Octave's own type-promotion logic with
addition operator to determine final type.  I asked for some tests to be run
to see how Matlab handles some of the edge cases.  See
https://octave.discourse.group/t/help-with-matlab-compatibility-of-blkdag/7520.

It seems that *blkdiag*, which is just assembling arrays, is more like the
function *cat* than it is like the addition operator *+*.  Test case:

A = ones (2,2, 'uint8');
B = ones (2,2, 'int32');
A + B
error: binary operator '+' not implemented for 'uint8 matrix' by 'int32
matrix' operations

BUT, you can concatenate these types and the output is the first integer type
seen.

C = cat (1, A, B);
class (C)
ans = uint8

D = cat (1, B, A);
class (D)
ans = int32


The results from testing in Matlab show that *blkdiag* seems to be doing the
same thing and choosing the first integer format found.





    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?68300>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/

signature.asc (application/pgp-signature, 228 B)