[Octave-bug-tracker] [bug #56232] Octave crash when inverting an empty sparse matrix.

[anonyme]

Follow-up Comment #8, bug #56232 (project octave):

David Bateman here (can't be bothered trying to recover my password) 

Several things I note here:

The function "tinverse" was written to handle triangular or permuted
triangular matrices. Though


C = sparse (ones (2)); C = C - C; matrix_type(C)
>> Full
C = sparse (2,2,0); Matrix_type(C)
>> Full


So something is screwing up in the detection of the matrix type of C-C and its
being detcted in the subtraction operation as triangular

I see to way of fixing this. Fix the matrix type detection code to detect "C -
C" as Full. In fact an empty sparse matrix could be considered as triangular,
so this seems wrong to me. Otherwise the function tinverse should be fixed to
allow for empty matrices. In that case dinverse might alos need to be adapted
for zeros en the diagonal

D.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?56232>

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