[with patch] hanke -- disturbing matrix constructor issue involving rows versus columns

[was]

Hi William, I just found a disturbing trait about Matrix constructions, and was wondering if you could include a rows/columns flag to address it.

When a matrix is constructed from tuples, it assumes that you want to use these numbers as *rows*, regardless of whether the tuples are appropriately sized. Since there is no way of deciding which is meant for square matrices, it seems reasonable to add an extra (mandatory) flag to the constructor for a list of tuples to ask which is meant.

I hope your semster is going well. See you,

-Jon (Hanke)

--------------------------------------------------------------------
 
sage: M2 = MatrixSpace(ZZ,2,4)(range(8)); M2
[0 1 2 3]
[4 5 6 7]
 
sage: M2.columns()
 [(0, 4), (1, 5), (2, 6), (3, 7)]
 
sage: MatrixSpace(ZZ,2,4)(M2.columns())
[0 4 1 5]
[2 6 3 7]
 
sage: M2 == MatrixSpace(ZZ,2,4)(M2.columns())
 False
 
sage: M2 == MatrixSpace(ZZ,2,4)(M2.rows())
 True

[mabshoff]

This is still a problem with Sage 2.8.2. The 2.8.3 release might cut it a little close because there are still a large number of tickets left (to be fixed in roughly 1 day).

Cheers,

Michael