generalized ufunc signature problem fix #2953

Merged
merged 8 commits into from Mar 2, 2013

Conversation

Projects
None yet
4 participants
@ovillellas
Contributor

ovillellas commented Jan 31, 2013

Fix for a problem in the generalized ufunc harness. When a generalized ufunc had a repeated cardinality within a single matrix in the signature, the arguments passed to the kernel function were incorrect.

Signatures like (m,m)->(m,m) didn't work. This pull-request fixes this

+ for (i = 0; i < nop; ++i) {
+ op_axes[i][0] = -1;
+ }
+ }

This comment has been minimized.

Show comment Hide comment
@njsmith

njsmith Jan 31, 2013

Member

Is this NpyIter's buggy behaviour with scalars striking again?

@njsmith

njsmith Jan 31, 2013

Member

Is this NpyIter's buggy behaviour with scalars striking again?

@njsmith

This comment has been minimized.

Show comment Hide comment
@njsmith

njsmith Jan 31, 2013

Member

I haven't reviewed the logic carefully, but skimming the code nothing jumped out at me, no-one uses this code, and there are tests, so... guess it's fine?

Not sure why Travis hasn't run tests... I'll try poking it.

Member

njsmith commented Jan 31, 2013

I haven't reviewed the logic carefully, but skimming the code nothing jumped out at me, no-one uses this code, and there are tests, so... guess it's fine?

Not sure why Travis hasn't run tests... I'll try poking it.

@njsmith njsmith closed this Jan 31, 2013

@njsmith njsmith reopened this Jan 31, 2013

charris added a commit that referenced this pull request Mar 2, 2013

Merge pull request #2953 from ContinuumIO/gufunc-fix
generalized ufunc signature problem fix

@charris charris merged commit 985b267 into numpy:master Mar 2, 2013

1 check passed

default The Travis build passed
Details
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment