r/l shift on uint64 scalar and python int failing #4151

Activity

1. charris commented on Feb 24, 2014

   

   charris

   on Feb 24, 2014

   Member

   More actions

   The problem is that an attempt is made cast the signed 0 and unit64 to a common type. The problem looks specific to array scalar types

   In [46]: a = np.array([123]*2, dtype='uint64')
   
   In [47]: a >> 2
   Out[47]: array([30, 30], dtype=uint64)
   

   Due to the difference in array/scalar and scalar/scalar casting rules. This isn't the only place this causes a problem.
2. charris

   added

   Defect

   on Feb 24, 2014
3. ianwlarson commented on Jul 28, 2017

   

   ianwlarson

   on Jul 28, 2017

   More actions

   This issue also affects me.
4. mattip

   removed

   priority: normal

   on Oct 21, 2018
5. mattip commented on Sep 18, 2023

   

   mattip

   on Sep 18, 2023

   MemberAuthor

   More actions

   On 1.26 this raises a (conceptually correct) TypeError("ufunc 'right_shift' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule 'safe'"). Closing.
6. mattip

   closed this as completedon Sep 18, 2023