Array.copy(): Also set the queue of the new array.

Also documents Array.copy(), bumps the version, and fixes a folding error in array.py.

This closes issue 184 on github.

changed title from Array.copy(): Set the queue of the new array. to Array.copy(): Also set the queue of the new array.

changed the description

added 1 commit

• 1ca7e583 - Whitespace fix.

Compare with previous version

restored source branch array-copy-associate-queue

added 1 commit

• 5ca953f8 - Grammar fix.

Compare with previous version

added 1 commit

• 373ac7b1 - test_array(): Check for queue identity; also add a missing queue check.

Compare with previous version

This seems to have killed two fancy indexing checks. Not sure why--could you take a look?

 test_fancy_indexing[ctx_factory=<context factory for <pyopencl.Device 'pthread-Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz' on 'Portable Computing Language' at 0x2304d60>>] 
Traceback (most recent call last):
  File "/var/lib/gitlab-runner/builds/569f1b25/0/inducer/pyopencl/test/test_array.py", line 1161, in test_fancy_indexing
    numpy_dest[numpy_idx] = numpy_src
ValueError: shape mismatch: value array of shape (2,) could not be broadcast to indexing result of shape (3,)
 test_multi_put[ctx_factory=<context factory for <pyopencl.Device 'pthread-Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz' on 'Portable Computing Language' at 0x2304d60>>] 
Traceback (most recent call last):
  File "/var/lib/gitlab-runner/builds/569f1b25/0/inducer/pyopencl/test/test_array.py", line 1201, in test_multi_put
    ary[idx.get()] = np.arange(0, 3, dtype=np.float32)
ValueError: shape mismatch: value array of shape (3,) could not be broadcast to indexing result of shape (6,)

I can't reproduce this on my machine (Python 3.4, numpy 1.11.1). However, I think that the new release of numpy could be a possible cause.

Argh. Here is a test balloon off of master. Good catch on the numpy release.

Yep, looks like it's a numpy thing.

mentioned in commit b42e7ebb

merged

mentioned in issue #2 (closed)

So this is good to go, and now there's #2 (closed).

Thanks for working on this!

You're welcome