Consolidation causes core dump

Code can reproduce:

import numpy as np
import tiledb

dense_array_name = 'test_1d'

ctx = tiledb.Ctx()

dom = tiledb.Domain(tiledb.Dim(ctx=ctx, domain=(0, 11), tile=12, dtype=np.int64), ctx=ctx)
schema = tiledb.ArraySchema(ctx=ctx, domain=dom, sparse=False,
                            attrs=[tiledb.Attr(ctx=ctx, dtype=np.int64)])

tiledb.DenseArray.create(dense_array_name, schema)

with tiledb.DenseArray(dense_array_name, ctx=ctx, mode='w') as A:
    A[:2] = np.arange(2)
    A[2:4] = np.arange(2, 4)
    A[4:6] = np.arange(4, 6)
    A[6:8] = np.arange(6, 8)
    A[8:10] = np.arange(8, 10)
    A[10:12] = np.arange(10, 12)
    
tiledb.consolidate(uri=dense_array_name)

I have just updated the tiledb-py version to 0.4.1 via conda. Problem still. My laptop is Macbook pro with Mac OS High Sierra. Python version is 3.7.3.

[ihnorton]

Hi @chineking thanks again for the report and the easy reproduction script. We just merged a fix in core tiledb: https://github.com/TileDB-Inc/TileDB/pull/1213

We are planning to do a point release at the end of this week for TileDB core, and will update the conda and PyPI packages to include the fix. I will leave this issue open until then.

[ihnorton]

@chineking thanks for the report and repro script. I reproduced in py and c++, and opened an issue in the core TileDB library.