[Numpy-svn] [numpy/numpy] e77b7b: BUG: In `norm`, always cast non-floating point arr...

[GitHub]

  Branch: refs/heads/master
  Home:   https://github.com/numpy/numpy
  Commit: e77b7b98df233f72a9d50934a4bf5b93c163b482
      https://github.com/numpy/numpy/commit/e77b7b98df233f72a9d50934a4bf5b93c163b482
  Author: John Kirkham <kirkhamj at janelia.hhmi.org>
  Date:   2016-01-22 (Fri, 22 Jan 2016)

  Changed paths:
    M numpy/linalg/linalg.py

  Log Message:
  -----------
  BUG: In `norm`, always cast non-floating point arrays to 64-bit floats. Otherwise, weird integer roundoff errors give faulty results in some cases.


  Commit: bc4a17ed89004ed63558a3e7f0bd035580777aa7
      https://github.com/numpy/numpy/commit/bc4a17ed89004ed63558a3e7f0bd035580777aa7
  Author: John Kirkham <kirkhamj at janelia.hhmi.org>
  Date:   2016-01-22 (Fri, 22 Jan 2016)

  Changed paths:
    M numpy/linalg/tests/test_linalg.py

  Log Message:
  -----------
  TST: Verify that `norm` is properly casting values to floats as needed.


  Commit: 43c6a89128347928c5fe26f67ba2a0a022f00822
      https://github.com/numpy/numpy/commit/43c6a89128347928c5fe26f67ba2a0a022f00822
  Author: John Kirkham <kirkhamj at janelia.hhmi.org>
  Date:   2016-01-22 (Fri, 22 Jan 2016)

  Changed paths:
    M numpy/linalg/linalg.py
    M numpy/linalg/tests/test_linalg.py

  Log Message:
  -----------
  BUG: Make sure that the `ord=0` case returns a float.


  Commit: 75d5b59bca181ee7e5ba872999014006c4b6c3f3
      https://github.com/numpy/numpy/commit/75d5b59bca181ee7e5ba872999014006c4b6c3f3
  Author: John Kirkham <kirkhamj at janelia.hhmi.org>
  Date:   2016-01-22 (Fri, 22 Jan 2016)

  Changed paths:
    M numpy/linalg/linalg.py

  Log Message:
  -----------
  DOC: Update `norm` docstring to include examples that reflect that all computations are done with floating point numbers.


  Commit: 66156a049aa833a06e6747dbe88676a61e034e8d
      https://github.com/numpy/numpy/commit/66156a049aa833a06e6747dbe88676a61e034e8d
  Author: Charles Harris <charlesr.harris at gmail.com>
  Date:   2016-01-22 (Fri, 22 Jan 2016)

  Changed paths:
    M numpy/linalg/linalg.py
    M numpy/linalg/tests/test_linalg.py

  Log Message:
  -----------
  Merge pull request #7088 from jakirkham/cast_float_linalg

BUG: Have `norm` cast non-floating point arrays to 64-bit float arrays


Compare: https://github.com/numpy/numpy/compare/d5cef016b336...66156a049aa8