Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on Jul 26, 2010
  1. @rgommers
Commits on Jul 24, 2010
  1. @rgommers
Commits on Jul 15, 2010
  1. @rgommers
Commits on Jul 14, 2010
  1. @rgommers
  2. @pv

    BUG: spatial/tests: fix spurious failures when int64 is not inherited…

    pv authored
    … from Python int (#1229)
    
    Thanks to Christoph Gohlke.
    
    (cherry picked from commit r6605)
  3. @pv

    BUG: fftpack/tests: fix spurious failures on machines where double==l…

    pv authored
    …ongdouble (fixes #1229)
    
    Thanks to Christoph Gohlke.
    
    (cherry picked from commit r6604)
Commits on Jul 13, 2010
  1. @rgommers
  2. @rgommers

    BUG: Fix and regression test for ticket #1228 (in Rbf, by Scott Sincl…

    rgommers authored
    …air).
    
    (cherry-picked from r6599)
  3. @rgommers
Commits on Jul 11, 2010
  1. @rgommers
  2. @rgommers

    TST: Increase relative precision in arccosh/arctanh tests in 0.8.x br…

    rgommers authored
    …anch.
    
    Cherry-picked from r6593.
Commits on Jul 5, 2010
  1. @pv

    0.8.x: DOC: get Scipy version number automatically

    pv authored
    (cherry picked from r6591)
Commits on Jul 4, 2010
  1. @rgommers
  2. @rgommers
  3. @rgommers

    BUG: optimize.fmin_ncg failed to converge occasionally on Windows due…

    rgommers authored
    … to (==
    
    0.0) comparison. Fixes #1150.
Commits on Jul 3, 2010
  1. @rgommers
  2. @rgommers
  3. @rgommers
  4. @rgommers

    BUG: Ticket #1214: optimize.newton could return an incorrect value if…

    rgommers authored
    … the function coefficients and initial guess were integers.
    
    Backport of r6579.
  5. @rgommers
Commits on Jun 27, 2010
  1. @rgommers

    DOC: Add curve_fit and changes to fsolve/leastsq to release notes.

    rgommers authored
    Also add the latter changes to their respective docstrings, and fix some
    asserts in their tests.
  2. @pv

    BUG: fftpack: use single-precision FFT implementation only for "easy"…

    pv authored
    … sizes
    
    For large "difficult" sizes, the current single precision implementation
    apparently falls back to a non-optimal algorithm, and can produce large
    rounding errors. (bug #1212)
    
    This commit provides a work-around, by enabling the single-precision
    algorithms only for sizes that are composite numbers of 2, 3, and 5, for
    which FFTPACK has explicit support. This is probably a rather
    conservative approach.
Commits on Jun 21, 2010
  1. Update nnlf method documentation in stats.

    Travis Oliphant authored
Commits on Jun 20, 2010
  1. @pv

    TST: special: add test for complex erf

    pv authored
    (cherry picked from commits r6549, r6550)
  2. @pv

    BUG: special: adjust complex erf switchover radius to improve accurac…

    pv authored
    …y (addresses #683)
    
    (cherry picked from commit r6548)
  3. @pv

    ENH: special: add assert_func_equal to clean up the way dataset tests…

    pv authored
    … are implemented
    
    (cherry picked from commit r6547)
Commits on Jun 19, 2010
  1. @rgommers
Commits on Jun 18, 2010
  1. @rgommers
  2. @rgommers
  3. @rgommers
  4. @rgommers
  5. @rgommers
  6. @rgommers

    Deprecate io.recaster module.

    rgommers authored
Something went wrong with that request. Please try again.