Script 'mail_helper' called by obssrc
Hello community,
here is the log from the commit of package python-numba for openSUSE:Factory
checked in at 2024-03-25 21:07:59
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/python-numba (Old)
and /work/SRC/openSUSE:Factory/.python-numba.new.1905 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "python-numba"
Mon Mar 25 21:07:59 2024 rev:46 rq:1160819 version:0.59.1
Changes:
--------
--- /work/SRC/openSUSE:Factory/python-numba/python-numba.changes
2024-03-09 20:56:44.595110375 +0100
+++ /work/SRC/openSUSE:Factory/.python-numba.new.1905/python-numba.changes
2024-03-25 21:09:50.308796571 +0100
@@ -1,0 +2,13 @@
+Fri Mar 22 20:05:25 UTC 2024 - Dirk Müller <dmuel...@suse.com>
+
+- update to 0.59.1:
+ * Fixed caching of kernels that use target-specific overloads
+ * Fixed a performance regression introduced in Numba 0.59 which
+ made ``np.searchsorted`` considerably slower.
+ * This patch fixes two issues with ``np.searchsorted``. First,
+ a regression is fixed in the support of ``np.datetime64``.
+ Second, adopt ``NAT``-aware comparisons to fix mishandling
+ of ``NAT`` value.
+ * Allow use of Python 3.12 PEP-695 type parameter syntax
+
+-------------------------------------------------------------------
Old:
----
numba-0.59.0.tar.gz
New:
----
numba-0.59.1.tar.gz
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Other differences:
------------------
++++++ python-numba.spec ++++++
--- /var/tmp/diff_new_pack.8ov4Ba/_old 2024-03-25 21:09:51.088825211 +0100
+++ /var/tmp/diff_new_pack.8ov4Ba/_new 2024-03-25 21:09:51.092825358 +0100
@@ -55,7 +55,7 @@
%endif
Name: python-numba%{?psuffix}
-Version: 0.59.0
+Version: 0.59.1
Release: 0
Summary: NumPy-aware optimizing compiler for Python using LLVM
License: BSD-2-Clause
++++++ numba-0.59.0.tar.gz -> numba-0.59.1.tar.gz ++++++
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/PKG-INFO new/numba-0.59.1/PKG-INFO
--- old/numba-0.59.0/PKG-INFO 2024-01-31 18:53:01.989844800 +0100
+++ new/numba-0.59.1/PKG-INFO 2024-03-18 14:32:45.775456000 +0100
@@ -1,6 +1,6 @@
Metadata-Version: 2.1
Name: numba
-Version: 0.59.0
+Version: 0.59.1
Summary: compiling Python code using LLVM
Home-page: https://numba.pydata.org
License: BSD
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/source/release/0.59.1-notes.rst
new/numba-0.59.1/docs/source/release/0.59.1-notes.rst
--- old/numba-0.59.0/docs/source/release/0.59.1-notes.rst 1970-01-01
01:00:00.000000000 +0100
+++ new/numba-0.59.1/docs/source/release/0.59.1-notes.rst 2024-03-18
14:32:37.000000000 +0100
@@ -0,0 +1,71 @@
+Version 0.59.1 (18 March 2024)
+------------------------------
+
+This is a bug-fix release to fix regressions in 0.59.0.
+
+
+CUDA API Changes
+~~~~~~~~~~~~~~~~
+
+Fixed caching of kernels that use target-specific overloads
+===========================================================
+
+Caching of kernels using target-specific overloads now works. This includes use
+of cooperative group sync, which is now implemented with a target-specific
overload.
+
+(`PR-#9447 <https://github.com/numba/numba/pull/9447>`__)
+
+
+Performance Improvements and Changes
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Improvement to ``np.searchsorted``
+==================================
+
+Fixed a performance regression introduced in Numba 0.59 which made
+``np.searchsorted`` considerably slower.
+
+(`PR-#9448 <https://github.com/numba/numba/pull/9448>`__)
+
+
+Bug Fixes
+~~~~~~~~~
+
+Fix issues with ``np.searchsorted`` not handling ``np.datetime64``
+==================================================================
+
+This patch fixes two issues with ``np.searchsorted``. First, a regression is
+fixed in the support of ``np.datetime64``. Second, adopt ``NAT``-aware
+comparisons to fix mishandling of ``NAT`` value.
+
+(`PR-#9445 <https://github.com/numba/numba/pull/9445>`__)
+
+Allow use of Python 3.12 PEP-695 type parameter syntax
+======================================================
+
+A patch is added to properly parse the PEP 695 syntax. While Numba
+does not yet take advantage of type parameters, it will no longer erroneously
+reject functions defined with the new Python 3.12 syntax.
+
+(`PR-#9459 <https://github.com/numba/numba/pull/9459>`__)
+
+
+Pull-Requests:
+~~~~~~~~~~~~~~
+
+* PR `#9445 <https://github.com/numba/numba/pull/9445>`_: Fix #9427
np.searchsorted on datetime64 (`sklam <https://github.com/sklam>`_)
+* PR `#9447 <https://github.com/numba/numba/pull/9447>`_: Fix Issue #9432,
caching of kernels using target-specific overloads (such as CG sync) (`gmarkall
<https://github.com/gmarkall>`_ `sklam <https://github.com/sklam>`_)
+* PR `#9448 <https://github.com/numba/numba/pull/9448>`_: Fix
`np.searchsorted` regression (`guilhermeleobas
<https://github.com/guilhermeleobas>`_)
+* PR `#9449 <https://github.com/numba/numba/pull/9449>`_: Remove deprecated
CondaEnvironment@1 (`sklam <https://github.com/sklam>`_)
+* PR `#9450 <https://github.com/numba/numba/pull/9450>`_: Fix gpuci versions
(`gmarkall <https://github.com/gmarkall>`_)
+* PR `#9459 <https://github.com/numba/numba/pull/9459>`_: Support pep695 type
param syntax (`sklam <https://github.com/sklam>`_)
+* PR `#9491 <https://github.com/numba/numba/pull/9491>`_: Fix
non-deterministic bug caused by unstableness in SSA (`sklam
<https://github.com/sklam>`_, `loicdtx <https://github.com/loicdtx>`_)
+
+
+Authors:
+~~~~~~~~
+
+* `gmarkall <https://github.com/gmarkall>`_
+* `guilhermeleobas <https://github.com/guilhermeleobas>`_
+* `sklam <https://github.com/sklam>`_
+* `loicdtx <https://github.com/loicdtx>`_
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/source/user/installing.rst
new/numba-0.59.1/docs/source/user/installing.rst
--- old/numba-0.59.0/docs/source/user/installing.rst 2024-01-31
18:52:55.000000000 +0100
+++ new/numba-0.59.1/docs/source/user/installing.rst 2024-03-18
14:32:37.000000000 +0100
@@ -262,6 +262,8 @@
+----------++--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
| Numba | Release date | Python | NumPy
| llvmlite | LLVM | TBB
|
+===========+==============+===========================+============================+==============================+===================+=============================+
+| 0.59.1 | 2024-03-18 | 3.9.x <= version < 3.13 | 1.22 <= version <
1.27 | 0.42.x | 14.x | 2021.6 <= version
|
++-----------+--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
| 0.59.0 | 2024-01-31 | 3.9.x <= version < 3.13 | 1.22 <= version <
1.27 | 0.42.x | 14.x | 2021.6 <= version
|
+-----------+--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
| 0.58.1 | 2023-10-17 | 3.8.x <= version < 3.12 | 1.22 <= version <
1.27 | 0.41.x | 14.x | 2021.6 <= version
|
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9048.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9048.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9048.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9048.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,17 +0,0 @@
-Dynamically Allocate Parfor Schedules
-=====================================
-
-This PR fixes an issue where a parallel region is executed in a loop
-many times. The previous code used an alloca to allocate the parfor
-schedule on the stack but if there are many such parfors in a loop
-then the stack will overflow. The new code does a pair of
-allocation/deallocation calls into the Numba parallel runtime before
-and after the parallel region respectively. At the moment, these
-calls redirect to malloc/free although other mechanisms such as
-pooling are possible and may be implemented later. This PR also
-adds a warning in cases where a prange loop is not converted to a
-parfor. This can happen if there is exceptional control flow in the
-loop. These are related in that the original issue had a prange
-loop that wasn't converted to a parfor and therefore all the parfors
-inside the body of the prange were running in parallel and adding to
-the stack each time.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9049.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9049.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9049.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9049.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Support multiple outputs in a ``@guvectorize`` function
-=======================================================
-
-This PR fixes `Numba #9058 <https://github.com/numba/numba/pull/9058>`_
-where it is now possible to call a guvectorize with multiple outputs.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9089.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9089.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9089.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9089.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Handling of `None` args fixed in `PythonAPI.call`.
-==================================================
-
-Fixing segfault when `args=None` was passed to `PythonAPI.call`.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9123.new_feature.rst
new/numba-0.59.1/docs/upcoming_changes/9123.new_feature.rst
--- old/numba-0.59.0/docs/upcoming_changes/9123.new_feature.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9123.new_feature.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Add support for ufunc attributes and reduce
-===========================================
-
-Support for ``ufunc.reduce`` and most ufunc attributes is added.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9126.np_support.rst
new/numba-0.59.1/docs/upcoming_changes/9126.np_support.rst
--- old/numba-0.59.0/docs/upcoming_changes/9126.np_support.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9126.np_support.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Added support for np.indices() function.
-========================================
-
-Support is added for ``numpy.indices()``.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9140.np_support.rst
new/numba-0.59.1/docs/upcoming_changes/9140.np_support.rst
--- old/numba-0.59.0/docs/upcoming_changes/9140.np_support.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9140.np_support.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Added support for `np.polynomial.polynomial.Polynomial` class.
-==============================================================
-
-Support is added for the `Polynomial` class from the package
`np.polynomial.polynomial`.
\ No newline at end of file
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9141.np_support.rst
new/numba-0.59.1/docs/upcoming_changes/9141.np_support.rst
--- old/numba-0.59.0/docs/upcoming_changes/9141.np_support.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9141.np_support.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Added support for functions ``np.polynomial.polyutils.as_series()``, as well
as functions ``polydiv()``, ``polyint()``, ``polyval()`` from
``np.polynomial.polynomial``.
-========================================================================================================================================================================
-
-Support is added for ``np.polynomial.polyutils.as_series()``,
``np.polynomial.polynomial.polydiv()``, ``np.polynomial.polynomial.polyint()``
(only the first 2 arguments), ``np.polynomial.polynomial.polyval()`` (only the
first 2 arguments).
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9144.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9144.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9144.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9144.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Fix propagation of literal values in PHI nodes.
-===============================================
-
-Fixed a bug in the literal propagation pass where a PHI node could be wrongly
-replaced by a constant.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9154.np_support.rst
new/numba-0.59.1/docs/upcoming_changes/9154.np_support.rst
--- old/numba-0.59.0/docs/upcoming_changes/9154.np_support.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9154.np_support.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Added support for np.unwrap() function.
-=======================================
-
-Support is added for ``numpy.unwrap()``. The ``axis`` argument is only
supported when its value equals -1.
\ No newline at end of file
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9169.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9169.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9169.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9169.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,6 +0,0 @@
-``numpy.digitize`` implementation behaviour aligned with numpy
-==============================================================
-
-The implementation of ``numpy.digitize`` is updated to behave per
-numpy in a wider set of cases, including where the supplied bins
-are not in fact monotonic.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9189.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9189.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9189.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9189.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,16 +0,0 @@
-``numpy.searchsorted`` and ``numpy.sort`` behaviour updates
-===========================================================
-
-* ``numpy.searchsorted`` implementation updated to produce
- identical outputs to numpy for a wider set of use cases,
- including where the provided array `a` is in fact not
- properly sorted.
-
-* ``numpy.searchsorted`` implementation bugfix for the case where
- side='right' and the provided array `a` contains NaN(s).
-
-* ``numpy.searchsorted`` implementation extended to support complex
- inputs.
-
-* ``numpy.sort`` (and ``array.sort``) implementation extended to
- support sorting of complex data.
\ No newline at end of file
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9223.cuda.rst
new/numba-0.59.1/docs/upcoming_changes/9223.cuda.rst
--- old/numba-0.59.0/docs/upcoming_changes/9223.cuda.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9223.cuda.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,6 +0,0 @@
-Added support for compiling device functions with a C ABI
-=========================================================
-
-Support for compiling device functions with a C ABI through the
-:func:`compile_ptx() <numba.cuda.compile_ptx>` API, for easier interoperability
-with CUDA C/C++ and other languages.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9235.cuda.rst
new/numba-0.59.1/docs/upcoming_changes/9235.cuda.rst
--- old/numba-0.59.0/docs/upcoming_changes/9235.cuda.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9235.cuda.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Make grid() and gridsize() use 64-bit integers
-==============================================
-
-`cuda.grid()` and `cuda.gridsize()` now use 64-bit integers, so they no longer
-overflow when the grid contains more than ``2 ** 31`` threads.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9242.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9242.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9242.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9242.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,6 +0,0 @@
-Fix SSA to consider variables where use is not dominated by the definition
-==========================================================================
-
-A SSA problem is fixed such that a conditionally defined variable will receive
-a phi node showing that there is a path where the variable is undefined.
-This affects extension code that relies on SSA behavior.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9244.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9244.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9244.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9244.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,26 +0,0 @@
-Fixed ``RecursionError`` in ``prange``
-======================================
-
-A problem with certain loop patterns using ``prange`` leading to
-``RecursionError`` in the compiler is fixed. An example of such loop is shown
-below. The problem would cause the compiler to fall into an infinite recursive
-cycle trying to determine the definition of ``var1`` and ``var2``. The pattern
-involves definitions of variables within an if-else tree and not all branches
-are defining the variables.
-
-.. code-block::
-
- for i in prange(N):
- for j in inner:
- if cond1:
- var1 = ...
- elif cond2:
- var1, var2 = ...
-
- elif cond3:
- pass
-
- if cond4:
- use(var1)
- use(var2)
-
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9245.performance.rst
new/numba-0.59.1/docs/upcoming_changes/9245.performance.rst
--- old/numba-0.59.0/docs/upcoming_changes/9245.performance.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9245.performance.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Improvement to IR copying speed
-===============================
-
-Improvements were made to the deepcopying of ``FunctionIR``.
-In one case, the ``InlineInlineables`` pass is 3x faster.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9246.highlight.rst
new/numba-0.59.1/docs/upcoming_changes/9246.highlight.rst
--- old/numba-0.59.0/docs/upcoming_changes/9246.highlight.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9246.highlight.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,13 +0,0 @@
-Python 3.12 Support
-===================
-
-The standout feature of this release is the official support for Python 3.12
-in Numba.
-
-Please note that profiling support is temporarily disabled in
-this release (for Python 3.12) and several known issues have been identified
-during development. The Numba team is actively working on resolving them.
-Please refer to the respective issue pages
-(`Numba #9289 <https://github.com/numba/numba/pull/9289>`_ and
-`Numba #9291 <https://github.com/numba/numba/pull/9291>`_)
-for a list of ongoing issues and updates on progress.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9249.np_support.rst
new/numba-0.59.1/docs/upcoming_changes/9249.np_support.rst
--- old/numba-0.59.0/docs/upcoming_changes/9249.np_support.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9249.np_support.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Adds support for checking if dtypes are equal.
-==============================================
-
-Support is added for checking if two dtype objects are equal, for example
`assert X.dtype == np.dtype(np.float64)`.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9267.cuda.rst
new/numba-0.59.1/docs/upcoming_changes/9267.cuda.rst
--- old/numba-0.59.0/docs/upcoming_changes/9267.cuda.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9267.cuda.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Prevent kernels being dropped by implementing the used list
-===========================================================
-
-Kernels are no longer dropped when being compiled and linked using nvJitLink,
-because they are added to the ``@"llvm.used"`` list.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9279.cuda.rst
new/numba-0.59.1/docs/upcoming_changes/9279.cuda.rst
--- old/numba-0.59.0/docs/upcoming_changes/9279.cuda.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9279.cuda.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Support for Windows CUDA 12.0 toolkit conda packages
-====================================================
-
-The library paths used in CUDA toolkit 12.0 conda packages on Windows are
-added to the search paths used when detecting CUDA libraries.
\ No newline at end of file
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9296.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9296.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9296.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9296.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,4 +0,0 @@
-Support negative axis in ufunc.reduce
-"""""""""""""""""""""""""""""""""""""
-
-Fixed a bug in ufunc.reduce to correctly handle negative axis values.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9309.change.rst
new/numba-0.59.1/docs/upcoming_changes/9309.change.rst
--- old/numba-0.59.0/docs/upcoming_changes/9309.change.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9309.change.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,8 +0,0 @@
-Make test listing not invoke CPU compilation.
----------------------------------------------
-
-Numba's test listing command ``python -m numba.runtests -l`` has historically
-triggered CPU target compilation due to the way in which certain test functions
-were declared within the test suite. It has now been made such that the CPU
-target compiler is not invoked on test listing and a test is added to ensure
-that it remains the case.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9310.highlight.rst
new/numba-0.59.1/docs/upcoming_changes/9310.highlight.rst
--- old/numba-0.59.0/docs/upcoming_changes/9310.highlight.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9310.highlight.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Move minimum supported Python version to 3.9.
-=============================================
-
-Support for Python 3.8 has been removed, Numba's minimum supported Python
-version is now Python 3.9.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9315.change.rst
new/numba-0.59.1/docs/upcoming_changes/9315.change.rst
--- old/numba-0.59.0/docs/upcoming_changes/9315.change.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9315.change.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,36 +0,0 @@
-Semantic differences due to Python 3.12 variable shadowing in comprehensions
-============================================================================
-
-Python 3.12 introduced a new bytecode ``LOAD_FAST_AND_CLEAR`` that is only
used
-in comprehensions. It has dynamic semantics that Numba cannot model.
-
-For example,
-
-.. code-block:: python
-
- def foo():
- if False:
- x = 1
- [x for x in (1,)]
- return x # This return uses undefined variable
-
-The variable `x` is undefined at the return statement. Instead of raising an
-``UnboundLocalError``, Numba will raise a ``TypingError`` at compile time if
an
-undefined variable is used.
-
-However, Numba cannot always detect undefined variables.
-
-For example,
-
-.. code-block:: python
-
- def foo(a):
- [x for x in (0,)]
- if a:
- x = 3 + a
- x += 10
- return x
-
-Calling ``foo(0)`` returns ``10`` instead of raising ``UnboundLocalError``.
-This is because Numba does not track variable liveness at runtime.
-The return value is ``0 + 10`` since Numba zero-initializes undefined
variables.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9330.change.rst
new/numba-0.59.1/docs/upcoming_changes/9330.change.rst
--- old/numba-0.59.0/docs/upcoming_changes/9330.change.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9330.change.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,24 +0,0 @@
-Refactor and remove legacy APIs/testing internals.
-""""""""""""""""""""""""""""""""""""""""""""""""""
-
-A number of internally used functions have been removed to aid with general
-maintenance by reducing the number of ways in which it is possible to invoke
-compilation, specifically:
-
-* ``numba.core.compiler.compile_isolated`` is removed.
-* ``numba.tests.support.TestCase::run_nullary_func`` is removed.
-* ``numba.tests.support.CompilationCache`` is removed.
-
-Additionally, the concept of "nested context" is removed from
-``numba.core.registry.CPUTarget`` along with the implementation details.
-Maintainers of target extensions (those using the
-API in ``numba.core.target_extension`` to extend Numba support to
-custom/synthetic hardware) should note that the same can be deleted from
-target extension implementations of ``numba.core.descriptor.TargetDescriptor``
-if it is present. i.e. the ``nested_context`` method and associated
-implementation details can just be removed from the custom target's
-``TargetDescriptor``.
-
-Further, a bug was discovered, during the refactoring, in the typing of record
-arrays. It materialised that two record types that only differed in their
-mutability could alias, this has now been fixed.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9334.bug_fix.rst
new/numba-0.59.1/docs/upcoming_changes/9334.bug_fix.rst
--- old/numba-0.59.0/docs/upcoming_changes/9334.bug_fix.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9334.bug_fix.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,11 +0,0 @@
-Fix issue with parfor reductions and Python 3.12.
-=================================================
-
-The parfor reduction code has certain expectations on the order of statements
-that it discovers, these are based on the code that previous versions of Numba
-generated. With Python 3.12, one assignment that used to follow the
-reduction operator statement, such as a binop, is now moved to its own basic
-block. This change reorders the set of discovered reduction nodes so that
-this assignment is right after the reduction operator as it was in previous
-Numba versions. This only affects internal parfor reduction code and
-doesn't actually change the Numba IR.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9335.infrastructure.rst
new/numba-0.59.1/docs/upcoming_changes/9335.infrastructure.rst
--- old/numba-0.59.0/docs/upcoming_changes/9335.infrastructure.rst
2024-01-31 18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9335.infrastructure.rst
1970-01-01 01:00:00.000000000 +0100
@@ -1,6 +0,0 @@
-Add validation capability for user generated towncrier ``.rst`` files.
-======================================================================
-
-Added a validation script for user generated towncrier ``.rst`` files.
-The script will run as a part of towncrier Github workflow automatically
-on every PR.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9336.improvement.rst
new/numba-0.59.1/docs/upcoming_changes/9336.improvement.rst
--- old/numba-0.59.0/docs/upcoming_changes/9336.improvement.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9336.improvement.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,11 +0,0 @@
-Add ``TargetLibraryInfo`` pass to CPU LLVM pipeline.
-""""""""""""""""""""""""""""""""""""""""""""""""""""
-
-The ``TargetLibraryInfo`` pass makes sure that the optimisations that take
place
-during call simplification are appropriate for the target, without this the
-target is assumed to be Linux and code will be optimised to produce e.g. math
-symbols that do not exit on Windows. Historically this issue has been avoided
-through the use of Numba internal libraries carrying wrapped symbols, but doing
-so potentially detriments performance. As a result of this change Numba
internal
-libraries are smaller and there is an increase in optimisation opportunity in
-code using ``exp2`` and ``log2`` functions.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9341.new_feature.rst
new/numba-0.59.1/docs/upcoming_changes/9341.new_feature.rst
--- old/numba-0.59.0/docs/upcoming_changes/9341.new_feature.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9341.new_feature.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,5 +0,0 @@
-Add a config variable to enable / disable the llvmlite memory manager
-=====================================================================
-
-A config variable to force enable or disable the llvmlite memory manager is
-added.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9346.deprecation.rst
new/numba-0.59.1/docs/upcoming_changes/9346.deprecation.rst
--- old/numba-0.59.0/docs/upcoming_changes/9346.deprecation.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9346.deprecation.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,8 +0,0 @@
-Explicitly setting ``NUMBA_CAPTURED_ERRORS=old_style`` will raise deprecation
warnings
-======================================================================================
-
-As per deprecation schedule of old-style error-capturing, explicitly setting
-``NUMBA_CAPTURED_ERRORS=old_style`` will raise deprecation warnings.
-This release is the last to use "old_style" as the default.
-Details are documented at
-https://numba.readthedocs.io/en/0.58.1/reference/deprecation.html#deprecation-of-old-style-numba-captured-errors
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/numba-0.59.0/docs/upcoming_changes/9347.improvement.rst
new/numba-0.59.1/docs/upcoming_changes/9347.improvement.rst
--- old/numba-0.59.0/docs/upcoming_changes/9347.improvement.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9347.improvement.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,9 +0,0 @@
-Numba deprecation warning classes are now subclasses of builtin ones
-====================================================================
-
-To help users manage and suppress deprecation warnings from Numba, the
-``NumbaDeprecationWarning`` and ``NumbaPendingDeprecationWarning`` classes are
-now subclasses of the builtin ``DeprecationWarning`` and
-``PendingDeprecationWarning`` respectively. Therefore, warning filters on
-``DeprecationWarning`` and ``PendingDeprecationWarning`` will apply to Numba
-deprecation warnings.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9352.expired.rst
new/numba-0.59.1/docs/upcoming_changes/9352.expired.rst
--- old/numba-0.59.0/docs/upcoming_changes/9352.expired.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9352.expired.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,8 +0,0 @@
-Object mode *fall-back* support has been removed.
-=================================================
-
-As per the deprecation schedule for Numba 0.59.0, support for
-"object mode fall-back" is removed from all Numba ``jit``-family decorators.
-Further, the default for the ``nopython`` key-word argument has been changed to
-``True``, this means that all Numba ``jit``-family decorated functions will now
-compile in ``nopython`` mode by default.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/docs/upcoming_changes/9353.expired.rst
new/numba-0.59.1/docs/upcoming_changes/9353.expired.rst
--- old/numba-0.59.0/docs/upcoming_changes/9353.expired.rst 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/docs/upcoming_changes/9353.expired.rst 1970-01-01
01:00:00.000000000 +0100
@@ -1,6 +0,0 @@
-Removal of deprecated API ``@numba.generated_jit``.
-===================================================
-
-As per the deprecation schedule for 0.59.0, support for
-``@numba.generated_jit`` has been removed. Use of ``@numba.extending.overload``
-and the high-level extension API is recommended as a replacement.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/_version.py
new/numba-0.59.1/numba/_version.py
--- old/numba-0.59.0/numba/_version.py 2024-01-31 18:53:01.989844800 +0100
+++ new/numba-0.59.1/numba/_version.py 2024-03-18 14:32:45.775456000 +0100
@@ -8,11 +8,11 @@
version_json = '''
{
- "date": "2024-01-31T10:35:08-0600",
+ "date": "2024-03-18T08:01:38-0500",
"dirty": false,
"error": null,
- "full-revisionid": "c6da26929c7fc18b06dd2e5fc86982cea828df6c",
- "version": "0.59.0"
+ "full-revisionid": "9ce83ef5c35d7f68a547bf2fd1266b9a88d3a00d",
+ "version": "0.59.1"
}
''' # END VERSION_JSON
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/core/ir.py
new/numba-0.59.1/numba/core/ir.py
--- old/numba-0.59.0/numba/core/ir.py 2024-01-31 18:52:54.000000000 +0100
+++ new/numba-0.59.1/numba/core/ir.py 2024-03-18 14:32:37.000000000 +0100
@@ -26,7 +26,7 @@
"""Source location
"""
- _defmatcher = re.compile(r'def\s+(\w+)\(.*')
+ _defmatcher = re.compile(r'def\s+(\w+)')
def __init__(self, filename, line, col=None, maybe_decorator=False):
""" Arguments:
@@ -82,10 +82,11 @@
def _raw_function_name(self):
defn = self._find_definition()
if defn:
- return self._defmatcher.match(defn.strip()).groups()[0]
- else:
- # Probably exec(<string>) or REPL.
- return None
+ m = self._defmatcher.match(defn.strip())
+ if m:
+ return m.groups()[0]
+ # Probably exec(<string>) or REPL.
+ return None
def get_lines(self):
if self.lines is None:
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/core/ssa.py
new/numba-0.59.1/numba/core/ssa.py
--- old/numba-0.59.0/numba/core/ssa.py 2024-01-31 18:52:54.000000000 +0100
+++ new/numba-0.59.1/numba/core/ssa.py 2024-03-18 14:32:37.000000000 +0100
@@ -17,6 +17,7 @@
from numba import config
from numba.core import ir, ir_utils, errors
+from numba.core.utils import OrderedSet
from numba.core.analysis import compute_cfg_from_blocks
@@ -154,8 +155,10 @@
states = dict(defs=defs, uses=uses)
_run_block_analysis(blocks, states, _GatherDefsHandler())
_logger.debug("defs %s", pformat(defs))
- # Gather violators by number of definitions
- violators = {k for k, vs in defs.items() if len(vs) > 1}
+ # Gather violators by number of definitions.
+ # The violators are added by the order that they are seen and the algorithm
+ # scan from the first to the last basic-block as they occur in bytecode.
+ violators = OrderedSet([k for k, vs in defs.items() if len(vs) > 1])
# Gather violators by uses not dominated by the one def
doms = cfg.dominators()
for k, use_blocks in uses.items():
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/core/utils.py
new/numba-0.59.1/numba/core/utils.py
--- old/numba-0.59.0/numba/core/utils.py 2024-01-31 18:52:54.000000000
+0100
+++ new/numba-0.59.1/numba/core/utils.py 2024-03-18 14:32:37.000000000
+0100
@@ -385,6 +385,28 @@
T = _tp.TypeVar('T')
+class OrderedSet(MutableSet[T]):
+
+ def __init__(self, iterable: _tp.Iterable[T] = ()):
+ # Just uses a dictionary under-the-hood to maintain insertion order.
+ self._data = dict.fromkeys(iterable, None)
+
+ def __contains__(self, key):
+ return key in self._data
+
+ def __iter__(self):
+ return iter(self._data)
+
+ def __len__(self):
+ return len(self._data)
+
+ def add(self, item):
+ self._data[item] = None
+
+ def discard(self, item):
+ self._data.pop(item, None)
+
+
class MutableSortedSet(MutableSet[T], _tp.Generic[T]):
"""Mutable Sorted Set
"""
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/cuda/cudadrv/nvvm.py
new/numba-0.59.1/numba/cuda/cudadrv/nvvm.py
--- old/numba-0.59.0/numba/cuda/cudadrv/nvvm.py 2024-01-31 18:52:54.000000000
+0100
+++ new/numba-0.59.1/numba/cuda/cudadrv/nvvm.py 2024-03-18 14:32:37.000000000
+0100
@@ -329,6 +329,7 @@
(12, 0): ((5, 0), (9, 0)),
(12, 1): ((5, 0), (9, 0)),
(12, 2): ((5, 0), (9, 0)),
+ (12, 3): ((5, 0), (9, 0)),
}
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/cuda/dispatcher.py
new/numba-0.59.1/numba/cuda/dispatcher.py
--- old/numba-0.59.0/numba/cuda/dispatcher.py 2024-01-31 18:52:54.000000000
+0100
+++ new/numba-0.59.1/numba/cuda/dispatcher.py 2024-03-18 14:32:37.000000000
+0100
@@ -564,6 +564,14 @@
"""
_impl_class = CUDACacheImpl
+ def load_overload(self, sig, target_context):
+ # Loading an overload refreshes the context to ensure it is
+ # initialized. To initialize the correct (i.e. CUDA) target, we need to
+ # enforce that the current target is the CUDA target.
+ from numba.core.target_extension import target_override
+ with target_override('cuda'):
+ return super().load_overload(sig, target_context)
+
class CUDADispatcher(Dispatcher, serialize.ReduceMixin):
'''
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/cuda/tests/cudapy/test_caching.py
new/numba-0.59.1/numba/cuda/tests/cudapy/test_caching.py
--- old/numba-0.59.0/numba/cuda/tests/cudapy/test_caching.py 2024-01-31
18:52:54.000000000 +0100
+++ new/numba-0.59.1/numba/cuda/tests/cudapy/test_caching.py 2024-03-18
14:32:37.000000000 +0100
@@ -1,6 +1,8 @@
import multiprocessing
import os
import shutil
+import subprocess
+import sys
import unittest
import warnings
@@ -170,6 +172,37 @@
# Check the code runs ok from another process
self.run_in_separate_process()
+ @skip_unless_cc_60
+ @skip_if_cudadevrt_missing
+ @skip_if_mvc_enabled('CG not supported with MVC')
+ def test_cache_cg_clean_run(self):
+ # See Issue #9432: https://github.com/numba/numba/issues/9432
+ # If a cached function using CG sync was the first thing to compile,
+ # the compile would fail.
+ self.check_pycache(0)
+
+ # This logic is modelled on run_in_separate_process(), but executes the
+ # CG usecase directly in the subprocess.
+ code = """if 1:
+ import sys
+
+ sys.path.insert(0, %(tempdir)r)
+ mod = __import__(%(modname)r)
+ mod.cg_usecase(0)
+ """ % dict(tempdir=self.tempdir, modname=self.modname)
+
+ popen = subprocess.Popen([sys.executable, "-c", code],
+ stdout=subprocess.PIPE,
+ stderr=subprocess.PIPE)
+ out, err = popen.communicate(timeout=60)
+ if popen.returncode != 0:
+ raise AssertionError(
+ "process failed with code %s: \n"
+ "stdout follows\n%s\n"
+ "stderr follows\n%s\n"
+ % (popen.returncode, out.decode(), err.decode()),
+ )
+
def _test_pycache_fallback(self):
"""
With a disabled __pycache__, test there is a working fallback
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/np/arraymath.py
new/numba-0.59.1/numba/np/arraymath.py
--- old/numba-0.59.0/numba/np/arraymath.py 2024-01-31 18:52:54.000000000
+0100
+++ new/numba-0.59.1/numba/np/arraymath.py 2024-03-18 14:32:37.000000000
+0100
@@ -3735,46 +3735,38 @@
return a < b
-def _searchsorted(func_1, func_2):
- # a facsimile of:
- #
https://github.com/numpy/numpy/blob/4f84d719657eb455a35fcdf9e75b83eb1f97024a/numpy/core/src/npysort/binsearch.cpp#L61
# noqa: E501
-
- def impl(a, v):
- min_idx = 0
- max_idx = len(a)
+@register_jitable
+def _less_then_datetime64(a, b):
+ # Original numpy code is at:
+ #
https://github.com/numpy/numpy/blob/3dad50936a8dc534a81a545365f69ee9ab162ffe/numpy/_core/src/npysort/npysort_common.h#L334-L346
+ if np.isnat(a):
+ return 0
- out = np.empty(v.size, np.intp)
+ if np.isnat(b):
+ return 1
- last_key_val = v.flat[0]
-
- for i in range(v.size):
- key_val = v.flat[i]
-
- if func_1(last_key_val, key_val):
- max_idx = len(a)
- else:
- min_idx = 0
- if max_idx < len(a):
- max_idx += 1
- else:
- max_idx = len(a)
-
- last_key_val = key_val
+ return a < b
- while min_idx < max_idx:
- # to avoid overflow
- mid_idx = min_idx + ((max_idx - min_idx) >> 1)
- mid_val = a[mid_idx]
+@register_jitable
+def _less_then_or_equal_datetime64(a, b):
+ return not _less_then_datetime64(b, a)
- if func_2(mid_val, key_val):
- min_idx = mid_idx + 1
- else:
- max_idx = mid_idx
- out[i] = min_idx
+def _searchsorted(cmp):
+ # a facsimile of:
+ #
https://github.com/numpy/numpy/blob/4f84d719657eb455a35fcdf9e75b83eb1f97024a/numpy/core/src/npysort/binsearch.cpp#L61
# noqa: E501
- return out.reshape(v.shape)
+ def impl(a, key_val, min_idx, max_idx):
+ while min_idx < max_idx:
+ # to avoid overflow
+ mid_idx = min_idx + ((max_idx - min_idx) >> 1)
+ mid_val = a[mid_idx]
+ if cmp(mid_val, key_val):
+ min_idx = mid_idx + 1
+ else:
+ max_idx = mid_idx
+ return min_idx, max_idx
return impl
@@ -3785,21 +3777,29 @@
def make_searchsorted_implementation(np_dtype, side):
assert side in VALID_SEARCHSORTED_SIDES
- lt = _less_than
- le = _less_than_or_equal
+ if np_dtype.char in 'mM':
+ # is datetime
+ lt = _less_then_datetime64
+ le = _less_then_or_equal_datetime64
+ else:
+ lt = _less_than
+ le = _less_than_or_equal
if side == 'left':
- _impl = _searchsorted(lt, lt)
+ _impl = _searchsorted(lt)
+ _cmp = lt
else:
if np.issubdtype(np_dtype, np.inexact) and numpy_version < (1, 23):
# change in behaviour for inexact types
# introduced by:
# https://github.com/numpy/numpy/pull/21867
- _impl = _searchsorted(lt, le)
+ _impl = _searchsorted(le)
+ _cmp = lt
else:
- _impl = _searchsorted(le, le)
+ _impl = _searchsorted(le)
+ _cmp = le
- return register_jitable(_impl)
+ return register_jitable(_impl), register_jitable(_cmp)
@overload(np.searchsorted)
@@ -3817,18 +3817,40 @@
v_dt = as_dtype(v)
np_dt = np.promote_types(as_dtype(a.dtype), v_dt)
- _impl = make_searchsorted_implementation(np_dt, side_val)
+ _impl, _cmp = make_searchsorted_implementation(np_dt, side_val)
if isinstance(v, types.Array):
def impl(a, v, side='left'):
- return _impl(a, v)
+ out = np.empty(v.size, dtype=np.intp)
+ last_key_val = v.flat[0]
+ min_idx = 0
+ max_idx = len(a)
+
+ for i in range(v.size):
+ key_val = v.flat[i]
+
+ if _cmp(last_key_val, key_val):
+ max_idx = len(a)
+ else:
+ min_idx = 0
+ if max_idx < len(a):
+ max_idx += 1
+ else:
+ max_idx = len(a)
+
+ last_key_val = key_val
+ min_idx, max_idx = _impl(a, key_val, min_idx, max_idx)
+ out[i] = min_idx
+
+ return out.reshape(v.shape)
elif isinstance(v, types.Sequence):
def impl(a, v, side='left'):
- return _impl(a, np.array(v))
+ v = np.asarray(v)
+ return np.searchsorted(a, v, side=side)
else: # presumably `v` is scalar
def impl(a, v, side='left'):
- return _impl(a, np.array([v]))[0]
-
+ r, _ = _impl(a, v, 0, len(a))
+ return r
return impl
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/tests/parfor_iss9490_usecase.py
new/numba-0.59.1/numba/tests/parfor_iss9490_usecase.py
--- old/numba-0.59.0/numba/tests/parfor_iss9490_usecase.py 1970-01-01
01:00:00.000000000 +0100
+++ new/numba-0.59.1/numba/tests/parfor_iss9490_usecase.py 2024-03-18
14:32:37.000000000 +0100
@@ -0,0 +1,77 @@
+"""
+This is a testcase for https://github.com/numba/numba/issues/9490.
+The bug is very sensitive to the control-flow and variable uses.
+It is impossible to shrink the reproducer in any meaningful way.
+
+The test is also sensitive to PYTHONHASHSEED.
+PYTHONHASHSEED=1 will trigger the bug.
+
+Example of traceback:
+
+ File "/numba/parfors/parfor.py", line 2070, in _arrayexpr_to_parfor
+ index_vars, loopnests = _mk_parfor_loops(pass_states.typemap, size_vars,
+ scope, loc)
+ File "/numba/parfors/parfor.py", line 1981, in _mk_parfor_loops
+ for size_var in size_vars:
+TypeError: Failed in nopython mode pipeline (step: convert to parfors)
+'NoneType' object is not iterable
+"""
+
+import numba
+import numpy as np
+
+
+@numba.jit(nopython=True, parallel=True)
+def stable_fit(X, y, threshold=3):
+ min_obs = int(X.shape[1] * 1.5)
+ beta = np.zeros((X.shape[1], y.shape[1]), dtype=np.float64)
+ residuals = np.full_like(y, np.nan)
+ stable = np.empty((y.shape[1]))
+ for idx in numba.prange(y.shape[1]):
+ y_sub = y[:, idx]
+ isna = np.isnan(y_sub)
+ X_sub = X[~isna]
+ y_sub = y_sub[~isna]
+ is_stable = False
+
+ # Run until minimum observations
+ # or until stability is reached
+ for jdx in range(len(y_sub), min_obs - 1, -2):
+ # Timeseries gets reduced by two elements
+ # each iteration
+ y_ = y_sub[-jdx:]
+ X_ = X_sub[-jdx:]
+ beta_sub = np.linalg.solve(np.dot(X_.T, X_), np.dot(X_.T, y_))
+ resid_sub = np.dot(X_, beta_sub) - y_
+ # Check for stability
+ rmse = np.sqrt(np.mean(resid_sub ** 2))
+ first = np.fabs(resid_sub[0]) / rmse < threshold
+ last = np.fabs(resid_sub[-1]) / rmse < threshold
+ # Break if stability is reached
+ is_stable = first & last
+ if is_stable:
+ break
+
+ beta[:, idx] = beta_sub
+ residuals[-jdx:, idx] = resid_sub
+ stable[idx] = is_stable
+ return beta, residuals, stable.astype(np.bool_)
+
+
+def check():
+ np.random.seed(0)
+ X_shape = (10, 4)
+ y_shape = (10, 5)
+ X = np.random.random(X_shape)
+ y = np.random.random(y_shape)
+
+ got_beta, got_residuals, got_stable = stable_fit(X, y)
+ exp_beta, exp_residuals, exp_stable = stable_fit.py_func(X, y)
+
+ np.testing.assert_allclose(got_beta, exp_beta)
+ np.testing.assert_allclose(got_residuals, exp_residuals)
+ np.testing.assert_allclose(got_stable, exp_stable)
+
+
+if __name__ == "__main__":
+ check()
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/tests/support.py
new/numba-0.59.1/numba/tests/support.py
--- old/numba-0.59.0/numba/tests/support.py 2024-01-31 18:52:55.000000000
+0100
+++ new/numba-0.59.1/numba/tests/support.py 2024-03-18 14:32:37.000000000
+0100
@@ -150,6 +150,14 @@
"parameter area' in "
"LowerCall_64SVR4"))
+# LLVM PPC issue.
+# Sample error message:
+# Invalid PPC CTR loop!
+# UNREACHABLE executed at /llvm/lib/Target/PowerPC/PPCCTRLoops.cpp:179!
+skip_ppc64le_invalid_ctr_loop = unittest.skipIf(
+ platform.machine() == 'ppc64le',
+ "Invalid PPC CTR loop")
+
# fenv.h on M1 may have various issues:
# https://github.com/numba/numba/issues/7822#issuecomment-1065356758
_uname = platform.uname()
@@ -549,7 +557,8 @@
_assertNumberEqual(first, second, delta)
def subprocess_test_runner(self, test_module, test_class=None,
- test_name=None, envvars=None, timeout=60):
+ test_name=None, envvars=None, timeout=60,
+ _subproc_test_env="1"):
"""
Runs named unit test(s) as specified in the arguments as:
test_module.test_class.test_name. test_module must always be supplied
@@ -560,7 +569,8 @@
environment variable name (str) -> value (str)
It is most convenient to use this method in conjunction with
@needs_subprocess as the decorator will cause the decorated test to be
- skipped unless the `SUBPROC_TEST` environment variable is set to 1
+ skipped unless the `SUBPROC_TEST` environment variable is set to
+ the same value of ``_subproc_test_env``
(this special environment variable is set by this method such that the
specified test(s) will not be skipped in the subprocess).
@@ -575,7 +585,7 @@
fully_qualified_test = '.'.join(x for x in parts if x is not None)
cmd = [sys.executable, '-m', 'numba.runtests', fully_qualified_test]
env_copy = os.environ.copy()
- env_copy['SUBPROC_TEST'] = '1'
+ env_copy['SUBPROC_TEST'] = _subproc_test_env
try:
env_copy['COVERAGE_PROCESS_START'] = os.environ['COVERAGE_RCFILE']
except KeyError:
@@ -601,15 +611,18 @@
subprocess_test_runner."""
def wrapper(func):
def inner(self, *args, **kwargs):
- if os.environ.get("SUBPROC_TEST", None) != "1":
+ if os.environ.get("SUBPROC_TEST", None) != func.__name__:
# Not in a subprocess test env, so stage the call to run
the
# test in a subprocess which will set the env var.
class_name = self.__class__.__name__
- self.subprocess_test_runner(test_module=self.__module__,
- test_class=class_name,
- test_name=func.__name__,
- timeout=timeout,
- envvars=envvars,)
+ self.subprocess_test_runner(
+ test_module=self.__module__,
+ test_class=class_name,
+ test_name=func.__name__,
+ timeout=timeout,
+ envvars=envvars,
+ _subproc_test_env=func.__name__,
+ )
else:
# env var is set, so we're in the subprocess, run the
# actual test.
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/tests/test_exceptions.py
new/numba-0.59.1/numba/tests/test_exceptions.py
--- old/numba-0.59.0/numba/tests/test_exceptions.py 2024-01-31
18:52:55.000000000 +0100
+++ new/numba-0.59.1/numba/tests/test_exceptions.py 2024-03-18
14:32:37.000000000 +0100
@@ -3,7 +3,7 @@
import traceback
from numba import jit, njit
-from numba.core import types, errors
+from numba.core import types, errors, utils
from numba.tests.support import (TestCase, expected_failure_py311,
expected_failure_py312,
)
@@ -309,9 +309,18 @@
# issue #3428
simple_raise = "def f(a):\n raise exc('msg', 10)"
assert_raise = "def f(a):\n assert a != 1"
- for f_text, exc in [(assert_raise, AssertionError),
- (simple_raise, UDEArgsToSuper),
- (simple_raise, UDENoArgSuper)]:
+ py312_pep695_raise = "def f[T: int](a: T) -> T:\n assert a != 1"
+ py312_pep695_raise_2 = "def f[T: int\n](a: T) -> T:\n assert a != 1"
+ test_cases = [
+ (assert_raise, AssertionError),
+ (simple_raise, UDEArgsToSuper),
+ (simple_raise, UDENoArgSuper),
+ ]
+ if utils.PYVERSION >= (3, 12):
+ # Added for https://github.com/numba/numba/issues/9443
+ test_cases.append((py312_pep695_raise, AssertionError))
+ test_cases.append((py312_pep695_raise_2, AssertionError))
+ for f_text, exc in test_cases:
loc = {}
exec(f_text, {'exc': exc}, loc)
pyfunc = loc['f']
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/tests/test_npdatetime.py
new/numba-0.59.1/numba/tests/test_npdatetime.py
--- old/numba-0.59.0/numba/tests/test_npdatetime.py 2024-01-31
18:52:55.000000000 +0100
+++ new/numba-0.59.1/numba/tests/test_npdatetime.py 2024-03-18
14:32:37.000000000 +0100
@@ -971,21 +971,7 @@
def test_sub_td_no_match(self):
self._test_add_sub_td_no_match(np.subtract)
- def _test_min_max(self, operation, parallel, method):
- if method:
- if operation is np.min:
- def impl(arr):
- return arr.min()
- else:
- def impl(arr):
- return arr.max()
- else:
- def impl(arr):
- return operation(arr)
-
- py_func = impl
- cfunc = njit(parallel=parallel)(impl)
-
+ def _get_testcases(self):
test_cases = [
np.array([
DT(0, "ns"),
@@ -1067,7 +1053,24 @@
TD("NaT", "ns"),
]),
]
+ return test_cases
+
+ def _test_min_max(self, operation, parallel, method):
+ if method:
+ if operation is np.min:
+ def impl(arr):
+ return arr.min()
+ else:
+ def impl(arr):
+ return arr.max()
+ else:
+ def impl(arr):
+ return operation(arr)
+
+ py_func = impl
+ cfunc = njit(parallel=parallel)(impl)
+ test_cases = self._get_testcases()
for arr in test_cases:
py_res = py_func(arr)
c_res = cfunc(arr)
@@ -1111,6 +1114,31 @@
def test_max_method_parallel(self):
self._test_min_max(np.max, True, True)
+ def test_searchsorted_datetime(self):
+ from .test_np_functions import (
+ searchsorted, searchsorted_left, searchsorted_right,
+ )
+ pyfunc_list = [searchsorted, searchsorted_left, searchsorted_right]
+ cfunc_list = [jit(fn) for fn in pyfunc_list]
+
+ def check(pyfunc, cfunc, a, v):
+ expected = pyfunc(a, v)
+ got = cfunc(a, v)
+ self.assertPreciseEqual(expected, got)
+
+ cases = self._get_testcases()
+ for pyfunc, cfunc in zip(pyfunc_list, cfunc_list):
+ for arr in cases:
+ arr = np.sort(arr)
+ for n in range(1, min(3, arr.size) + 1):
+ idx = np.random.randint(0, arr.size, n)
+ vs = arr[idx]
+ if n == 1:
+ [v] = vs
+ check(pyfunc, cfunc, arr, v)
+ check(pyfunc, cfunc, arr, vs)
+
+
if __name__ == '__main__':
unittest.main()
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/tests/test_parfors.py
new/numba-0.59.1/numba/tests/test_parfors.py
--- old/numba-0.59.0/numba/tests/test_parfors.py 2024-01-31
18:52:55.000000000 +0100
+++ new/numba-0.59.1/numba/tests/test_parfors.py 2024-03-18
14:32:37.000000000 +0100
@@ -44,7 +44,8 @@
override_env_config, linux_only, tag,
skip_parfors_unsupported, _32bit, needs_blas,
needs_lapack, disabled_test,
skip_unless_scipy,
- needs_subprocess)
+ needs_subprocess,
+ skip_ppc64le_invalid_ctr_loop)
from numba.core.extending import register_jitable
from numba.core.bytecode import _fix_LOAD_GLOBAL_arg
from numba.core import utils
@@ -2368,6 +2369,25 @@
self.assertEqual(expected, njit(parallel=False)(def_in_loop)(4))
self.assertEqual(expected, njit(parallel=True)(def_in_loop)(4))
+ @needs_lapack # use of np.linalg.solve
+ @skip_ppc64le_invalid_ctr_loop
+ def test_issue9490_non_det_ssa_problem(self):
+ cmd = [
+ sys.executable,
+ "-m",
+ "numba.tests.parfor_iss9490_usecase",
+ ]
+
+ try:
+ subp.check_output(cmd, env={**os.environ,
+ "PYTHONHASHSEED": "1",
+ "NUMBA_NUM_THREADS": "1"},
+ stderr=subp.STDOUT,
+ encoding='utf-8')
+ except subp.CalledProcessError as e:
+ msg = f"subprocess failed with output:\n{e.output}"
+ self.fail(msg=msg)
+
@skip_parfors_unsupported
class TestParforsLeaks(MemoryLeakMixin, TestParforsBase):
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba/tests/test_unicode.py
new/numba-0.59.1/numba/tests/test_unicode.py
--- old/numba-0.59.0/numba/tests/test_unicode.py 2024-01-31
18:52:55.000000000 +0100
+++ new/numba-0.59.1/numba/tests/test_unicode.py 2024-03-18
14:32:37.000000000 +0100
@@ -2667,11 +2667,11 @@
"""
# requires formatting (FORMAT_VALUE) and concatenation (BUILD_STRINGS)
def impl1(a):
- return f"AA_{a+3}_B"
+ return f"AA_{a + 3}_B"
# does not require concatenation
def impl2(a):
- return f"{a+2}"
+ return f"{a + 2}"
# no expression
def impl3(a):
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba.egg-info/PKG-INFO
new/numba-0.59.1/numba.egg-info/PKG-INFO
--- old/numba-0.59.0/numba.egg-info/PKG-INFO 2024-01-31 18:53:01.000000000
+0100
+++ new/numba-0.59.1/numba.egg-info/PKG-INFO 2024-03-18 14:32:45.000000000
+0100
@@ -1,6 +1,6 @@
Metadata-Version: 2.1
Name: numba
-Version: 0.59.0
+Version: 0.59.1
Summary: compiling Python code using LLVM
Home-page: https://numba.pydata.org
License: BSD
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/numba-0.59.0/numba.egg-info/SOURCES.txt
new/numba-0.59.1/numba.egg-info/SOURCES.txt
--- old/numba-0.59.0/numba.egg-info/SOURCES.txt 2024-01-31 18:53:01.000000000
+0100
+++ new/numba-0.59.1/numba.egg-info/SOURCES.txt 2024-03-18 14:32:45.000000000
+0100
@@ -106,6 +106,7 @@
docs/source/release/0.58.0-notes.rst
docs/source/release/0.58.1-notes.rst
docs/source/release/0.59.0-notes.rst
+docs/source/release/0.59.1-notes.rst
docs/source/user/5minguide.rst
docs/source/user/cfunc.rst
docs/source/user/cli.rst
@@ -126,39 +127,6 @@
docs/source/user/troubleshoot.rst
docs/source/user/vectorize.rst
docs/source/user/withobjmode.rst
-docs/upcoming_changes/9048.bug_fix.rst
-docs/upcoming_changes/9049.bug_fix.rst
-docs/upcoming_changes/9089.bug_fix.rst
-docs/upcoming_changes/9123.new_feature.rst
-docs/upcoming_changes/9126.np_support.rst
-docs/upcoming_changes/9140.np_support.rst
-docs/upcoming_changes/9141.np_support.rst
-docs/upcoming_changes/9144.bug_fix.rst
-docs/upcoming_changes/9154.np_support.rst
-docs/upcoming_changes/9169.bug_fix.rst
-docs/upcoming_changes/9189.bug_fix.rst
-docs/upcoming_changes/9223.cuda.rst
-docs/upcoming_changes/9235.cuda.rst
-docs/upcoming_changes/9242.bug_fix.rst
-docs/upcoming_changes/9244.bug_fix.rst
-docs/upcoming_changes/9245.performance.rst
-docs/upcoming_changes/9246.highlight.rst
-docs/upcoming_changes/9249.np_support.rst
-docs/upcoming_changes/9267.cuda.rst
-docs/upcoming_changes/9279.cuda.rst
-docs/upcoming_changes/9296.bug_fix.rst
-docs/upcoming_changes/9309.change.rst
-docs/upcoming_changes/9310.highlight.rst
-docs/upcoming_changes/9315.change.rst
-docs/upcoming_changes/9330.change.rst
-docs/upcoming_changes/9334.bug_fix.rst
-docs/upcoming_changes/9335.infrastructure.rst
-docs/upcoming_changes/9336.improvement.rst
-docs/upcoming_changes/9341.new_feature.rst
-docs/upcoming_changes/9346.deprecation.rst
-docs/upcoming_changes/9347.improvement.rst
-docs/upcoming_changes/9352.expired.rst
-docs/upcoming_changes/9353.expired.rst
docs/upcoming_changes/README.rst
docs/upcoming_changes/template.rst
numba/__init__.py
@@ -695,6 +663,7 @@
numba/tests/matmul_usecase.py
numba/tests/orphaned_semaphore_usecase.py
numba/tests/overload_usecases.py
+numba/tests/parfor_iss9490_usecase.py
numba/tests/parfors_cache_usecases.py
numba/tests/parfors_max_label_error.py
numba/tests/pdlike_usecase.py
