Updated graphics/libjpeg-turbo, devel/blosc2

(adam)

blosc2: updated to 2.13.0

Changes from 2.12.0 to 2.13.0

* Added a new BLOSC_FILTER_INT_TRUNC filter for truncating integers to a
  given number of bits.  This is useful for compressing integers that are
  not using all the bits of the type.

* Optimized zstd, specially when using dicts.

* Initialize grok library when loading the plugin. This is needed for other plugins
  to be able to use it without the need of importing the package.

(adam)

libjpeg-turbo: updated to 3.0.2

3.0.2
=====

Significant changes relative to 3.0.1:

1. Fixed a signed integer overflow in the `tj3CompressFromYUV8()`,
`tj3DecodeYUV8()`, `tj3DecompressToYUV8()`, and `tj3EncodeYUV8()` functions,
detected by the Clang and GCC undefined behavior sanitizers, that could be
triggered by setting the `align` parameter to an unreasonably large value.
This issue did not pose a security threat, but removing the warning made it
easier to detect actual security issues, should they arise in the future.

2. Introduced a new parameter (`TJPARAM_MAXMEMORY` in the TurboJPEG C API and
`TJ.PARAM_MAXMEMORY` in the TurboJPEG Java API) and a corresponding TJBench
option (`-maxmemory`) for specifying the maximum amount of memory (in
megabytes) that will be allocated for intermediate buffers, which are used with
progressive JPEG compression and decompression, optimized baseline entropy
coding, lossless JPEG compression, and lossless transformation.  The new
parameter and option serve the same purpose as the `max_memory_to_use` field in
the `jpeg_memory_mgr` struct in the libjpeg API, the `JPEGMEM` environment
variable, and the cjpeg/djpeg/jpegtran `-maxmemory` option.

3. Introduced a new parameter (`TJPARAM_MAXPIXELS` in the TurboJPEG C API and
`TJ.PARAM_MAXPIXELS` in the TurboJPEG Java API) and a corresponding TJBench
option (`-maxpixels`) for specifying the maximum number of pixels that the
decompression, lossless transformation, and packed-pixel image loading
functions/methods will process.

4. Fixed an error ("Unsupported color conversion request") that occurred when
attempting to decompress a 3-component lossless JPEG image without an Adobe
APP14 marker.  The decompressor now assumes that a 3-component lossless JPEG
image without an Adobe APP14 marker uses the RGB colorspace if its component
IDs are 1, 2, and 3.

(adam)

Added time/py-backports-datetime-fromisoformat; Updated multimedia/py-m3u8

(adam)

py-m3u8: updated to 4.0.0

4.0.0

Remove class for UTC compat for Python 2
Use built-in ISO 8601 parse for Python 3.11 and later
Run pyupgrade for 3.7
Upgrade actions to latest stable version
Add 3.12 to the list of versions to test
Pass timespec all the way down in dumps
Ruff cleanup
Run format with ruff

(adam)

py-backports-datetime-fromisoformat: added version 2.0.1

A backport of Python 3.11's datetime.fromisoformat methods to earlier versions
of Python 3.

(adam)

attr: add --sysconfdir

(wiz)

doc: add upper bound for minizip vuln

(wiz)

doc: Updated archivers/minizip to 1.3.1

(wiz)

minizip: update to 1.3.1.

Fixes a heap-overflow.

(wiz)

doc/TODO: + minizip-1.3.1, zlib-1.3.1.

(wiz)

biology/libpll: only works on Linux

(thor)

hfstospell: replace patch that forced c++14 with FORCE_CXX_STD

Patch didn't apply any longer after autoconf update

(wiz)

py-ghp-import: add missing tools

(wiz)

libgltf: needs c++11 now due to boost

Fix HOMEPAGE while here.

(wiz)

breezy: add missing tools

(wiz)

libpll: fix pkglint

(wiz)

libppl: fix unportable test operator

(wiz)

libppl: fix buildlink3 file

(wiz)

py-pandas: needs at least gcc 10 because of __has_builtin

(wiz)

py-pbr: Add missing tools

(wiz)

doc: Updated math/py-numba to 0.58.1

(thor)

math/py-numba: update to 0.58.1

This is the first version I tested with the re-vived py-llvmlite. This version
works with Pythons below 3.12 so far. Upstream changes since 0.55.2:

Version 0.58.1 (17 October 2023)

This is a maintenance release that adds support for NumPy 1.26 and fixes a bug.
NumPy Support
Support NumPy 1.26

Support for NumPy 1.26 is added.

(PR-#9227)
Bug Fixes
Fixed handling of float default arguments in inline closures

Float default arguments in inline closures would produce incorrect results since updates for Python 3.11 - these are now handled correctly again.

(PR-#9222)
Pull-Requests

    PR #9220: Support passing arbitrary flags to NVVM (gmarkall)

    PR #9227: Support NumPy 1.26 (PR aimed at review / merge) (Tialo gmarkall)

    PR #9228: Fix #9222 - Don’t replace . with _ in func arg names in inline closures (gmarkall)

Authors

    gmarkall

    Tialo

Version 0.58.0 (20 September 2023)

Table of Contents

    Version 0.58.0 (20 September 2023)

        Highlights

        New Features

        Improvements

        NumPy Support

        CUDA Changes

        Bug Fixes

        Changes

        Deprecations

        Pull-Requests

        Authors

This is a major Numba release. Numba now uses towncrier to create the release notes, so please find a summary of all noteworthy items below.
Highlights
Added towncrier

This PR adds towncrier as a GitHub workflow for checking release notes. From this PR onwards every PR made in Numba will require a appropriate release note associated with it. The reviewer may decide to skip adding release notes in smaller PRs with minimal impact by addition of a skip_release_notes label to the PR.

(PR-#8792)
The minimum supported NumPy version is 1.22.

Following NEP-0029, the minimum supported NumPy version is now 1.22.

(PR-#9093)
Add support for NumPy 1.25

Extend Numba to support new and changed features released in NumPy 1.25.

(PR-#9011)
Remove NVVM 3.4 and CTK 11.0 / 11.1 support

Support for CUDA toolkits < 11.2 is removed.

(PR-#9040)
Removal of Windows 32-bit Support

This release onwards, Numba has discontinued support for Windows 32-bit operating systems.

(PR-#9083)
The minimum llvmlite version is now 0.41.0.

The minimum required version of llvmlite is now version 0.41.0.

(PR-#8916)
Added RVSDG-frontend

This PR is a preliminary work on adding a RVSDG-frontend for processing bytecode. RVSDG (Regionalized Value-State Dependence Graph) allows us to have a dataflow-centric view instead of a traditional SSA-CFG view. This allows us to simplify the compiler in the future.

(PR-#9012)
New Features
numba.experimental.jitclass gains support for __*matmul__ methods.

numba.experimental.jitclass now has support for the following methods:

    __matmul__

    __imatmul__

    __rmatmul__

(PR-#8892)
numba.experimental.jitclass gains support for reflected “dunder” methods.

numba.experimental.jitclass now has support for the following methods:

    __radd__

    __rand_

    __rfloordiv__

    __rlshift__

    __ror_

    __rmod_

    __rmul_

    __rpow_

    __rrshift_

    __rsub_

    __rtruediv_

    __rxor_

(PR-#8906)
Add support for value max to NUMBA_OPT.

The optimisation level that Numba applies when compiling can be set through the environment variable NUMBA_OPT. This has historically been a value between 0 and 3 (inclusive). Support for the value max has now been added, this is a Numba-specific optimisation level which indicates that the user would like Numba to try running the most optimisation possible, potentially trading a longer compilation time for better run-time performance. In practice, use of the max level of optimisation may or may not benefit the run-time or compile-time performance of user code, but it has been added to present an easy to access option for users to try if they so wish.

(PR-#9094)
Improvements
Updates to numba.core.pythonapi.

Support for Python C-API functions PyBytes_AsString and PyBytes_AsStringAndSize is added to numba.core.pythonapi.PythonAPI as bytes_as_string and bytes_as_string_and_size methods respectively.

(PR-#8462)
Support for isinstance is now non-experimental.

Support for the isinstance built-in function has moved from being considered an experimental feature to a fully supported feature.

(PR-#8911)
NumPy Support
All modes are supported in numpy.correlate and numpy.convolve.

All values for the mode argument to numpy.correlate and numpy.convolve are now supported.

(PR-#7543)
@vectorize accommodates arguments implementing __array_ufunc__.

Universal functions (ufuncs) created with numba.vectorize will now respect arguments implementing __array_ufunc__ (NEP-13) to allow pre- and post-processing of arguments and return values when the ufunc is called from the interpreter.

(PR-#8995)
Added support for np.geomspace function.

This PR improves on #4074 by adding support for np.geomspace. The current implementation only supports scalar start and stop parameters.

(PR-#9068)
Added support for np.vsplit, np.hsplit, np.dsplit.

This PR improves on #4074 by adding support for np.vsplit, np.hsplit, and np.dsplit.

(PR-#9082)
Added support for np.row_stack function.

Support is added for numpy.row_stack.

(PR-#9085)
Added support for functions np.polynomial.polyutils.trimseq, as well as functions polyadd, polysub, polymul from np.polynomial.polynomial.

Support is added for np.polynomial.polyutils.trimseq, np.polynomial.polynomial.polyadd, np.polynomial.polynomial.polysub, np.polynomial.polynomial.polymul.

(PR-#9087)
Added support for np.diagflat function.

Support is added for numpy.diagflat.

(PR-#9113)
Added support for np.resize function.

Support is added for numpy.resize.

(PR-#9118)
Add np.trim_zeros

Support for np.trim_zeros() is added.

(PR-#9074)
CUDA Changes
Bitwise operation ufunc support for the CUDA target.

Support is added for some ufuncs associated with bitwise operation on the CUDA target. Namely:

    numpy.bitwise_and

    numpy.bitwise_or

    numpy.bitwise_not

    numpy.bitwise_xor

    numpy.invert

    numpy.left_shift

    numpy.right_shift

(PR-#8974)
Add support for the latest CUDA driver codes.

Support is added for the latest set of CUDA driver codes.

(PR-#8988)
Add NumPy comparison ufunc in CUDA

this PR adds support for comparison ufuncs for the CUDA target (eg. numpy.greater, numpy.greater_equal, numpy.less_equal, etc.).

(PR-#9007)
Report absolute path of libcuda.so on Linux

numba -s now reports the absolute path to libcuda.so on Linux, to aid troubleshooting driver issues, particularly on WSL2 where a Linux driver can incorrectly be installed in the environment.

(PR-#9034)
Add debuginfo support to nvdisasm output.

Support is added for debuginfo (source line and inlining information) in functions that make calls through nvdisasm. For example the CUDA dispatcher .inspect_sass method output is now augmented with this information.

(PR-#9035)
Add CUDA SASS CFG Support

This PR adds support for getting the SASS CFG in dot language format. It adds an inspect_sass_cfg() method to CUDADispatcher and the -cfg flag to the nvdisasm command line tool.

(PR-#9051)
Support NVRTC using the ctypes binding

NVRTC can now be used when the ctypes binding is in use, enabling float16, and linking CUDA C / C++ sources without needing the NVIDIA CUDA Python bindings.

(PR-#9086)
Fix CUDA atomics tests with toolkit 12.2

CUDA 12.2 generates slightly different PTX for some atomics, so the relevant tests are updated to look for the correct instructions when 12.2 is used.

(PR-#9088)
Bug Fixes
Handling of different sized unsigned integer indexes are fixed in numba.typed.List.

An issue with the order of truncation/extension and casting of unsigned integer indexes in numba.typed.List has been fixed.

(PR-#7262)
Prevent invalid fusion

This PR fixes an issue in which an array first read in a parfor and later written in the same parfor would only be classified as used in the parfor. When a subsequent parfor also used the same array then fusion of the parfors was happening which should have been forbidden given that that the first parfor was also writing to the array. This PR treats such arrays in a parfor as being both used and defined so that fusion will be prevented.

(PR-#7582)
The numpy.allclose implementation now correctly handles default arguments.

The implementation of numpy.allclose is corrected to use TypingError to report typing errors.

(PR-#8885)
Add type validation to numpy.isclose.

Type validation is added to the implementation of numpy.isclose.

(PR-#8944)
Fix support for overloading dispatcher with non-compatible first-class functions

Fixes an error caused by not handling compilation error during casting of Dispatcher objects into first-class functions. With the fix, users can now overload a dispatcher with non-compatible first-class functions. Refer to https://github.com/numba/numba/issues/9071 for details.

(PR-#9072)
Support dtype keyword argument in numpy.arange with parallel=True

Fixes parfors transformation to support the use of dtype keyword argument in numpy.arange(..., dtype=dtype).

(PR-#9095)
Fix all @overloads to use parameter names that match public APIs.

Some of the Numba @overloads for functions in NumPy and Python’s built-ins were written using parameter names that did not match those used in API they were overloading. The result of this being that calling a function with such a mismatch using the parameter names as key-word arguments at the call site would result in a compilation error. This has now been universally fixed throughout the code base and a unit test is running with a best-effort attempt to prevent reintroduction of similar mistakes in the future. Fixed functions include:

From Python built-ins:

    complex

From the Python random module:

    random.seed

    random.gauss

    random.normalvariate

    random.randrange

    random.randint

    random.uniform

    random.shuffle

From the numpy module:

    numpy.argmin

    numpy.argmax

    numpy.array_equal

    numpy.average

    numpy.count_nonzero

    numpy.flip

    numpy.fliplr

    numpy.flipud

    numpy.iinfo

    numpy.isscalar

    numpy.imag

    numpy.real

    numpy.reshape

    numpy.rot90

    numpy.swapaxes

    numpy.union1d

    numpy.unique

From the numpy.linalg module:

    numpy.linalg.norm

    numpy.linalg.cond

    numpy.linalg.matrix_rank

From the numpy.random module:

    numpy.random.beta

    numpy.random.chisquare

    numpy.random.f

    numpy.random.gamma

    numpy.random.hypergeometric

    numpy.random.lognormal

    numpy.random.pareto

    numpy.random.randint

    numpy.random.random_sample

    numpy.random.ranf

    numpy.random.rayleigh

    numpy.random.sample

    numpy.random.shuffle

    numpy.random.standard_gamma

    numpy.random.triangular

    numpy.random.weibull

(PR-#9099)
Changes
Support for @numba.extending.intrinsic(prefer_literal=True)

In the high level extension API, the prefer_literal option is added to the numba.extending.intrinsic decorator to prioritize the use of literal types when available. This has the same behavior as in the prefer_literal option in the numba.extending.overload decorator.

(PR-#6647)
Deprecations
Deprecation of old-style NUMBA_CAPTURED_ERRORS

Added deprecation schedule of NUMBA_CAPTURED_ERRORS=old_style. NUMBA_CAPTURED_ERRORS=new_style will become the default in future releases. Details are documented at https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-old-style-numba-captured-errors

(PR-#9090)
Pull-Requests

    PR #6647: Support prefer_literal option for intrinsic decorator (ashutoshvarma sklam)

    PR #7262: fix order of handling and casting (esc)

    PR #7543: Support for all modes in np.correlate and np.convolve (jeertmans)

    PR #7582: Use get_parfor_writes to detect illegal array access that prevents fusion. (DrTodd13)

    PR #8371: Added binomial distribution (esc kc611)

    PR #8462: Add PyBytes_AsString and PyBytes_AsStringAndSize (ianna)

    PR #8633: DOC: Convert vectorize and guvectorize examples to doctests (Matt711)

    PR #8730: Update dev-docs (sgbaird esc)

    PR #8792: Added towncrier as a github workflow (kc611)

    PR #8854: Updated mk_alloc to support Numba-Dpex compute follows data. (mingjie-intel)

    PR #8861: CUDA: Don’t add device kwarg for jit registry (gmarkall)

    PR #8871: Don’t return the function in CallConv.decorate_function() (gmarkall)

    PR #8885: Fix np.allclose not handling default args (guilhermeleobas)

    PR #8892: Add support for __*matmul__ methods in jitclass (louisamand)

    PR #8895: CUDA: Enable caching functions that use CG (gmarkall)

    PR #8906: Add support for reflected dunder methods in jitclass (louisamand)

    PR #8911: Remove isinstance experimental feature warning (guilhermeleobas)

    PR #8916: Bump llvmlite requirement to 0.41.0dev0 (sklam)

    PR #8925: Update release checklist template (sklam)

    PR #8937: Remove old Website development documentation (esc gmarkall)

    PR #8944: Add exceptions to np.isclose (guilhermeleobas)

    PR #8974: CUDA: Add binary ufunc support (Matt711)

    PR #8976: Fix index URL for ptxcompiler/cubinlinker packages. (bdice)

    PR #8978: Import MVC packages when using MVCLinker. (bdice)

    PR #8983: Fix typo in deprecation.rst (dsgibbons)

    PR #8988: support for latest CUDA driver codes #8363 (s1Sharp)

    PR #8995: Allow libraries that implement __array_ufunc__ to override DUFunc.__c… (jpivarski)

    PR #9007: CUDA: Add comparison ufunc support (Matt711)

    PR #9012: RVSDG-frontend (sklam)

    PR #9021: update the release checklist following 0.57.1rc1 (esc)

    PR #9022: fix: update the C++ ABI repo reference (emmanuel-ferdman)

    PR #9028: Replace use of imp module removed in 3.12 (hauntsaninja)

    PR #9034: CUDA libs test: Report the absolute path of the loaded libcuda.so on Linux, + other improvements (gmarkall)

    PR #9035: CUDA: Allow for debuginfo in nvdisasm output (Matt711)

    PR #9037: Recognize additional functions as being pure or not having side effects. (DrTodd13)

    PR #9039: Correct git clone link in installation instructions. (ellifteria)

    PR #9040: Remove NVVM 3.4 and CTK 11.0 / 11.1 support (gmarkall)

    PR #9046: copy the change log changes for 0.57.1 to main (esc)

    PR #9050: Update CODEOWNERS (sklam)

    PR #9051: Add CUDA CFG support (Matt711)

    PR #9056: adding weekly meeting notes script (esc)

    PR #9068: Adding np.geomspace (KrisMinchev)

    PR #9069: Fix towncrier error due to importlib_resources upgrade (sklam)

    PR #9072: Fix support for overloading dispatcher with non-compatible first-class functions (gmarkall sklam)

    PR #9074: Add np.trim_zeros (sungraek guilhermeleobas)

    PR #9082: Add np.vsplit, np.hsplit, and np.dsplit (KrisMinchev)

    PR #9083: Removed windows 32 references from code and documentation (kc611)

    PR #9085: Add tests for np.row_stack (KrisMinchev)

    PR #9086: Support NVRTC using ctypes binding (testhound gmarkall)

    PR #9087: Add trimseq from np.polynomial.polyutils and polyadd, polysub, polymul from np.polynomial.polynomial (KrisMinchev)

    PR #9088: Fix: Issue 9063 - CUDA atomics tests failing with CUDA 12.2 (gmarkall)

    PR #9090: Add deprecation notice for old_style error capturing. (esc sklam)

    PR #9094: Add support for a ‘max’ level to NUMBA_OPT environment variable. (stuartarchibald)

    PR #9095: Support dtype keyword in arange_parallel_impl (DrTodd13 sklam)

    PR #9105: NumPy 1.25 support (PR #9011) continued (gmarkall apmasell)

    PR #9111: Fixes ReST syntax error in PR#9099 (stuartarchibald gmarkall sklam apmasell)

    PR #9112: Fixups for PR#9100 (stuartarchibald sklam)

    PR #9113: Add support for np.diagflat (KrisMinchev)

    PR #9114: update np min to 122 (stuartarchibald esc)

    PR #9117: Fixed towncrier template rendering (kc611)

    PR #9118: Add support for np.resize() (KrisMinchev)

    PR #9120: Update conda-recipe for numba-rvsdg (sklam)

    PR #9127: Fix accidental cffi test deps, refactor cffi skipping (gmarkall)

    PR #9128: Merge rvsdg_frontend branch to main (esc sklam)

    PR #9152: Fix old_style error capturing deprecation warnings (sklam)

    PR #9159: Fix uncaught exception in find_file() (gmarkall)

    PR #9173: Towncrier fixups (Continue #9158 and retarget to main branch) (sklam)

    PR #9181: Remove extra decrefs in RNG (sklam)

    PR #9190: Fix issue with incompatible multiprocessing context in test. (stuartarchibald)

Authors

    apmasell

    ashutoshvarma

    bdice

    DrTodd13

    dsgibbons

    ellifteria

    emmanuel-ferdman

    esc

    gmarkall

    guilhermeleobas

    hauntsaninja

    ianna

    jeertmans

    jpivarski

    jtilly

    kc611

    KrisMinchev

    louisamand

    Matt711

    mingjie-intel

    s1Sharp

    sgbaird

    sklam

    stuartarchibald

    sungraek

    testhound

Version 0.57.1 (21 June, 2023)

Pull-Requests:

    PR #8964: fix missing nopython keyword in cuda random module (esc)

    PR #8965: fix return dtype for np.angle (guilhermeleobas esc)

    PR #8982: Don’t do the parfor diagnostics pass for the parfor gufunc. (DrTodd13)

    PR #8996: adding a test for 8940 (esc)

    PR #8958: resurrect the import, this time in the registry initialization (esc)

    PR #8947: Introduce internal _isinstance_no_warn (guilhermeleobas esc)

    PR #8998: Fix 8939 (second attempt) (esc)

    PR #8978: Import MVC packages when using MVCLinker. (bdice)

    PR #8895: CUDA: Enable caching functions that use CG (gmarkall)

    PR #8976: Fix index URL for ptxcompiler/cubinlinker packages. (bdice)

    PR #9004: Skip MVC test when libraries unavailable (gmarkall esc)

    PR #9006: link to version support table instead of using explicit versions (esc)

    PR #9005: Fix: Issue #8923 - avoid spurious device-to-host transfers in CUDA ufuncs (gmarkall)

Authors:

    bdice

    DrTodd13

    esc

    gmarkall

Version 0.57.0 (1 May, 2023)

This release continues to add new features, bug fixes and stability improvements to Numba. Please note that this release contains a significant number of both deprecation and pending-deprecation notices with view of making it easier to develop new technology for Numba in the future. Also note that this will be the last release to support Windows 32-bit packages produced by the Numba team.

Highlights of core dependency upgrades:

    Support for Python 3.11 (minimum is moved to 3.8)

    Support for NumPy 1.24 (minimum is moved to 1.21)

Python language support enhancements:

    Exception classes now support arguments that are not compile time constant.

    The built-in functions hasattr and getattr are supported for compile time constant attributes.

    The built-in functions str and repr are now implemented similarly to their Python implementations. Custom __str__ and __repr__ functions can be associated with types and work as expected.

    Numba’s unicode functionality in str.startswith now supports kwargs start and end.

    min and max now support boolean types.

    Support is added for the dict(iterable) constructor.

NumPy features/enhancements:

    The largest set of new features is within the numpy.random.Generator support, the vast majority of commonly used distributions are now supported. Namely:

        Generator.beta

        Generator.chisquare

        Generator.exponential

        Generator.f

        Generator.gamma

        Generator.geometric

        Generator.integers

        Generator.laplace

        Generator.logistic

        Generator.lognormal

        Generator.logseries

        Generator.negative_binomial

        Generator.noncentral_chisquare

        Generator.noncentral_f

        Generator.normal

        Generator.pareto

        Generator.permutation

        Generator.poisson

        Generator.power

        Generator.random

        Generator.rayleigh

        Generator.shuffle

        Generator.standard_cauchy

        Generator.standard_exponential

        Generator.standard_gamma

        Generator.standard_normal

        Generator.standard_t

        Generator.triangular

        Generator.uniform

        Generator.wald

        Generator.weibull

        Generator.zipf

    The nbytes property on NumPy ndarray types is implemented.

    Nesting of nested-array types is now supported.

    datetime and timedelta types can be cast to int.

    F-order iteration is supported in ufunc generation for increased performance when using combinations of predominantly F-order arrays.

    The following functions are also now supported:

        np.argpartition

        np.isclose

        np.nan_to_num

        np.new_axis

        np.union1d

Highlights of core changes:

    A large amount of refactoring has taken place to convert many of Numba’s internal implementations, of both Python and NumPy functions, from the low-level extension API to the high-level extension API (numba.extending).

    The __repr__ method is supported for Numba types.

    The default target for applicable functions in the extension API (numba.extending) is now "generic". This means that @overload* and @intrinsic functions will by default be accepted by both the CPU and CUDA targets.

    The use of __getitem__ on Numba types is now supported in compiled code. i.e. types.float64[:, ::1] is now compilable.

Performance:

    The performance of str.find() and str.rfind() has been improved.

    Unicode support for __getitem__ now avoids allocation and returns a view.

    The numba.typed.Dict dictionary now accepts an n_keys option to enable allocating the dictionary instance to a predetermined initial size (useful to avoid resizes!).

    The Numba Run-time (NRT) has been improved in terms of performance and safety:

        The NRT internal statistics counters are now off by default (removes atomic lock contentions).

        Debug cache line filling is off by default.

        The NRT is only compiled once a compilation starts opposed to at function decoration time, this improves import speed.

        The NRT allocation calls are all made through a “checked” layer by default.

CUDA:

    New NVIDIA hardware and software compatibility / support:

        Toolkits: CUDA 11.8 and 12, with Minor Version Compatibility for 11.x.

        Packaging: NVIDIA-packaged CUDA toolkit conda packages.

        Hardware: Hopper, Ada Lovelace, and AGX Orin.

    float16 support:

        Arithmetic operations are now fully supported.

        A new method, is_fp16_supported(), and device property, supports_float16, for checking the availability of float16 support.

    Functionality:

        The high-level extension API is now fully-supported in the CUDA target.

        Eager compilation of multiple signatures, multiple outputs from generalized ufuncs, and specifying the return type of ufuncs are now supported.

        A limited set of NumPy ufuncs (trigonometric functions) can now be called inside kernels.

    Lineinfo quality improvement: enabling lineinfo no longer results in any changes to generated code.

Deprecations:

    The numba.pycc module and everything in it is now pending deprecation.

    The long awaited full deprecation of object mode fall-back is underway. This change means @jit with no keyword arguments will eventually alias @njit.

    The @generated_jit decorator is deprecated as the Numba extension API provides a better supported superset of the same functionality, particularly through @numba.extending.overload.

Version support/dependency changes:

    The setuptools package is now an optional run-time dependency opposed to a required run-time dependency.

    The TBB threading-layer now requires version 2021.6 or later.

    LLVM 14 is now supported on all platforms via llvmlite.

Pull-Requests:

    PR #5113: Fix error handling in the Interval extending example (esc eric-wieser)

    PR #5544: Add support for np.union1d (shangbol gmarkall)

    PR #7009: Add writable args (dmbelov)

    PR #7067: Implement np.isclose (guilhermeleobas)

    PR #7255: CUDA: Support CUDA Toolkit conda packages from NVIDIA (gmarkall)

    PR #7622: Support fortran loop ordering for ufunc generation (sklam)

    PR #7733: fix for /tmp/tmp access issues (ChiCheng45)

    PR #7884: Implement getattr builtin. (stuartarchibald)

    PR #7885: Adds CUDA FP16 arithmetic operators (testhound)

    PR #7920: Drop pre-3.7 code path (CPU only) (sklam)

    PR #8001: CUDA fp16 math functions (testhound gmarkall)

    PR #8010: Add support for fp16 comparison native operators (testhound)

    PR #8024: Allow converting NumPy datetimes to int (apmasell)

    PR #8038: Support for Numpy BitGenerators PR#2: Standard Distributions support (kc611)

    PR #8040: Support for Numpy BitGenerators PR#3: Advanced Distributions Support. (kc611)

    PR #8041: Support for Numpy BitGenerators PR#4: Generator().integers() Support. (kc611)

    PR #8042: Support for NumPy BitGenerators PR#5: Generator Shuffling Methods. (kc611)

    PR #8061: Migrate random glue_lowering to overload where easy (apmasell)

    PR #8106: Remove injection of atomic JIT functions into NRT memsys. (stuartarchibald)

    PR #8120: Support nesting of nested array types (gmarkall)

    PR #8134: Support non-constant exception values in JIT (guilhermeleobas sklam)

    PR #8147: Adds size variable at runtime for arrays that cannot be inferred (njriasan)

    PR #8154: Testhound/native cast 8138 (testhound)

    PR #8158: adding -pthread for linux-ppc64le in setup.py (esc)

    PR #8164: remove myself from automatic reviewer assignment (esc)

    PR #8167: CUDA: Facilitate and document passing arrays / pointers to foreign functions (gmarkall)

    PR #8180: CUDA: Initial support for Minor Version Compatibility (gmarkall)

    PR #8183: Add n_keys option to Dict.empty() (stefanfed gmarkall)

    PR #8198: Update the release template to include updating the version table. (stuartarchibald)

    PR #8200: Make the NRT use the “unsafe” allocation API by default. (stuartarchibald)

    PR #8201: Bump llvmlite dependency to 0.40.dev0 for Numba 0.57.0dev0 (stuartarchibald)

    PR #8207: development tag should be in monofont (esc)

    PR #8212: release checklist: include a note to ping @RC_testers on discourse (esc)

    PR #8216: chore: Set permissions for GitHub actions (naveensrinivasan)

    PR #8217: Fix syntax in docs (jorgepiloto)

    PR #8220: Added the interval example as doctest (kc611)

    PR #8221: CUDA stubs docstring: Replace illegal escape sequence (gmarkall)

    PR #8228: Fix typo in @vectorize docstring and a NumPy spelling. (stuartarchibald)

    PR #8229: Remove mk_unique_var in inline_closurecall.py (sklam)

    PR #8234: Replace @overload_glue by @overload for 20 NumPy functions (guilhermeleobas)

    PR #8235: Make the NRT stats counters optional. (stuartarchibald)

    PR #8238: Advanced Indexing Support #1 (kc611)

    PR #8240: Add get_shared_mem_per_block method to Dispatcher (testhound)

    PR #8241: Reorder typeof checks to avoid infinite loops on StructrefProxy __hash__ (DannyWeitekamp)

    PR #8243: Add a note to reference/numpysupported.rst ()

    PR #8245: Fix links in CONTRIBUTING.md ()

    PR #8247: Fix issue 8127 (bszollosinagy)

    PR #8250: Fix issue 8161 (bszollosinagy)

    PR #8253: CUDA: Verify NVVM IR prior to compilation (gmarkall)

    PR #8255: CUDA: Make numba.cuda.tests.doc_examples.ffi a module to fix #8252 (gmarkall)

    PR #8256: Migrate linear algebra functions from glue_lowering (apmasell)

    PR #8258: refactor np.where to use overload (guilhermeleobas)

    PR #8259: Add np.broadcast_to(scalar_array, ()) (guilhermeleobas)

    PR #8264: remove mk_unique_var from parfor_lowering_utils.py (guilhermeleobas)

    PR #8265: Remove mk_unique_var from array_analysis.py (guilhermeleobas)

    PR #8266: Remove mk_unique_var in untyped_passes.py (guilhermeleobas)

    PR #8267: Fix segfault for invalid axes in np.split (aseyboldt)

    PR #8271: Implement some CUDA intrinsics with @overload, @overload_attribute, and @intrinsic (gmarkall)

    PR #8274: Update version support table doc for 0.56. (stuartarchibald)

    PR #8275: Update CHANGE_LOG for 0.56.0 final (stuartarchibald)

    PR #8283: Clean up / remove support for old NumPy versions (gmarkall)

    PR #8287: Drop CUDA 10.2 (gmarkall)

    PR #8289: Revert #8265. (stuartarchibald)

    PR #8290: CUDA: Replace use of deprecated NVVM IR features, questionable constructs (gmarkall)

    PR #8292: update checklist (esc)

    PR #8294: CUDA: Add trig ufunc support (gmarkall)

    PR #8295: Add get_const_mem_size method to Dispatcher (testhound gmarkall)

    PR #8297: Add __name__ attribute to CUDAUFuncDispatcher and test case (testhound)

    PR #8299: Fix build for mingw toolchain (Biswa96)

    PR #8302: CUDA: Revert numba_nvvm intrinsic name workaround (gmarkall)

    PR #8308: CUDA: Support for multiple signatures (gmarkall)

    PR #8315: Add get_local_mem_per_thread method to Dispatcher (testhound)

    PR #8319: Bump minimum supported Python version to 3.8 (esc stuartarchibald jamesobutler)

    PR #8320: Add __name__ support for GUFuncs (testhound)

    PR #8321: Fix literal_unroll pass erroneously exiting on non-conformant loop. (stuartarchibald)

    PR #8325: Remove use of mk_unique_var in stencil.py (bszollosinagy)

    PR #8326: Remove mk_unique_var from parfor_lowering.py (guilhermeleobas)

    PR #8331: Extend docs with info on how to call C functions from Numba (guilhermeleobas)

    PR #8334: Add dict(*iterable) constructor (guilhermeleobas)

    PR #8335: Remove deprecated pycc script and related source. (stuartarchibald)

    PR #8336: Fix typos of “Generalized” in GUFunc-related code (gmarkall)

    PR #8338: Calculate reductions before fusion so that use of reduction vars can stop fusion. (DrTodd13)

    PR #8339: Fix #8291 parfor leak of redtoset variable (sklam)

    PR #8341: CUDA: Support multiple outputs for Generalized Ufuncs (gmarkall)

    PR #8343: Eliminate references to type annotation in compile_ptx (testhound)

    PR #8348: Add get_max_threads_per_block method to Dispatcher (testhound)

    PR #8354: pin setuptools to < 65 and switch from mamba to conda on RTD (esc gmarkall)

    PR #8357: Clean up the buildscripts directory. (stuartarchibald)

    PR #8359: adding warnings about cache behaviour (luk-f-a)

    PR #8368: Remove glue_lowering in random math that requires IR (apmasell)

    PR #8376: Fix issue 8370 (bszollosinagy)

    PR #8387: Add support for compute capability in IR Lowering (testhound)

    PR #8388: Remove more references to the pycc binary. (stuartarchibald)

    PR #8389: Make C++ extensions compile with correct compiler (apmasell)

    PR #8390: Use NumPy logic for lessthan in sort to move NaNs to the back. (sklam)

    PR #8401: Remove Cuda toolkit version check (testhound)

    PR #8415: Refactor numba.np.arraymath methods from lower_builtins to overloads (kc611)

    PR #8418: Fixes ravel failure on 1d arrays (#5229) (cako)

    PR #8421: Update release checklist: add a task to check dependency pinnings on subsequent releases (e.g. PATCH) (esc)

    PR #8422: Switch public CI builds to use gdb from conda packages. (stuartarchibald)

    PR #8423: Remove public facing and CI references to 32 bit linux support. (stuartarchibald, in addition, we are grateful for the contribution of jamesobutler towards a similar goal in PR #8319)

    PR #8425: Post 0.56.2 cleanup (esc)

    PR #8427: Shorten the time to verify test discovery. (stuartarchibald)

    PR #8429: changelog generator script (esc)

    PR #8431: Replace @overload_glue by @overload for np.linspace and np.take (guilhermeleobas)

    PR #8432: Refactor carray/farray to use @overload (guilhermeleobas)

    PR #8435: Migrate np.atleast_? functions from glue_lowering to overload (apmasell)

    PR #8438: Make the initialisation of the NRT more lazy for the njit decorator. (stuartarchibald)

    PR #8439: Update the contributing docs to include a policy on formatting changes. (stuartarchibald)

    PR #8440: [DOC]: Replaces icc_rt with intel-cmplr-lib-rt (oleksandr-pavlyk)

    PR #8442: Implement hasattr(), str() and repr(). (stuartarchibald)

    PR #8446: add version info in ImportError’s (raybellwaves)

    PR #8450: remove GitHub username from changelog generation script (esc)

    PR #8467: Convert implementations using generated_jit to overload (gmarkall)

    PR #8468: Reference test suite in installation documentation (apmasell)

    PR #8469: Correctly handle optional types in parfors lowering (apmasell)

    PR #8473: change the include style in _pymodule.h and remove unused or duplicate headers in two header files ()

    PR #8476: Make setuptools optional at runtime. (stuartarchibald)

    PR #8490: Restore installing SciPy from defaults instead of conda-forge on public CI (esc)

    PR #8494: Remove context.compile_internal where easy on numba/cpython/cmathimpl.py (guilhermeleobas)

    PR #8495: Removes context.compile_internal where easy on numba/cpython/listobj.py (guilhermeleobas)

    PR #8496: Rewrite most of the set API to use overloads (guilhermeleobas)

    PR #8499: Deprecate numba.generated_jit (stuartarchibald)

    PR #8508: This updates the release checklists to capture some more checks. (stuartarchibald)

    PR #8513: Added support for numpy.newaxis (kc611)

    PR #8517: make some typedlist C-APIs public ()

    PR #8518: Adjust stencil tests to use hardcoded python source opposed to AST. (stuartarchibald)

    PR #8520: Added noncentral-chisquared, noncentral-f and logseries distributions (kc611)

    PR #8522: Import jitclass from numba.experimental in jitclass documentation (armgabrielyan)

    PR #8524: Fix grammar in stencil.rst (armgabrielyan)

    PR #8525: Making CUDA specific datamodel manager (sklam)

    PR #8526: Fix broken url (Nimrod0901)

    PR #8527: Fix grammar in troubleshoot.rst (armgabrielyan)

    PR #8532: Vary NumPy version on gpuCI (gmarkall)

    PR #8535: LLVM14 (apmasell)

    PR #8536: Fix fusion bug. (DrTodd13)

    PR #8539: Fix #8534, np.broadcast_to should update array size attr. (stuartarchibald)

    PR #8541: Remove restoration of “free” channel in Azure CI windows builds. (stuartarchibald)

    PR #8542: CUDA: Make arg optional for Stream.add_callback() (gmarkall)

    PR #8544: Remove reliance on npy_<impl> ufunc loops. (stuartarchibald)

    PR #8545: Py3.11 basic support (esc sklam)

    PR #8547: [Unicode] Add more string view usages for unicode operations ()

    PR #8549: Fix rstcheck in Azure CI builds, update sphinx dep and docs to match (stuartarchibald)

    PR #8550: Changes how tests are split between test instances (apmasell)

    PR #8554: Make target for @overload have ‘generic’ as default. (stuartarchibald gmarkall)

    PR #8557: [Unicode] support startswith with args, start and end. ()

    PR #8566: Update workqueue abort message on concurrent access. (stuartarchibald)

    PR #8572: CUDA: Reduce memory pressure from local memory tests (gmarkall)

    PR #8579: CUDA: Add CUDA 11.8 / Hopper support and required fixes (gmarkall)

    PR #8580: adding note about doing a wheel test build prior to tagging (esc)

    PR #8583: Skip tests that contribute to M1 RuntimeDyLd Assertion error (sklam)

    PR #8587: Remove unused refcount removal code, clean core/cpu.py module. (stuartarchibald)

    PR #8588: Remove lowering extension hooks, replace with pass infrastructure. (stuartarchibald)

    PR #8590: Py3.11 support continues (sklam)

    PR #8592: fix failure of test_cache_invalidate due to read-only install (tpwrules)

    PR #8593: Adjusted ULP precesion for noncentral distribution test (kc611)

    PR #8594: Fix various CUDA lineinfo issues (gmarkall)

    PR #8597: Prevent use of NumPy’s MaskedArray. (stuartarchibald)

    PR #8598: Setup Azure CI to test py3.11 (sklam)

    PR #8600: Chrome trace timestamp should be in microseconds not seconds. (sklam)

    PR #8602: Throw error for unsupported dunder methods (apmasell)

    PR #8605: Support for CUDA fp16 math functions (part 1) (testhound)

    PR #8606: [Doc] Make the RewriteArrayExprs doc more precise ()

    PR #8619: Added flat iteration logic for random distributions (kc611)

    PR #8623: Adds support for np.nan_to_num (thomasjpfan)

    PR #8624: DOC: Add guvectorize scalar return example (Matt711)

    PR #8625: Refactor test_ufuncs (gmarkall)

    PR #8626: [unicode-PERF]: use optmized BM algorithm to replace the brute-force finder (dlee992)

    PR #8630: Fix #8628: Don’t test math.trunc with non-float64 NumPy scalars (gmarkall)

    PR #8634: Add new method is_fp16_supported (testhound)

    PR #8636: CUDA: Skip test_ptds on Windows (gmarkall)

    PR #8639: Python 3.11 - fix majority of remaining test failures. (stuartarchibald)

    PR #8644: Fix bare reraise support (sklam)

    PR #8649: Remove numba.core.overload_glue module. (apmasell)

    PR #8659: Preserve module name of jitted class (neilflood)

    PR #8661: Make external compiler discovery lazy in the test suite. (stuartarchibald)

    PR #8662: Add support for .nbytes accessor for numpy arrays (alanhdu)

    PR #8666: Updates for Python 3.8 baseline/Python 3.11 migration (stuartarchibald)

    PR #8673: Enable the CUDA simulator tests on Windows builds in Azure CI. (stuartarchibald)

    PR #8675: Make always_run test decorator a tag and improve shard tests. (stuartarchibald)

    PR #8677: Add support for min and max on boolean types. (DrTodd13)

    PR #8680: Adjust flake8 config to be compatible with flake8=6.0.0 (thomasjpfan)

    PR #8685: Implement __repr__ for numba types (luk-f-a)

    PR #8691: NumPy 1.24 (gmarkall)

    PR #8697: Close stale issues after 7 days (gmarkall)

    PR #8701: Relaxed ULP testing precision for NumPy Generator tests across all systems (kc611)

    PR #8702: Supply concrete timeline for objmode fallback deprecation/removal. (stuartarchibald)

    PR #8706: Fix doctest for @vectorize (sklam)

    PR #8711: Python 3.11 tracing support (continuation of #8670). (AndrewVallette sklam)

    PR #8716: CI: Use set -e in “Before Install” step and fix install (gmarkall)

    PR #8720: Enable coverage for subprocess testing (sklam)

    PR #8723: Check for void return type in cuda.compile_ptx (brandonwillard)

    PR #8726: Make Numba dependency check run ahead of Numba internal imports. (stuartarchibald)

    PR #8728: Fix flake8 checks since upgrade to flake8=6.x (stuartarchibald)

    PR #8729: Run flake8 CI step in multiple processes. (stuartarchibald)

    PR #8732: Add numpy argpartition function support ()

    PR #8735: Update bot to close PRs waiting on authors for more than 3 months (guilhermeleobas)

    PR #8736: Implement np.lib.stride_tricks.sliding_window_view ()

    PR #8744: Update CtypesLinker::add_cu error message to include fp16 usage (testhound gmarkall)

    PR #8746: Fix failing test_dispatcher test case (testhound)

    PR #8748: Suppress known test failures for py3.11 (sklam)

    PR #8751: Recycle test runners more aggressively (apmasell)

    PR #8752: Flake8 fixes for py311 branch (esc sklam)

    PR #8760: Bump llvmlite PR in py3.11 branch testing (sklam)

    PR #8764: CUDA tidy-up: remove some unneeded methods (gmarkall)

    PR #8765: BLD: remove distutils (fangchenli)

    PR #8766: Stale bot: Use abandoned - stale label for closed PRs (gmarkall)

    PR #8771: Update vendored Versioneer from 0.14 to 0.28 (oscargus gmarkall)

    PR #8775: Revert PR#8751 for buildfarm stability (sklam)

    PR #8780: Improved documentation for Atomic CAS (MiloniAtal)

    PR #8781: Ensure gc.collect() is called before checking refcount in tests. (sklam)

    PR #8782: Changed wording of the escape error (MiloniAtal)

    PR #8786: Upgrade stale GitHub action (apmasell)

    PR #8788: CUDA: Fix returned dtype of vectorized functions (Issue #8400) (gmarkall)

    PR #8790: CUDA compare and swap with index (ianthomas23)

    PR #8795: Add pending-deprecation warnings for numba.pycc (stuartarchibald)

    PR #8802: Move the minimum supported NumPy version to 1.21 (stuartarchibald)

    PR #8803: Attempted fix to #8789 by changing compile_ptx to accept a signature instead of argument tuple (KyanCheung)

    PR #8804: Split parfor pass into 3 parts (DrTodd13)

    PR #8809: Update LLVM versions for 0.57 release (apmasell)

    PR #8810: Fix llvmlite dependency in meta.yaml (sklam)

    PR #8816: Fix some buildfarm test failures (sklam)

    PR #8819: Support “static” __getitem__ on Numba types in @njit code. (stuartarchibald)

    PR #8822: Merge py3.11 branch to main (esc AndrewVallette stuartarchibald sklam)

    PR #8826: CUDA CFFI test: conditionally require cffi module (gmarkall)

    PR #8831: Redo py3.11 sync branch with main (sklam)

    PR #8833: Fix typeguard import hook location. (stuartarchibald)

    PR #8836: Fix failing typeguard test. (stuartarchibald)

    PR #8837: Update AzureCI matrix for Python 3.11/NumPy 1.21..1.24 (stuartarchibald)

    PR #8839: Add Dynamic Shared Memory example. (k1m190r)

    PR #8842: Fix buildscripts, setup.py, docs for setuptools becoming optional. (stuartarchibald)

    PR #8843: Pin typeguard to 3.0.1 in AzureCI. (stuartarchibald)

    PR #8848: added lifted loops to glossary term (cherieliu)

    PR #8852: Disable SLP vectorisation due to miscompilations. (stuartarchibald)

    PR #8855: DOC: pip into double backticks in installing.rst (F3eQnxN3RriK)

    PR #8856: Update TBB to use >= 2021.6 by default. (kozlov-alexey stuartarchibald)

    PR #8858: Update deprecation notice for objmode fallback RE @jit use. (stuartarchibald)

    PR #8864: Remove obsolete deprecation notices (gmarkall)

    PR #8866: Revise CUDA deprecation notices (gmarkall)

    PR #8869: Update CHANGE_LOG for 0.57.0rc1 (stuartarchibald esc gmarkall)

    PR #8870: Fix opcode “spelling” change since Python 3.11 in CUDA debug test. (stuartarchibald)

    PR #8879: Remove use of compile_isolated from generator tests. (stuartarchibald)

    PR #8880: Fix missing dependency guard on pyyaml in test_azure_config. (stuartarchibald)

    PR #8881: Replace use of compile_isolated in test_obj_lifetime (sklam)

    PR #8884: Pin llvmlite and NumPy on release branch (sklam)

    PR #8887: Update PyPI supported version tags (bryant1410)

    PR #8896: Remove codecov install (now deleted from PyPI) (gmarkall)

    PR #8902: Enable CALL_FUNCTION_EX fix for py3.11 (sklam)

    PR #8907: Work around issue #8898. Defer exp2 (and log2) calls to Numba internal symbols. (stuartarchibald)

    PR #8909: Fix #8903. NumbaDeprecationWarning``s raised from ``@{gu,}vectorize. (stuartarchibald)

    PR #8929: Update CHANGE_LOG for 0.57.0 final. (stuartarchibald)

    PR #8930: Fix year in change log (jtilly)

    PR #8932: Fix 0.57 release changelog (sklam)

Authors:

    alanhdu

    AndrewVallette

    apmasell

    armgabrielyan

    aseyboldt

    Biswa96

    brandonwillard

    bryant1410

    bszollosinagy

    cako

    cherieliu

    ChiCheng45

    DannyWeitekamp

    dlee992

    dmbelov

    DrTodd13

    eric-wieser

    esc

    F3eQnxN3RriK

    fangchenli

    gmarkall

    guilhermeleobas

    ianthomas23

    jamesobutler

    jorgepiloto

    jtilly

    k1m190r

    kc611

    kozlov-alexey

    KyanCheung

    luk-f-a

    Matt711

    MiloniAtal

    naveensrinivasan

    neilflood

    Nimrod0901

    njriasan

    oleksandr-pavlyk

    oscargus

    raybellwaves

    shangbol

    sklam

    stefanfed

    stuartarchibald

    testhound

    thomasjpfan

    tpwrules

Version 0.56.4 (3 November, 2022)

This is a bugfix release to fix a regression in the CUDA target in relation to the .view() method on CUDA device arrays that is present when using NumPy version 1.23.0 or later.

Pull-Requests:

    PR #8537: Make ol_compatible_view accessible on all targets (gmarkall)

    PR #8552: Update version support table for 0.56.4. (stuartarchibald)

    PR #8553: Update CHANGE_LOG for 0.56.4 (stuartarchibald)

    PR #8570: Release 0.56 branch: Fix overloads with target="generic" for CUDA (gmarkall)

    PR #8571: Additional update to CHANGE_LOG for 0.56.4 (stuartarchibald)

Authors:

    gmarkall

    stuartarchibald

Version 0.56.3 (13 October, 2022)

This is a bugfix release to remove the version restriction applied to the setuptools package and to fix a bug in the CUDA target in relation to copying zero length device arrays to zero length host arrays.

Pull-Requests:

    PR #8475: Remove setuptools version pin (gmarkall)

    PR #8482: Fix #8477: Allow copies with different strides for 0-length data (gmarkall)

    PR #8486: Restrict the TBB development package to supported version in Azure. (stuartarchibald)

    PR #8503: Update version support table for 0.56.3 (stuartarchibald)

    PR #8504: Update CHANGE_LOG for 0.56.3 (stuartarchibald)

Authors:

    gmarkall

    stuartarchibald

Version 0.56.2 (1 September, 2022)

This is a bugfix release that supports NumPy 1.23 and fixes CUDA function caching.

Pull-Requests:

    PR #8239: Add decorator to run a test in a subprocess (stuartarchibald)

    PR #8276: Move Azure to use macos-11 (stuartarchibald)

    PR #8310: CUDA: Fix Issue #8309 - atomics don’t work on complex components (Graham Markall)

    PR #8342: Upgrade to ubuntu-20.04 for azure pipeline CI (jamesobutler)

    PR #8356: Update setup.py, buildscripts, CI and docs to require setuptools<60 (stuartarchibald)

    PR #8374: Don’t pickle LLVM IR for CUDA code libraries (Graham Markall)

    PR #8377: Add support for NumPy 1.23 (stuartarchibald)

    PR #8384: Move strace() check into tests that actually need it (stuartarchibald)

    PR #8386: Fix the docs for numba.get_thread_id (stuartarchibald)

    PR #8407: Pin NumPy version to 1.18-1.24 (Andre Masella)

    PR #8411: update version support table for 0.56.1 (esc)

    PR #8412: Create changelog for 0.56.1 (Andre Masella)

    PR #8413: Fix Azure CI for NumPy 1.23 and use conda-forge scipy (Siu Kwan Lam)

    PR #8414: Hotfix for 0.56.2 (Siu Kwan Lam)

Authors:

    Andre Masella

    esc

    Graham Markall

    jamesobutler

    Siu Kwan Lam

    stuartarchibald

Version 0.56.1 (NO RELEASE)

The release was skipped due to issues during the release process.
Version 0.56.0 (25 July, 2022)

This release continues to add new features, bug fixes and stability improvements to Numba. Please note that this will be the last release that has support for Python 3.7 as the next release series (Numba 0.57) will support Python 3.11! Also note that, this will be the last release to support linux-32 packages produced by the Numba team.

Python language support enhancements:

    Previously missing support for large, in-line dictionaries and internal calls to functions with large numbers of keyword arguments in Python 3.10 has been added.

    operator.mul now works for list s.

    Literal slices, e.g. slice(1, 10, 2) can be returned from nopython mode functions.

    The len function now works on dict_keys, dict_values and dict_items .

    Numba’s set implementation now supports reference counted items e.g. strings.

Numba specific feature enhancements:

    The experimental jitclass feature gains support for a large number of builtin methods e.g. declaring __hash__ or __getitem__ for a jitclass type.

    It’s now possible to use @vectorize on an already @jit family decorated function.

    Name mangling has been updated to emit compiled function names that exactly match the function name in Python. This means debuggers, like GDB, can be set to break directly on Python function names.

    A GDB “pretty printing” support module has been added, when loaded into GDB Numba’s internal representations of Python/NumPy types are rendered inside GDB as they would be in Python.

    An experimental option is added to the @jit family decorators to entirely turn off LLVM’s optimisation passes for a given function (see _dbg_optnone kwarg in the @jit decorator family).

    A new environment variable is added NUMBA_EXTEND_VARIABLE_LIFETIMES, which if set will extend the lifetime of variables to the end of their basic block, this to permit a debugging experience in GDB similar to that found in compiled C/C++/Fortran code.

NumPy features/enhancements:

    Initial support for passing, using and returning numpy.random.Generator instances has been added, this currently includes support for the random distribution.

    The broadcasting functions np.broadcast_shapes and np.broadcast_arrays are now supported.

    The min and max functions now work with np.timedelta64 and np.datetime64 types.

    Sorting multi-dimensional arrays along the last axis is now supported in np.sort().

    The np.clip function is updated to accept NumPy arrays for the a_min and a_max arguments.

    The NumPy allocation routines (np.empty , np.ones etc.) support shape arguments specified using members of enum.IntEnum s.

    The function np.random.noncentral_chisquare is now supported.

    The performance of functions np.full and np.ones has been improved.

Parallel Accelerator enhancements:

    The parallel=True functionality is enhanced through the addition of the functions numba.set_parallel_chunksize and numba.get_parallel_chunksize to permit a more fine grained scheduling of work defined in a parallel region. There is also support for adjusting the chunksize via a context manager.

    The ID of a thread is now defined to be predictable and within a known range, it is available through calling the function numba.get_thread_id.

    The performance of @stencil s has been improved in both serial and parallel execution.

CUDA enhancements:

    New functionality:

        Self-recursive device functions.

        Vector type support (float4, int2, etc.).

        Shared / local arrays of extension types can now be created.

        Support for linking CUDA C / C++ device functions into Python kernels.

        PTX generation for Compute Capabilities 8.6 and 8.7 - e.g. RTX A series, GTX 3000 series.

        Comparison operations for float16 types.

    Performance improvements:

        Context queries are no longer made during launch configuration.

        Launch configurations are now LRU cached.

        On-disk caching of CUDA kernels is now supported.

    Documentation: many new examples added.

Docs:

    Numba now has an official “mission statement”.

    There’s now a “version support table” in the documentation to act as an easy to use, single reference point, for looking up information about Numba releases and their required/supported dependencies.

General Enhancements:

    Numba imports more quickly in environments with large numbers of packages as it now uses importlib-metadata for querying other packages.

    Emission of chrome tracing output is now supported for the internal compilation event handling system.

    This release is tested and known to work when using the Pyston Python interpreter.

Pull-Requests:

    PR #5209: Use importlib to load numba extensions (Stepan Rakitin Graham Markall stuartarchibald)

    PR #5877: Jitclass builtin methods (Ethan Pronovost Graham Markall)

    PR #6490: Stencil output allocated with np.empty now and new code to initialize the borders. (Todd A. Anderson)

    PR #7005: Make numpy.searchsorted match NumPy when first argument is unsorted (Brandon T. Willard)

    PR #7363: Update cuda.local.array to clarify “simple constant expression” (e.g. no NumPy ints) (Sterling Baird)

    PR #7364: Removes an instance of signed integer overflow undefined behaviour. (Tobias Sargeant)

    PR #7537: Add chrome tracing (Hadia Ahmed Siu Kwan Lam)

    PR #7556: Testhound/fp16 comparison (Michael Collison Graham Markall)

    PR #7586: Support for len on dict.keys, dict.values, and dict.items (Nick Riasanovsky)

    PR #7617: Numba gdb-python extension for printing (stuartarchibald)

    PR #7619: CUDA: Fix linking with PTX when compiling lazily (Graham Markall)

    PR #7621: Add support for linking CUDA C / C++ with @cuda.jit kernels (Graham Markall)

    PR #7625: Combined parfor chunking and caching PRs. (stuartarchibald Todd A. Anderson Siu Kwan Lam)

    PR #7651: DOC: pypi and conda-forge badges (Ray Bell)

    PR #7660: Add support for np.broadcast_arrays (Guilherme Leobas)

    PR #7664: Flatten mangling dicts into a single dict (Graham Markall)

    PR #7680: CUDA Docs: include example calling slow matmul (Graham Markall)

    PR #7682: performance improvements to np.full and np.ones (Rishi Kulkarni)

    PR #7684: DOC: remove incorrect warning in np.random reference (Rishi Kulkarni)

    PR #7685: Don’t convert setitems that have dimension mismatches to parfors. (Todd A. Anderson)

    PR #7690: Implemented np.random.noncentral_chisquare for all size arguments (Rishi Kulkarni)

    PR #7695: IntEnumMember support for np.empty, np.zeros, and np.ones (Benjamin Graham)

    PR #7699: CUDA: Provide helpful error if the return type is missing for declare_device (Graham Markall)

    PR #7700: Support for scalar arguments in Np.ascontiguousarray (Dhruv Patel)

    PR #7703: Ignore unsupported types in ShapeEquivSet._getnames() (Benjamin Graham)

    PR #7704: Move the type annotation pass to post legalization. (stuartarchibald)

    PR #7709: CUDA: Fixes missing type annotation pass following #7704 (stuartarchibald)

    PR #7712: Fixing issue 7693 (stuartarchibald Graham Markall luk-f-a)

    PR #7714: Support for boxing SliceLiteral type (Nick Riasanovsky)

    PR #7718: Bump llvmlite dependency to 0.39.0dev0 for Numba 0.56.0dev0 (stuartarchibald)

    PR #7724: Update URLs in error messages to refer to RTD docs. (stuartarchibald)

    PR #7728: Document that AOT-compiled functions do not check arg types (Graham Markall)

    PR #7729: Handle Omitted/OmittedArgDataModel in DI generation. (stuartarchibald)

    PR #7732: update release checklist following 0.55.0 RC1 (esc)

    PR #7736: Update CHANGE_LOG for 0.55.0 final. (stuartarchibald)

    PR #7740: CUDA Python 11.6 support (Graham Markall)

    PR #7744: Fix issues with locating/parsing source during DebugInfo emission. (stuartarchibald)

    PR #7745: Fix the release year for Numba 0.55 change log entry. (stuartarchibald)

    PR #7748: Fix #7713: Ensure _prng_random_hash return has correct bitwidth (Graham Markall)

    PR #7749: Refactor threading layer priority tests to not use stdout/stderr (stuartarchibald)

    PR #7752: Fix #7751: Use original filename for array exprs (Graham Markall)

    PR #7755: CUDA: Deprecate support for CC < 5.3 and CTK < 10.2 (Graham Markall)

    PR #7763: Update Read the Docs configuration (automatic) (readthedocs-assistant)

    PR #7764: Add dbg_optnone and dbg_extend_lifetimes flags (Siu Kwan Lam)

    PR #7771: Move function unique ID to abi-tags (stuartarchibald Siu Kwan Lam)

    PR #7772: CUDA: Add Support to Creating StructModel Array (Michael Wang)

    PR #7776: Updates coverage.py config (stuartarchibald)

    PR #7777: Remove reference existing issue from GH template. (stuartarchibald)

    PR #7778: Remove long deprecated flags from the CLI. (stuartarchibald)

    PR #7780: Fix sets with reference counted items (Benjamin Graham)

    PR #7782: adding reminder to check on deprecations (esc)

    PR #7783: remove upper limit on Python version (esc)

    PR #7786: Remove dependency on intel-openmp for OSX (stuartarchibald)

    PR #7788: Avoid issue with DI gen for arrayexprs. (stuartarchibald)

    PR #7796: update change-log for 0.55.1 (esc)

    PR #7797: prune README (esc)

    PR #7799: update the release checklist post 0.55.1 (esc)

    PR #7801: add sdist command and umask reminder (esc)

    PR #7804: update local references from master -> main (esc)

    PR #7805: Enhance source line finding logic for debuginfo (Siu Kwan Lam)

    PR #7809: Updates the gdb configuration to accept a binary name or a path. (stuartarchibald)

    PR #7813: Extend parfors test timeout for aarch64. (stuartarchibald)

    PR #7814: CUDA Dispatcher refactor (Graham Markall)

    PR #7815: CUDA Dispatcher refactor 2: inherit from dispatcher.Dispatcher (Graham Markall)

    PR #7817: Update intersphinx URLs for NumPy and llvmlite. (stuartarchibald)

    PR #7823: Add renamed vars to callee scope such that it is self consistent. (stuartarchibald)

    PR #7829: CUDA: Support Enum/IntEnum in Kernel (Michael Wang)

    PR #7833: Add version support information table to docs. (stuartarchibald)

    PR #7835: Fix pickling error when module cannot be imported (idorrington)

    PR #7836: min() and max() support for np.datetime and np.timedelta (Benjamin Graham)

    PR #7837: Initial refactoring of parfor reduction lowering (Siu Kwan Lam)

    PR #7845: change time.time() to time.perf_counter() in docs (Nopileos2)

    PR #7846: Fix CUDA enum vectorize test on Windows (Graham Markall)

    PR #7848: Support for int * list (Nick Riasanovsky)

    PR #7850: CUDA: Pass fastmath compiler flag down to compile_ptx and compile_device; Improve fastmath tests (Michael Wang)

    PR #7855: Ensure np.argmin/no.argmax return type is intp (stuartarchibald)

    PR #7858: CUDA: Deprecate ptx Attribute and Update Tests (Graham Markall Michael Wang)

    PR #7861: Fix a spelling mistake in README (Zizheng Guo)

    PR #7864: Fix cross_iter_dep check. (Todd A. Anderson)

    PR #7865: Remove add_user_function (Graham Markall)

    PR #7866: Support for large numbers of args/kws with Python 3.10 (Nick Riasanovsky)

    PR #7878: CUDA: Remove some deprecated support, add CC 8.6 and 8.7 (Graham Markall)

    PR #7893: Use uuid.uuid4() as the key in serialization. (stuartarchibald)

    PR #7895: Remove use of llvmlite.llvmpy (Andre Masella)

    PR #7898: Skip test_ptds under cuda-memcheck (Graham Markall)

    PR #7901: Pyston compatibility for the test suite (Kevin Modzelewski)

    PR #7904: Support m1 (esc)

    PR #7911: added sys import (Nightfurex)

    PR #7915: CUDA: Fix test checking debug info rendering. (stuartarchibald)

    PR #7918: Add JIT examples to CUDA docs (brandon-b-miller Graham Markall)

    PR #7919: Disallow //= reductions in pranges. (Todd A. Anderson)

    PR #7924: Retain non-modified index tuple components. (Todd A. Anderson)

    PR #7939: Fix rendering in feature request template. (stuartarchibald)

    PR #7940: Implemented np.allclose in numba/np/arraymath.py (Gagandeep Singh)

    PR #7941: Remove debug dump output from closure inlining pass. (stuartarchibald)

    PR #7946: instructions for creating a build environment were outdated (esc)

    PR #7949: Add Cuda Vector Types (Michael Wang)

    PR #7950: mission statement (esc)

    PR #7956: Stop using pip for 3.10 on public ci (Revert “start testing Python 3.10 on public CI”) (esc)

    PR #7957: Use cloudpickle for disk caches (Siu Kwan Lam)

    PR #7958: numpy.clip accept numpy.array for a_min, a_max (Gagandeep Singh)

    PR #7959: Permit a new array model to have a super set of array model fields. (stuartarchibald)

    PR #7961: numba.typed.typeddict.Dict.get uses castedkey to avoid returning default value even if the key is present (Gagandeep Singh)

    PR #7963: remove the roadmap from the sphinx based docs (esc)

    PR #7964: Support for large constant dictionaries in Python 3.10 (Nick Riasanovsky)

    PR #7965: Use uuid4 instead of PID in cache temp name to prevent collisions. (stuartarchibald)

    PR #7971: lru cache for configure call (Tingkai Liu)

    PR #7972: Fix fp16 support for cuda shared array (Michael Collison Graham Markall)

    PR #7986: Small caching refactor to support target cache implementations (Graham Markall)

    PR #7994: Supporting multidimensional arrays in quick sort (Gagandeep Singh Siu Kwan Lam)

    PR #7996: Fix binding logic in @overload_glue. (stuartarchibald)

    PR #7999: Remove @overload_glue for NumPy allocators. (stuartarchibald)

    PR #8003: Add np.broadcast_shapes (Guilherme Leobas)

    PR #8004: CUDA fixes for Windows (Graham Markall)

    PR #8014: Fix support for {real,imag} array attrs in Parfors. (stuartarchibald)

    PR #8016: [Docs] [Very Minor] Make numba.jit boundscheck doc line consistent (Kyle Martin)

    PR #8017: Update FAQ to include details about using debug-only option (Guilherme Leobas)

    PR #8027: Support for NumPy 1.22 (stuartarchibald)

    PR #8031: Support for Numpy BitGenerators PR#1 - Core Generator Support (Kaustubh)

    PR #8035: Fix a couple of typos RE implementation (stuartarchibald)

    PR #8037: CUDA self-recursion tests (Graham Markall)

    PR #8044: Make Python 3.10 kwarg peephole less restrictive (Nick Riasanovsky)

    PR #8046: Fix caching test failures (Siu Kwan Lam)

    PR #8049: support str(bool) syntax (LI Da)

    PR #8052: Ensure pthread is linked in when building for ppc64le. (Siu Kwan Lam)

    PR #8056: Move caching tests from test_dispatcher to test_caching (Graham Markall)

    PR #8057: Fix coverage checking (Graham Markall)

    PR #8064: Rename “nb:run_pass” to “numba:run_pass” and document it. (Siu Kwan Lam)

    PR #8065: Fix PyLowering mishandling starargs (Siu Kwan Lam)

    PR #8068: update changelog for 0.55.2 (esc)

    PR #8077: change return type of np.broadcast_shapes to a tuple (Guilherme Leobas)

    PR #8080: Fix windows test failure due to timeout when the machine is slow poss… (Siu Kwan Lam)

    PR #8081: Fix erroneous array count in parallel gufunc kernel generation. (stuartarchibald)

    PR #8089: Support on-disk caching in the CUDA target (Graham Markall)

    PR #8097: Exclude libopenblas 0.3.20 on osx-arm64 (esc)

    PR #8099: Fix Py_DECREF use in case of error state (for devicearray). (stuartarchibald)

    PR #8102: Combine numpy run_constrained in meta.yaml to the run requirements (Siu Kwan Lam)

    PR #8109: Pin TBB support with respect to incompatible 2021.6 API. (stuartarchibald)

    PR #8118: Update release checklists post 0.55.2 (esc)

    PR #8123: Fix CUDA print tests on Windows (Graham Markall)

    PR #8124: Add explicit checks to all allocators in the NRT. (stuartarchibald)

    PR #8126: Mark gufuncs as having mutable outputs (Andre Masella)

    PR #8133: Fix #8132. Regression in Record.make_c_struct for handling nestedarray (Siu Kwan Lam)

    PR #8137: CUDA: Fix #7806, Division by zero stops the kernel (Graham Markall)

    PR #8142: CUDA: Fix some missed changes from dropping 9.2 (Graham Markall)

    PR #8144: Fix NumPy capitalisation in docs. (stuartarchibald)

    PR #8145: Allow ufunc builder to use previously JITed function (Andre Masella)

    PR #8151: pin NumPy to build 0 of 1.19.2 on public CI (esc)

    PR #8163: CUDA: Remove context query in launch config (Graham Markall)

    PR #8165: Restrict strace based tests to be linux only via support feature. (stuartarchibald)

    PR #8170: CUDA: Fix missing space in low occupancy warning (Graham Markall)

    PR #8175: make build and upload order consistent (esc)

    PR #8181: Fix various typos (luzpaz)

    PR #8187: Update CHANGE_LOG for 0.55.2 (stuartarchibald esc)

    PR #8189: updated version support information for 0.55.2/0.57 (esc)

    PR #8191: CUDA: Update deprecation notes for 0.56. (Graham Markall)

    PR #8192: Update CHANGE_LOG for 0.56.0 (stuartarchibald esc Siu Kwan Lam)

    PR #8195: Make the workqueue threading backend once again fork safe. (stuartarchibald)

    PR #8196: Fix numerical tolerance in parfors caching test. (stuartarchibald)

    PR #8197: Fix isinstance warning check test. (stuartarchibald)

    PR #8203: pin llvmlite 0.39 for public CI builds (esc)

    PR #8255: CUDA: Make numba.cuda.tests.doc_examples.ffi a module to fix #8252 (Graham Markall)

    PR #8274: Update version support table doc for 0.56. (stuartarchibald)

    PR #8275: Update CHANGE_LOG for 0.56.0 final (stuartarchibald)

Authors:

    Andre Masella

    Benjamin Graham

    brandon-b-miller

    Brandon T. Willard

    Gagandeep Singh

    Dhruv Patel

    LI Da

    Todd A. Anderson

    Ethan Pronovost

    esc

    Tobias Sargeant

    Graham Markall

    Guilherme Leobas

    Zizheng Guo

    Hadia Ahmed

    idorrington

    Michael Wang

    Kaustubh

    Kevin Modzelewski

    luk-f-a

    luzpaz

    Kyle Martin

    Nightfurex

    Nick Riasanovsky

    Nopileos2

    Ray Bell

    readthedocs-assistant

    Rishi Kulkarni

    Sterling Baird

    Siu Kwan Lam

    stuartarchibald

    Stepan Rakitin

    Michael Collison

    Tingkai Liu

(thor)

doc: Updated filesystems/gnunet-fuse to 0.20.0

(nikita)

gnunet-fuse: update to version 0.20.0

(nikita)

doc: Updated net/gnunet-gtk to 0.20.0

(nikita)

gnunet-gtk: update to version 0.20.0

(nikita)

doc: Updated net/gnunet to 0.20.0

(nikita)

gnunet: update to version 0.20.0

Changelog (taken from NEWS):

v0.20.0:
  - GNUNET_TESTING_get_testname_from_underscore renamed to GNUNET_STRINGS_get_suffix_from_binary_name and moved from libgnunettesting to libgnuneutil
  - Move GNUNET_s into libgnunetutil.
  - re-introduce compiler annotation for array size in signature
  - function-signature adjustment due to compiler error
  - GNUNET_PQ_get_oid removed, GNUNET_PQ_get_oid_by_name improved
  - Added GNUNET_PQ_get_oid_by_name
  - added GNUNET_PQ_get_oid()
  - Added new CCA-secure KEM and use in IDENTITY encryption
  - Add KEM API to avoid ephemeral private key management
  - Add new GNUNET_PQ_event_do_poll() API to gnunet_pq_lib.h
  - Added API to support arrays in query results
  - Improve PQ API documentation.
  - API for array types extended for times
  - API extended for array query types
  - relevant array-types in queries (not results) in postgresql added
  - just style fixes, int to enum
  - initial steps towards support of array-types in posgresql
  - adds GNUNET_JSON_spec_object_const() and GNUNET_JSON_spec_array_const()

(nikita)

doc: Updated devel/py-llvmlite to 0.41.1

(thor)

devel/py-llvmlite: un-break at least on Linux, update to 0.41.1 with static LLVM

This now builds a patched LLVM that is statically linked, with llvmlite patches,
as upstream wants and supports as only variant.

This has not been tested widely, but has been uncondtionally BROKEN before.

v0.41.1 (Oct 17, 2023)¶

This is a maintenance release that includes a workaround in the test suite for ORCJit issues on the aarch64 platform. Also, this is the last release to support the Windows 32-bit platform (win32).

Pull-Requests:

    PR #996: fix typos found by codespell (esc)

    PR #997: Fix issue #880 by ensuring all sources are compiled under FreeBSD. (ke6jjj)

    PR #998: adding sphinx_rtd_theme to RTD build to fix build (esc)

    PR #1001: Fix / workaround for OrcJIT blocking issues (gmarkall)

Authors:

    esc

    ke6jjj

    gmarkall

v0.41.0 (Sept 20, 2023)¶

Pull-Requests:

    PR #871: Refactor native library loading (folded sklam)

    PR #896: drop upper limit on Python for conda recipe (esc)

    PR #904: Create GitHub Action for llvmlite release (apmasell)

    PR #934: Expose TargetLibraryInfo pass (sklam)

    PR #935: Disable zlib for LLVM on Windows (apmasell)

    PR #936: Enable querying constants and value kinds (tbennun)

    PR #939: Bump llvmdev build number to include the nozlib change for windows (sklam)

    PR #940: Update CHANGE_LOG for 0.40.0 final. (stuartarchibald)

    PR #942: Add ORCJITv2 support (apmasell)

    PR #951: Add a type hint for IntType.width (apmasell)

    PR #952: Fix CI failing due to unsupported target triple on non-x86 platforms. (sklam)

    PR #958: fixup LLVM versions in version compat table (esc)

    PR #959: Remove support for LLVM < 14 (apmasell)

    PR #960: add various bullets to release checklists and sync (esc)

    PR #963: Allow adding comments to generated IR (apmasell)

    PR #966: build: support building on GNU/Hurd (pinotree)

    PR #967: Expose library name in OrcJIT tracker (apmasell)

    PR #968: Update LLVM manual build instructions (apmasell)

    PR #969: update changelog on main for v0.40.1 (esc)

    PR #983: adding RTD conf file V2 as per request (esc)

    PR #985: Update release checklist post 0.41.0rc1 (esc)

    PR #988: Fix FreeBsd build (sklam)

Authors:

    apmasell

    esc

    folded

    pinotree

    sklam

    stuartarchibald

    tbennun

v0.40.1 (June 21, 2023)¶

Pull-Requests:

    PR #945: Fix #944. Add .argtypes to prevent errors in pypy. (Siu Kwan Lam)

    PR #947: Update SVML patch for LLVM 14 (Andre Masella)

    PR #949: Handle PowerPC synonyms (Andre Masella)

    PR #950: Fix incorrect byval and other attributes on LLVM 14 (Andre Masella)

Authors:

    Andre Masella

    Siu Kwan Lam

v0.40.0 (May 1, 2023)¶

This release predominantly upgrades to LLVM 14 and Python 3.11. Bindings to a large number of passes are added. The minimum supported Python version is now Python 3.8.

Note: A bug was discovered in LLVM’s RuntimeDyldELF on the Aarch64 platform that can cause segfaults when cross module symbols are linked. It is necessary for JIT users to build LLVM with the patch added in PR#926.

Pull-Requests:

    PR #827: Add more LLVM pass bindings (apmasell)

    PR #830: Add LLVM 14 support (apmasell)

    PR #860: the git tag for the RC needs an rc1 suffix (esc)

    PR #869: bump max Python version to 3.11 (esc sklam)

    PR #876: Remove llvmlite.llvmpy after deprecation (apmasell)

    PR #883: Adds support for calling functions with ‘tail’, ‘notail’, or ‘musttail’ markers. (bslatkin)

    PR #886: Simplify setup.py Python version guard (mbargull)

    PR #892: Bump minimum supported Python version to 3.8 (jamesobutler)

    PR #893: Upgrade to ubuntu-20.04 for azure pipeline CI (jamesobutler)

    PR #899: Run Minconda install with bash (gmarkall)

    PR #903: Fix flake8 config and style for flake8 6 (gmarkall)

    PR #905: Add YouCompleteMe configuration file and ignore vim swap files (gmarkall)

    PR #906: Replace importlib-resources legacy API use (sklam)

    PR #910: Aarch64 split build for LLVM14 (sklam)

    PR #921: Setup AzureCI to use py311 and llvm14 (sklam)

    PR #922: Fix AzureCI not using llvm14 on windows (sklam)

    PR #926: llvmdev recipe: Add patch that clears GOTOffsetMap (apmasell gmarkall sklam)

    PR #930: Update changelog for 0.40.0rc1 (sklam stuartarchibald)

    PR #931: Remove maximum Python version limit (sklam apmasell)

    PR #932: Fix wheel builds (sklam)

    PR #935: Disable zlib for LLVM on Windows (apmasell)

    PR #939: Bump llvmdev build number to include the nozlib change for windows (sklam)

    PR #940: Update CHANGE_LOG for 0.40.0 final. (stuartarchibald)

Authors:

    apmasell

    bslatkin

    esc

    gmarkall

    jamesobutler

    mbargull

    sklam

    stuartarchibald

v0.39.1 (September 1, 2022)¶

This is a maintenance release to fix build issues on MacOS.

Pull-Requests:

    PR #752: Skip test if libm is not found (Siu Kwan Lam)

    PR #865: Move Azure to use macos-11 (stuartarchibald)

    PR #874: Add zlib as a dependency for aarch64 (esc)

    PR #878: Update changelog (Andre Masella)

v0.39.0 (July 25, 2022)¶

This release predominantly adds new features and improves functionality.

    It’s now possible to directly set LLVM metadata on global variables.

    Functions and global variables now support the specification of a section in which they should be placed.

    The attribute source_file had been added to the ModuleRef class, it returns the module’s original file name.

    The FFI library binding to LLVM is now loaded with importlib to increase compatibility with other projects and improve start-up times.

    Linux builds now use the parallel option to make to speed up building the FFI library.

    Preliminary work to expose LLVM’s optimization-remarks interface has been undertaken. The bindings are exposed and tested, but not yet documented for general use (additional work is needed).

Deprecations:

    The llvmlite.llvmpy module has been deprecated as the functionality it provides is available through the llvmlite.ir module. See the deprecation guide in the user documentation for details and recommendations regarding replacement.

Pull-Requests:

    PR #328: Build C files separately on Linux and support parallel make (Michał Górny)

    PR #754: manylinux2014 aarch64 wheels with system compilers (esc)

    PR #760: add support for attaching metadata to global variables (Graham Markall John Törnblom)

    PR #786: Update Windows and OSX CI images. (stuartarchibald)

    PR #801: Update ffi.py (franzhaas)

    PR #803: llvm::Module::GetSourceFileName (J. Aaron Pendergrass)

    PR #806: Bump to v0.39.0dev (esc)

    PR #807: Exclude ExecutionEngine tests on linux 32 (esc)

    PR #809: Update CHANGE_LOG for 0.38.0 (stuartarchibald)

    PR #813: Add m1 support to conda build scripts (esc Stan Seibert)

    PR #815: update local references (esc)

    PR #816: remove configuration landscape service as it is no longer used (esc)

    PR #817: remove uppper limit on Python requires (esc)

    PR #819: adding rc and final release checklist templates (esc)

    PR #823: Add section to globals (Andreas Wrisley)

    PR #824: add GitHub URL for PyPi (Andrii Oriekhov)

    PR #825: Add flag handling to more instructions. (stuartarchibald Andre Masella)

    PR #826: Deprecated llvmlite.llvmpy (Andre Masella)

    PR #831: Format C++ code (Andre Masella)

    PR #832: DOC: Fix the syntax for the llvmlite discourse topic link. (stuartarchibald)

    PR #835: Add pre-commit hooks for clang-format (Andre Masella)

    PR #837: Add support for optimization remarks in pass managers (Siu Kwan Lam Andre Masella)

    PR #846: Cherry-Pick: #842 –> main :Changelog for 0.38.1 (esc)

    PR #851: adding the llvm_11_consecutive_registers.patch (esc)

    PR #857: Delegate passmanager remarks methods (Andre Masella)

    PR #858: Update CHANGE_LOG for 0.39.0 (esc Graham Markall stuartarchibald)

    PR #863: Update changelog for 0.39.0 release (Siu Kwan Lam)

    PR #864: Update release date for 0.39.0 release. (stuartarchibald)

    PR #867: Update CHANGE_LOG 0.39.0 final. (stuartarchibald)

Authors:

    Andrii Oriekhov

    Andreas Wrisley

    Andre Masella

    esc

    franzhaas

    Graham Markall

    J. Aaron Pendergrass

    John Törnblom

    Michał Górny

    Stan Seibert

    Siu Kwan Lam

    stuartarchibald

(thor)