Package: numba / 0.56.4+dfsg-2
Metadata
Package | Version | Patches format |
---|---|---|
numba | 0.56.4+dfsg-2 | 3.0 (quilt) |
Patch series
view the series filePatch | File delta | Description |
---|---|---|
no git in generate_lower_listing.patch | (download) |
numba/scripts/generate_lower_listing.py |
6 4 + 2 - 0 ! |
prevent to query the git history whatsoever |
doc no git.patch | (download) |
docs/source/_ext/ghfiles.py |
2 1 + 1 - 0 ! |
--- |
llvm_long_name_PR4373.patch | (download) |
numba/__init__.py |
5 5 + 0 - 0 ! |
[patch] set maximum name size to maximum allowable value Fix for #3876 without needing to patch LLVM. This is upstream PR4373 for fixing LLVM handling of long names. See also upstream Issue#5771, and LLVM review at https://reviews.llvm.org/D41296 |
32bit_skip_64bit_hashing_tests.patch | (download) |
numba/tests/test_hashing.py |
3 3 + 0 - 0 ! |
--- |
skip armhf tests on arm64.patch | (download) |
numba/tests/support.py |
2 1 + 1 - 0 ! |
debian builds packages on armv7l systems which don't support unaligned memory access and upstream tested for armv7l to avoid the tests that would trigger that problem. Unfortunately Debian runs the CI tests in a 32-bit user space on arm64 machines. The archecture reported on those machines is armv8l and so the skipIf test wasn't being triggered. |
skip unicode buserror armhf.patch | (download) |
numba/tests/test_unicode.py |
4 3 + 1 - 0 ! |
skip tests triggering bus errors when running armhf on arm64 when hash() accesses complex multibyte unicode strings. Bug: https://github.com/numba/numba/issues/6659 |
use python3 test nonsense gdb binary.patch | (download) |
numba/tests/test_cli.py |
2 1 + 1 - 0 ! |
by default debian doesn't have a python executable installed use the default python3 name instead. |
reorder setuptools import.patch | (download) |
setup.py |
3 2 + 1 - 0 ! |
setup.py build_ext failed, but earlier on there was an warning message about distutils being imported before setuptools, but setuptools replaces parts of distutils and undesierable behavior may occur. Changing the order did fix my problem with running the build_ext step. |
add test tags.patch | (download) |
numba/tests/npyufunc/test_caching.py |
6 5 + 1 - 0 ! |
--- |
add gdb test check.patch | (download) |
numba/tests/gdb/test_array_arg.py |
3 2 + 1 - 0 ! |
--- |
numpy 1 24 pr8691.patch | (download) |
docs/source/user/installing.rst |
4 2 + 2 - 0 ! |
[patch 01/13] cuda intrinsics tests: correct np.float -> np.float16 I believe this was written in error and should always have been float16. |
python3.11/0011 Comment out all C level changes.patch | (download) |
numba/_dispatcher.cpp |
188 94 + 94 - 0 ! |
comment out all c-level changes This will make Numba compile against the Python 3.11 code-base and C-level API. Note that some of the Numba functionality may not be available as a result. |
python3.11/0012 implement new Python 3.11 opcode RESUME.patch | (download) |
numba/core/byteflow.py |
3 3 + 0 - 0 ! |
implement new python 3.11 opcode: resume With Python 3.11 a new Bytecode has been introduced: `RESUME`. The documentation states that: ``` RESUME has been added. It is a no-op. Performs internal tracing, debugging and optimization checks. ``` And so, this is implemented as a `NOP` (no-operation) in Numba. |
python3.11/0013 bump max python version.patch | (download) |
setup.py |
2 1 + 1 - 0 ! |
bump max python version As title |
python3.11/0014 Minimal changes to get a basic integer sum reduction.patch | (download) |
numba/core/bytecode.py |
3 3 + 0 - 0 ! |
minimal changes to get a basic integer sum reduction for loop working in py3.11 |
python3.11/0015 Get more jump opcodes.patch | (download) |
numba/core/byteflow.py |
3 3 + 0 - 0 ! |
get more jump opcodes |
python3.11/0016 Ignore the NULL push from LOAD_GLOBAL.patch | (download) |
numba/core/byteflow.py |
6 4 + 2 - 0 ! |
ignore the null push from load_global. It's only used in CALL which checks if the TOS is a NULL. |
python3.11/0017 Add more BACKWARD jump and PUSH_NULL.patch | (download) |
numba/core/bytecode.py |
5 4 + 1 - 0 ! |
add more backward jump and push_null. |
python3.11/0018 Fix backward jump offset.patch | (download) |
numba/core/bytecode.py |
2 1 + 1 - 0 ! |
fix backward jump offset |
python3.11/0019 Support KW_NAMES and getting looplifting to work.patch | (download) |
numba/core/byteflow.py |
27 25 + 2 - 0 ! |
support kw_names and getting looplifting to work |
python3.11/0020 Support RETURN_GENERATOR.patch | (download) |
numba/core/byteflow.py |
27 27 + 0 - 0 ! |
support return_generator |
python3.11/0021 Don t dump CACHE bytecode.patch | (download) |
numba/core/bytecode.py |
3 2 + 1 - 0 ! |
don't dump cache bytecode |
python3.11/0022 Partially support MAKE_FUNCTION.patch | (download) |
numba/core/byteflow.py |
5 4 + 1 - 0 ! |
partially support make_function Not sure why MakeFunctionToJitFunction is picking up the <listcomp> in the failing test: - numba.tests.test_looplifting.TestLoopLiftingInAction.test_lift_listcomp_block0 |
python3.11/0023 Leave comment about co_qualname changes in MAKE_FUNC.patch | (download) |
numba/core/byteflow.py |
2 2 + 0 - 0 ! |
leave comment about co_qualname changes in make_function |
python3.11/0024 Support BEFORE_WITH.patch | (download) |
numba/core/bytecode.py |
11 10 + 1 - 0 ! |
support before_with |
python3.11/0025 Support with lifting.patch | (download) |
numba/core/bytecode.py |
21 17 + 4 - 0 ! |
support with-lifting test_withlifting tests report errors=31, skipped=2, expected failures=2 out of 68 tests. A lot of the tests are failing due to cloudpickle incompatibility. |
python3.11/0026 Support JUMP_IS_NONE.patch | (download) |
numba/core/byteflow.py |
6 6 + 0 - 0 ! |
support jump_is_none |
python3.11/0027 Upgrade cloudpickle to version v2.2.0.patch | (download) |
numba/cloudpickle/__init__.py |
4 2 + 2 - 0 ! |
upgrade cloudpickle to version v2.2.0 from commit https://github.com/cloudpipe/cloudpickle/commit/f31859b1dd83fa691f4f7f797166b262c9acb8e7 |
python3.11/0028 Support py3.11 try except.patch | (download) |
numba/core/bytecode.py |
5 3 + 2 - 0 ! |
support py3.11 try-except |
python3.11/0029 Flake8 fixes.patch | (download) |
numba/core/bytecode.py |
1 1 + 0 - 0 ! |
flake8 fixes |
python3.11/0030 More flake8 fix.patch | (download) |
numba/core/typeinfer.py |
3 2 + 1 - 0 ! |
more flake8 fix |
python3.11/0031 Fix older python.patch | (download) |
numba/core/bytecode.py |
30 20 + 10 - 0 ! |
fix older python |
python3.11/0032 Temporarily disable compilation warnings.patch | (download) |
buildscripts/incremental/build.sh |
11 6 + 5 - 0 ! |
temporarily disable compilation warnings |
python3.11/0033 Unbreak python 3.11.patch | (download) |
numba/core/byteflow.py |
105 69 + 36 - 0 ! |
unbreak python <3.11 |
python3.11/0034 Unbreak py3.10 builds.patch | (download) |
numba/_dispatcher.cpp |
202 106 + 96 - 0 ! |
unbreak py3.10 builds |
python3.11/0035 Fix py3.10 failure.patch | (download) |
numba/core/interpreter.py |
44 26 + 18 - 0 ! |
fix py3.10 failure failng test numba.tests.test_parfors.TestParforChunksizing.test_all_iterations_reset_chunksize |
python3.11/0036 Restore 7388 fixes.patch | (download) |
numba/cloudpickle/cloudpickle.py |
6 6 + 0 - 0 ! |
restore #7388 fixes |
python3.11/0037 Fix try except in py 3.11.patch | (download) |
numba/_helperlib.c |
13 12 + 1 - 0 ! |
fix try except in py<3.11 |
python3.11/0038 Fix inplace op.patch | (download) |
numba/core/interpreter.py |
2 1 + 1 - 0 ! |
fix inplace op |
python3.11/0039 Fix hashing support in py 3.11.patch | (download) |
numba/cpython/hashing.py |
4 3 + 1 - 0 ! |
fix hashing support in py<3.11 |
python3.11/0040 Skip testing by git diff for now.patch | (download) |
buildscripts/incremental/test.sh |
4 2 + 2 - 0 ! |
skip testing by git diff for now |
python3.11/0041 Fix CI test run.patch | (download) |
buildscripts/incremental/test.sh |
22 12 + 10 - 0 ! |
fix ci test run |
python3.11/0042 Copy in struct _frame from cpython 3.11 source code .patch | (download) |
numba/_helperlib.c |
28 25 + 3 - 0 ! |
copy in struct _frame from cpython 3.11 source code to regain access to the f_lineno |
python3.11/0043 Adds compiler warning for stub code.patch | (download) |
numba/_dispatcher.cpp |
3 2 + 1 - 0 ! |
adds compiler warning for stub code |
python3.11/0044 Use _FIXED_OFFSET instead of a inline literal number.patch | (download) |
numba/core/bytecode.py |
5 4 + 1 - 0 ! |
use _fixed_offset instead of a inline literal number |
python3.11/0045 Rename base ByteCode class to _ByteCode.patch | (download) |
numba/core/bytecode.py |
6 4 + 2 - 0 ! |
rename base bytecode class to _bytecode |
python3.11/0046 Apply some review suggestions in byteflow.py.patch | (download) |
numba/core/byteflow.py |
17 9 + 8 - 0 ! |
apply some review suggestions in byteflow.py |
python3.11/0047 Address reviews on byteflow.py.patch | (download) |
numba/core/byteflow.py |
20 13 + 7 - 0 ! |
address reviews on byteflow.py |
python3.11/0048 Make PYVERSION check more consistent.patch | (download) |
numba/core/bytecode.py |
12 8 + 4 - 0 ! |
make pyversion check more consistent |
python3.11/0049 Rewrite peep_hole_split_at_pop_block. so it does not.patch | (download) |
numba/core/interpreter.py |
69 39 + 30 - 0 ! |
rewrite peep_hole_split_at_pop_block. so it does not mutate in-place, does better at assigning labels, and does better with multiple POP_BLOCK. |
python3.11/0050 Address reviews in interpreter.py.patch | (download) |
numba/core/interpreter.py |
84 48 + 36 - 0 ! |
address reviews in interpreter.py |
python3.11/0051 Remove dead code StaticReraise.patch | (download) |
numba/core/ir.py |
13 0 + 13 - 0 ! |
remove dead code: staticreraise |
python3.11/0052 Address review in transforms.py.patch | (download) |
numba/core/transforms.py |
4 2 + 2 - 0 ! |
address review in transforms.py |
python3.11/0053 One more PYVERSION guard for newer versions.patch | (download) |
numba/core/byteflow.py |
10 7 + 3 - 0 ! |
one more pyversion guard for newer versions |
python3.11/0054 Apply suggestions from code review.patch | (download) |
numba/core/interpreter.py |
5 2 + 3 - 0 ! |
apply suggestions from code review Co-authored-by: stuartarchibald <stuartarchibald@users.noreply.github.com> |
python3.11/0055 Fix arrayexprs for py3.11.patch | (download) |
numba/np/ufunc/array_exprs.py |
24 23 + 1 - 0 ! |
[patch 01/14] fix arrayexprs for py3.11 - fixes all of numba.tests.test_array_exprs - fixes some of numba.tests.test_parfors |
python3.11/0056 Fix LOAD_GLOBAL arg use in test_parfors for py3.11.patch | (download) |
numba/tests/test_parfors.py |
22 17 + 5 - 0 ! |
[patch 02/14] fix load_global arg use in test_parfors for py3.11 otherwise, sometests may segfault |
python3.11/0057 Move _fix_LOAD_GLOBAL arg to bytecode module and use it in _compute_used_globals.patch | (download) |
numba/core/bytecode.py |
10 9 + 1 - 0 ! |
[patch 03/14] move _fix_load_global_arg to bytecode module and use it in _compute_used_globals |
python3.11/0058 Fix text_flow_control for py3.11.patch | (download) |
numba/core/controlflow.py |
7 7 + 0 - 0 ! |
[patch 04/14] fix test_flow_control for py3.11 |
python3.11/0059 Not a py311 change stop printing source code.patch | (download) |
numba/core/ir.py |
8 6 + 2 - 0 ! |
[patch 05/14] (not a py311 change) stop printing source code in errmsg when lineno is unavail |