Hi all, On behalf of the SciPy development team, I'm pleased to announce the release of SciPy 1.8.1, which is a bug fix release that restores Pythran usage on Windows. Sources and binary wheels can be found at: https://pypi.org/project/scipy/ and at: https://github.com/scipy/scipy/releases/tag/v1.8.1 One of a few ways to install this release with pip: pip install scipy==1.8.1 ===================== SciPy 1.8.1 Release Notes ===================== SciPy 1.8.1 is a bug-fix release with no new features compared to 1.8.0. Notably, usage of Pythran has been restored for Windows builds/binaries. Authors ====== * Henry Schreiner * Maximilian Nöthe * Sebastian Berg (1) * Sameer Deshmukh (1) + * Niels Doucet (1) + * DWesl (4) * Isuru Fernando (1) * Ralf Gommers (4) * Matt Haberland (1) * Andrew Nelson (1) * Dimitri Papadopoulos Orfanos (1) + * Tirth Patel (3) * Tyler Reddy (46) * Pamphile Roy (7) * Niyas Sait (1) + * H. Vetinari (2) * Warren Weckesser (1) A total of 17 people contributed to this release. People with a "+" by their names contributed a patch for the first time. This list of names is automatically generated, and may not be fully complete. Issues closed for 1.8.1 ------------------------------- * `#15258 <https://github.com/scipy/scipy/issues/15258>`__: BUG: sparse \`dot\` method should accept scalars * `#15433 <https://github.com/scipy/scipy/issues/15433>`__: BUG: optimize: minimize: \`ValueError\` when \`np.all(lb==ub)\` * `#15539 <https://github.com/scipy/scipy/issues/15539>`__: BUG: Questionable macOS wheel contents * `#15543 <https://github.com/scipy/scipy/issues/15543>`__: REL: list contributors using GitHub handles * `#15552 <https://github.com/scipy/scipy/issues/15552>`__: BUG: MacOS universal2 wheels have two gfortran shared libraries,... * `#15636 <https://github.com/scipy/scipy/issues/15636>`__: BUG: DOCS incorrect \`source\` link on docs * `#15678 <https://github.com/scipy/scipy/issues/15678>`__: BUG: scipy.stats.skew does not work with scipy.stats.bootstrap * `#16174 <https://github.com/scipy/scipy/issues/16174>`__: Failure of \`TestCorrelateComplex.test_rank0\` in CI with NumPy... Pull requests for 1.8.1 ------------------------------ * `#15167 <https://github.com/scipy/scipy/pull/15167>`__: CI: make sure CI stays on VS2019 unless changed explicitly * `#15306 <https://github.com/scipy/scipy/pull/15306>`__: Revert "BLD Respect the --skip-build flag in setup.py" * `#15504 <https://github.com/scipy/scipy/pull/15504>`__: MAINT: np.all(lb == ub) for optimize.minimize * `#15530 <https://github.com/scipy/scipy/pull/15530>`__: REL: prep for SciPy 1.8.1 * `#15531 <https://github.com/scipy/scipy/pull/15531>`__: [BUG] Fix importing scipy.lib._pep440 * `#15558 <https://github.com/scipy/scipy/pull/15558>`__: CI: re-enable Pythran in Azure Windows CI jobs * `#15566 <https://github.com/scipy/scipy/pull/15566>`__: BUG: fix error message * `#15580 <https://github.com/scipy/scipy/pull/15580>`__: BUG: Avoid C Preprocessor symbol in _hypotests_pythran.py. * `#15614 <https://github.com/scipy/scipy/pull/15614>`__: REL: filter out @ in authors name and add count * `#15637 <https://github.com/scipy/scipy/pull/15637>`__: DOC, MAINT: fix links to wrapped functions and SciPy's distributions * `#15669 <https://github.com/scipy/scipy/pull/15669>`__: BUG: stats: fix a bug in UNU.RAN error handler * `#15691 <https://github.com/scipy/scipy/pull/15691>`__: MAINT: stats: bootstrap: fix bug with \`method="BCa"\` when \`statistic\`... * `#15798 <https://github.com/scipy/scipy/pull/15798>`__: MAINT,BUG: stats: update to UNU.RAN 1.9.0 * `#15870 <https://github.com/scipy/scipy/pull/15870>`__: TST: signal: Convert a test with 'assert_array_less' to 'less... * `#15910 <https://github.com/scipy/scipy/pull/15910>`__: make sure CI stays on VS2019 unless changed explicitly * `#15926 <https://github.com/scipy/scipy/pull/15926>`__: MAINT: 1.8.1 backports/prep * `#16035 <https://github.com/scipy/scipy/pull/16035>`__: BUG: allow scalar input to the \`.dot\` method of sparse matrices * `#16041 <https://github.com/scipy/scipy/pull/16041>`__: MAINT: add include dir explicitly for PROPACK to build with classic... * `#16139 <https://github.com/scipy/scipy/pull/16139>`__: WIP, BLD, MAINT: git security/version shim * `#16152 <https://github.com/scipy/scipy/pull/16152>`__: TST: Fortify invalid-value warning filters to small changes in... * `#16155 <https://github.com/scipy/scipy/pull/16155>`__: MAINT: correct wrong license of Biasedurn * `#16158 <https://github.com/scipy/scipy/pull/16158>`__: MAINT: better UNU.RAN licensing information * `#16163 <https://github.com/scipy/scipy/pull/16163>`__: MAINT: update UNU.RAN copyright information * `#16172 <https://github.com/scipy/scipy/pull/16172>`__: CI: pin Pip to 22.0.4 to avoid issues with \`--no-build-isolation\` * `#16175 <https://github.com/scipy/scipy/pull/16175>`__: TST: fix test failure due to changes in numpy scalar behavior. Checksums ========= MD5 ~~~ fb1d0c039789080b79eb7f5cfbe2ed6c scipy-1.8.1-cp310-cp310-macosx_10_9_x86_64.whl 5cd8466411b660852075ded9609ceb49 scipy-1.8.1-cp310-cp310-macosx_12_0_arm64.whl 1f5e72c7d91fda43cddcbd3f272b9b37 scipy-1.8.1-cp310-cp310-macosx_12_0_universal2.macosx_10_9_x86_64.whl 9cc2f082eeea8b9a75dccfe9d21d2cec scipy-1.8.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl 2cdec5543380dc0c39b061f28893d448 scipy-1.8.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl e8965ab9b1665d1afab0acc6ff716294 scipy-1.8.1-cp310-cp310-win_amd64.whl 2806a86294cff1208ce008515cc114a0 scipy-1.8.1-cp38-cp38-macosx_10_9_x86_64.whl 6fa1abe98efc2b70de1b166a469bb253 scipy-1.8.1-cp38-cp38-macosx_12_0_arm64.whl 4996536b5017a35942c04ac452645983 scipy-1.8.1-cp38-cp38-macosx_12_0_universal2.macosx_10_9_x86_64.whl 7c626d9c3eb2fcddc0be78b27d75f6a9 scipy-1.8.1-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl 8291385cb9fed5df2b99447d67c3d2af scipy-1.8.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl dbd563d80b403c0e0efaebca344f2336 scipy-1.8.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl 6925e10377659b20a5499202df4ec2d8 scipy-1.8.1-cp38-cp38-win32.whl 821c51f0d16b0b33e2fb9600db83f41e scipy-1.8.1-cp38-cp38-win_amd64.whl ab397feec1d9479b3d603e0314e8876a scipy-1.8.1-cp39-cp39-macosx_10_9_x86_64.whl 251ac02266cf7b1f39a479e29c1b2b31 scipy-1.8.1-cp39-cp39-macosx_12_0_arm64.whl cec549c3321d98233d5e25b99bf8a7de scipy-1.8.1-cp39-cp39-macosx_12_0_universal2.macosx_10_9_x86_64.whl 21cba6f710c4065cf602af55e661928e scipy-1.8.1-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl c44c79b4447cd9f7b77a6172e1d6a71a scipy-1.8.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl 44c33e63e2e810891bfd7aebcd87bf4d scipy-1.8.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl 9a7e5cd1440b4d506acc2105501d8643 scipy-1.8.1-cp39-cp39-win32.whl 12952c28f9b1b045598f04628d97c64b scipy-1.8.1-cp39-cp39-win_amd64.whl df5ce79288fc457238aeef18e8f70dfc scipy-1.8.1.tar.gz 15c3e57656dbd1626ae33971f2e80ab1 scipy-1.8.1.tar.xz 51a26cfa0ca569cffd940138814ba23d scipy-1.8.1.zip SHA256 ~~~~~~ 65b77f20202599c51eb2771d11a6b899b97989159b7975e9b5259594f1d35ef4 scipy-1.8.1-cp310-cp310-macosx_10_9_x86_64.whl e013aed00ed776d790be4cb32826adb72799c61e318676172495383ba4570aa4 scipy-1.8.1-cp310-cp310-macosx_12_0_arm64.whl 02b567e722d62bddd4ac253dafb01ce7ed8742cf8031aea030a41414b86c1125 scipy-1.8.1-cp310-cp310-macosx_12_0_universal2.macosx_10_9_x86_64.whl 1da52b45ce1a24a4a22db6c157c38b39885a990a566748fc904ec9f03ed8c6ba scipy-1.8.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl a0aa8220b89b2e3748a2836fbfa116194378910f1a6e78e4675a095bcd2c762d scipy-1.8.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl 4e53a55f6a4f22de01ffe1d2f016e30adedb67a699a310cdcac312806807ca81 scipy-1.8.1-cp310-cp310-win_amd64.whl 28d2cab0c6ac5aa131cc5071a3a1d8e1366dad82288d9ec2ca44df78fb50e649 scipy-1.8.1-cp38-cp38-macosx_10_9_x86_64.whl 6311e3ae9cc75f77c33076cb2794fb0606f14c8f1b1c9ff8ce6005ba2c283621 scipy-1.8.1-cp38-cp38-macosx_12_0_arm64.whl 3b69b90c9419884efeffaac2c38376d6ef566e6e730a231e15722b0ab58f0328 scipy-1.8.1-cp38-cp38-macosx_12_0_universal2.macosx_10_9_x86_64.whl 6cc6b33139eb63f30725d5f7fa175763dc2df6a8f38ddf8df971f7c345b652dc scipy-1.8.1-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl 9c4e3ae8a716c8b3151e16c05edb1daf4cb4d866caa385e861556aff41300c14 scipy-1.8.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl 23b22fbeef3807966ea42d8163322366dd89da9bebdc075da7034cee3a1441ca scipy-1.8.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl 4b93ec6f4c3c4d041b26b5f179a6aab8f5045423117ae7a45ba9710301d7e462 scipy-1.8.1-cp38-cp38-win32.whl 70ebc84134cf0c504ce6a5f12d6db92cb2a8a53a49437a6bb4edca0bc101f11c scipy-1.8.1-cp38-cp38-win_amd64.whl f3e7a8867f307e3359cc0ed2c63b61a1e33a19080f92fe377bc7d49f646f2ec1 scipy-1.8.1-cp39-cp39-macosx_10_9_x86_64.whl 2ef0fbc8bcf102c1998c1f16f15befe7cffba90895d6e84861cd6c6a33fb54f6 scipy-1.8.1-cp39-cp39-macosx_12_0_arm64.whl 83606129247e7610b58d0e1e93d2c5133959e9cf93555d3c27e536892f1ba1f2 scipy-1.8.1-cp39-cp39-macosx_12_0_universal2.macosx_10_9_x86_64.whl 93d07494a8900d55492401917a119948ed330b8c3f1d700e0b904a578f10ead4 scipy-1.8.1-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl d3b3c8924252caaffc54d4a99f1360aeec001e61267595561089f8b5900821bb scipy-1.8.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl 70de2f11bf64ca9921fda018864c78af7147025e467ce9f4a11bc877266900a6 scipy-1.8.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl 1166514aa3bbf04cb5941027c6e294a000bba0cf00f5cdac6c77f2dad479b434 scipy-1.8.1-cp39-cp39-win32.whl 9dd4012ac599a1e7eb63c114d1eee1bcfc6dc75a29b589ff0ad0bb3d9412034f scipy-1.8.1-cp39-cp39-win_amd64.whl 9e3fb1b0e896f14a85aa9a28d5f755daaeeb54c897b746df7a55ccb02b340f33 scipy-1.8.1.tar.gz 0140dd24d14ce194fdb3df496c2190f0896ca84ac98bf5d69c3b3fb7bde1a1e9 scipy-1.8.1.tar.xz 0d27d1a8644ca83b86a370bcc8dfb73bd7d200843d0c327fef1127e8be95512d scipy-1.8.1.zip
Hello Tylor, Currently some of the projects require scipy win32 wheel packages in order to support python 3.10. Recently numpy released py310 win32 packages. we are expecting the win32 python310 scipy package also. I had posted in scipy user thread [SciPy-User] Reg scipy-1.8.1-cp310-cp310-win32.whl regarding the same. Could you please let us know the limitations for py310 win32 releases?
On Tue, May 24, 2022 at 7:04 PM V R Krishna Ambati < rameshkrishna.ambati@gmail.com> wrote:
Hello Tylor, Currently some of the projects require scipy win32 wheel packages in order to support python 3.10. Recently numpy released py310 win32 packages. we are expecting the win32 python310 scipy package also. I had posted in scipy user thread [SciPy-User] Reg scipy-1.8.1-cp310-cp310-win32.whl regarding the same. Could you please let us know the limitations for py310 win32 releases?
Hi, I planned to write a longer email about this soon, but: 32-bit Python on Windows is on its way out. It's highly unlikely that we will have 32-bit Windows wheels for 1.9.x releases (there's no suitable compiler toolchain). Hence spending time on CI now to build 32-bit 1.8.1 wheels is probably also not a great way to spend time. Why do you need these wheels? If you're stuck with 32-bit Python, can you use an older SciPy version? Cheers, Ralf
Hello Ralf, Thank you for your response. We have a framework that supports win32, win64, Linux, and it is in use by many users. We would like to add python 3.10 support to our framework, Initially, we were stuck with the supported version (win32 python 3.10) of numpy and scipy dependencies. Numpy has released win32 python 3.10 supported whl packages recently. we expected the same with scipy. If the win32 users want to use python3.10, then we will have an issue and our framework will complain about scipy dependency. Currently, we have the following options 1) Compiling win32 for python3.10 locally, which is very hard and we don't have experience 2) Drop win32 support, which is not a viable option for our customers. Regards Krishna On Tue, May 24, 2022 at 6:27 PM Ralf Gommers <ralf.gommers@gmail.com> wrote:
On Tue, May 24, 2022 at 7:04 PM V R Krishna Ambati < rameshkrishna.ambati@gmail.com> wrote:
Hello Tylor, Currently some of the projects require scipy win32 wheel packages in order to support python 3.10. Recently numpy released py310 win32 packages. we are expecting the win32 python310 scipy package also. I had posted in scipy user thread [SciPy-User] Reg scipy-1.8.1-cp310-cp310-win32.whl regarding the same. Could you please let us know the limitations for py310 win32 releases?
Hi, I planned to write a longer email about this soon, but: 32-bit Python on Windows is on its way out. It's highly unlikely that we will have 32-bit Windows wheels for 1.9.x releases (there's no suitable compiler toolchain). Hence spending time on CI now to build 32-bit 1.8.1 wheels is probably also not a great way to spend time.
Why do you need these wheels? If you're stuck with 32-bit Python, can you use an older SciPy version?
Cheers, Ralf
_______________________________________________ SciPy-Dev mailing list -- scipy-dev@python.org To unsubscribe send an email to scipy-dev-leave@python.org https://mail.python.org/mailman3/lists/scipy-dev.python.org/ Member address: rameshkrishna.ambati@gmail.com
On Thu, May 26, 2022 at 2:33 PM ambati venkata ramesh krishna < rameshkrishna.ambati@gmail.com> wrote:
Hello Ralf,
Thank you for your response. We have a framework that supports win32, win64, Linux, and it is in use by many users. We would like to add python 3.10 support to our framework, Initially, we were stuck with the supported version (win32 python 3.10) of numpy and scipy dependencies. Numpy has released win32 python 3.10 supported whl packages recently. we expected the same with scipy.
If the win32 users want to use python3.10, then we will have an issue and our framework will complain about scipy dependency.
Currently, we have the following options 1) Compiling win32 for python3.10 locally, which is very hard and we don't have experience 2) Drop win32 support, which is not a viable option for our customers.
Thanks for sharing your use case Krishna, that is useful data. One more question if you don't mind: how do you ship your framework? Do you expect users to install from PyPI, or do you ship a single installer yourself (e.g. made with PyInstaller or similar)? Thanks, Ralf
Regards Krishna
On Tue, May 24, 2022 at 6:27 PM Ralf Gommers <ralf.gommers@gmail.com> wrote:
On Tue, May 24, 2022 at 7:04 PM V R Krishna Ambati < rameshkrishna.ambati@gmail.com> wrote:
Hello Tylor, Currently some of the projects require scipy win32 wheel packages in order to support python 3.10. Recently numpy released py310 win32 packages. we are expecting the win32 python310 scipy package also. I had posted in scipy user thread [SciPy-User] Reg scipy-1.8.1-cp310-cp310-win32.whl regarding the same. Could you please let us know the limitations for py310 win32 releases?
Hi, I planned to write a longer email about this soon, but: 32-bit Python on Windows is on its way out. It's highly unlikely that we will have 32-bit Windows wheels for 1.9.x releases (there's no suitable compiler toolchain). Hence spending time on CI now to build 32-bit 1.8.1 wheels is probably also not a great way to spend time.
Why do you need these wheels? If you're stuck with 32-bit Python, can you use an older SciPy version?
Cheers, Ralf
_______________________________________________ SciPy-Dev mailing list -- scipy-dev@python.org To unsubscribe send an email to scipy-dev-leave@python.org https://mail.python.org/mailman3/lists/scipy-dev.python.org/ Member address: rameshkrishna.ambati@gmail.com
_______________________________________________ SciPy-Dev mailing list -- scipy-dev@python.org To unsubscribe send an email to scipy-dev-leave@python.org https://mail.python.org/mailman3/lists/scipy-dev.python.org/ Member address: ralf.gommers@googlemail.com
participants (4)
-
ambati venkata ramesh krishna -
Ralf Gommers -
Tyler Reddy -
V R Krishna Ambati