Search Criteria
Package Details: python37 3.7.13-1
Package Actions
| Git Clone URL: | https://aur.archlinux.org/python37.git (read-only, click to copy) |
|---|---|
| Package Base: | python37 |
| Description: | Major release 3.7 of the Python high-level programming language |
| Upstream URL: | https://www.python.org/ |
| Keywords: | python python3 |
| Licenses: | custom |
| Provides: | python |
| Submitter: | 5long |
| Maintainer: | greut (jrd) |
| Last Packager: | greut |
| Votes: | 28 |
| Popularity: | 0.82 |
| First Submitted: | 2019-11-14 21:23 (UTC) |
| Last Updated: | 2022-03-17 08:28 (UTC) |
Dependencies (16)
- bzip2 (bzip2-git, bzip2-rustify-git, bzip2-with-lbzip2-symlinks)
- expat (expat-git)
- gdbm (gdbm-git)
- libffi (libffi-minimal-git, libffi-git)
- libnsl (libnsl-git)
- libxcrypt (libxcrypt-git)
- openssl (libressl-git, openssl-static, openssl-hardened, quictls-openssl, openssl-git)
- zlib (zlib-static, zlib-git, zlib-ng, zlib-ng-compat-git)
- bluez-libs (bluez-git) (make)
- mpdecimal (make)
- sqlite (sqlite-replication, sqlite-minimal-git, sqlite-fts3-parenthesis, sqlite-fossil) (make)
- tk (tk85) (make)
- mpdecimal (optional) – for decimal
- sqlite (sqlite-replication, sqlite-minimal-git, sqlite-fts3-parenthesis, sqlite-fossil) (optional)
- tk (tk85) (optional) – for tkinter
- xz (xz-git) (optional) – for lzma
Required by (10904)
- 0x0uploader (requires python)
- 10ff-git (requires python)
- 10moons-driver-t503 (requires python)
- 10moons-driver-t503 (requires python) (make)
- 2048-py (requires python)
- 3to2 (requires python)
- 4scanner (requires python)
- 64gram-desktop (requires python) (make)
- 64gram-desktop-no-ad (requires python) (make)
- a2ln (requires python)
- a7800 (requires python) (make)
- aarch64-none-elf-gdb (requires python)
- aard2-slob (requires python)
- abeceda_convertor-git (requires python)
- abraca (requires python) (make)
- absolutely-proprietary (requires python)
- abv-calc-git (requires python)
- abzer-git (requires python)
- acalc-git (requires python)
- accountable2you-bin
Latest Comments
tinytangent commented on 2022-04-30 05:15 (UTC) (edited on 2022-04-30 05:15 (UTC) by tinytangent)
The build process succeeded but I noticed that some unit tests seem to fail ([37/40] test_time -- test_tabnanny failed):
greut commented on 2022-03-24 06:07 (UTC)
@hadian90 I'll let you search how to deal with importing the provided GPG key.
hadian90 commented on 2022-03-24 05:05 (UTC) (edited on 2022-03-24 05:08 (UTC) by hadian90)
failed install on Manjaro
hwangh95 commented on 2021-11-22 14:03 (UTC)
_decimal module doesn't build without the patch in python38 https://aur.archlinux.org/cgit/aur.git/tree/mpdecimal-2.5.1.patch?h=python38
anatolik commented on 2021-11-19 02:34 (UTC) (edited on 2021-11-19 02:35 (UTC) by anatolik)
It needs rebuild due to libldap bump
bacteriostat commented on 2021-10-20 06:20 (UTC) (edited on 2021-10-20 06:47 (UTC) by bacteriostat)
libffi was upgraded 3.4.2-4 and Jupyter notebook no more launches for me. It is looking for libffi.so.7 while the system has libffi.so.8.
I am not sure if this issue goes here or somewhere upstream.
FIXED: Rebuilding the package fixed the issue for me. Leaving it here for noobs out there like me.
avzb_arch commented on 2021-07-08 06:48 (UTC) (edited on 2021-07-08 06:50 (UTC) by avzb_arch)
Use this works well
gpg --keyserver keyserver.ubuntu.com --recv-key 0D96DF4D4110E5C43FBFB17F2D347EA6AA65421Djrd commented on 2021-07-03 09:01 (UTC)
I had no problem with this package (in french):
greut commented on 2021-07-03 07:48 (UTC)
@Krtko some pacman wrappers are able to import the PGP/GPG keys for you. https://wiki.archlinux.org/title/AUR_helpers#Pacman_wrappers
Afaik, 3.7.11 was signed by the same person/key than the previous one. Sorry for the trouble.
robibert commented on 2021-07-02 13:01 (UTC)
I specified the key server
gpg --keyserver hkp://pgp.mit.edu:80 --recv-keys 0D96DF4D4110E5C43FBFB17F2D347EA6AA65421Dand it worked for meKrtko commented on 2021-07-01 04:58 (UTC)
...
greut commented on 2021-01-20 15:42 (UTC)
@user20159 me neither. I've made it more like Python38, let me know if it breaks anything on your side. Thanks for the heads up.
user20159 commented on 2021-01-19 16:46 (UTC)
I'm not sure if this question has been asked already, I've checked the comments but couldn't see anything. How can it be that python37 has llvm gdb xorg-server-xvfb as additional dependencies over python38 or python36? Furthermore, build time is around twice as long as for python38 or python36. Are the other ones missing something?
jhincks commented on 2020-09-28 20:15 (UTC)
build failed with "/usr/bin/xvfb-run: line 186: kill: (1576374) - No such process"
xvfb-run says in its help, that "-a" is deprecated, one should use "-d" instead. I replaced the two occurences in the PKGBUILD, then the build process was successfull.
RealGecko commented on 2020-09-22 08:59 (UTC)
I used this command to import the key
gpg --keyserver pool.sks-keyservers.net --receive-keys 0D96DF4D4110E5C43FBFB17F2D347EA6AA65421D
Leeo97one commented on 2020-09-18 20:14 (UTC) (edited on 2020-09-18 20:15 (UTC) by Leeo97one)
I have imported and signed the GPG key:
pacman-key --list-keysBut I'm still getting a GPG error and I'm not sure why:
greut commented on 2020-09-10 19:45 (UTC)
@DMolt, it make sense to follow what is done for python35 and python36 in that regard and go without ensurepip.
DMolt commented on 2020-09-10 07:59 (UTC)
Hi, I use python 3.7 for some projects with virtualenv and 3.8 for other and system. To install this I have to change --with-ensurepip with --without-ensurepip
Davide
greut commented on 2020-09-07 09:56 (UTC)
@surrealistic above in the logs, you should see why the _posixsubprocess module is not building successfully.
surrealistic commented on 2020-09-07 09:27 (UTC) (edited on 2020-09-07 09:31 (UTC) by surrealistic)
Hi! I'm having this issue with subprocess while tying to upgrade python37 and also after uninstalling it to try a clean installation of it.
It's like it can't find or recognize subprocess. I searched on Google about the error and I've tried to install subprocess32 with python3.7 -m pip install subprocess32 but it returns the same error. After all those attempts I decided to try a clean installation so I removed python3.7 to install it again instead of upgrading it, but it keeps returning the same error.
jrd commented on 2020-09-03 10:46 (UTC)
Thanks greut, and thanks 5long for initial package.
greut commented on 2020-09-02 19:43 (UTC)
@jrd I've taken the patch from the bug mentioned (https://bugs.python.org/issue41302) and added you as a co-maintainer.
GeorgP commented on 2020-09-02 08:18 (UTC)
You can click "Adopt Package" in "Package Actions" in the top right. Then it's yours. :D
jrd commented on 2020-09-01 12:11 (UTC)
If 5long don't/won't update the package, I could take over if that's possible (I don't know how to transfer AUR package owership…)
jrd commented on 2020-09-01 09:48 (UTC) (edited on 2020-09-01 10:11 (UTC) by jrd)
I made it compile and install with the following diff for
PKGBUILD:This is because libmpec is now at version 2.5+ in Arch and this is not supported for python version prior to 3.8. So the one bundled within the tarball should be used.
The issue that let me understand the problem: https://bugs.python.org/issue41302
PRAGMA commented on 2020-08-28 19:07 (UTC)
I'm also having the same issue as jrd.
jrd commented on 2020-08-21 12:51 (UTC)
Build fails for me, this seems related to
fakeroot. Probably an incompatibility between twoLD_LIBRARY_PATH/LD_PRELOADusages.First weird messages:
And last messages before stopping
buildstage:Edgeworth commented on 2020-08-14 02:13 (UTC)
To elaborate on alexz comment, I had to follow the instructions from keybase/nad to import the GPG key
alexz commented on 2020-08-12 20:14 (UTC)
I guess, another fix is to import the public key:
I don't think this package is Out-of-Date ? How do you unflag it?
alexz commented on 2020-08-12 20:03 (UTC) (edited on 2020-08-12 20:10 (UTC) by alexz)
Failing with GPG error, too:
seems to be related to:
And this release was signed by somebody else, so for me it works with:
justicehunter commented on 2020-08-11 08:47 (UTC)
@GhostThreader139 yup, same here. Signature problem
GhostThreader139 commented on 2020-08-09 06:52 (UTC)
Anyone else having a problem with the key, as if it's unknown???
holishing commented on 2020-08-05 12:54 (UTC)
@emanuelfontelles What's your gpg signature's problem? It seems to work ok in my machine
timo42 commented on 2020-07-15 22:25 (UTC) (edited on 2020-07-15 22:26 (UTC) by timo42)
I had exactly the same build error and re-installing the package didn't work for me. However, replacing
--with-ensurepipby--without-ensurepip(the other way round as @seyn6gk suggested) in the PKGBUILD fixed it for me. Not sure if changing this option has any unwanted side-effects though.manofsteal123 commented on 2020-07-12 22:25 (UTC)
"I got the permission denied error reported below when attempting an upgrade but removing the package and reinstalling it worked for me."
I had to remove python37-pillow & python37 in that order and used "makepkg -s" to build the Download snapshot of python37 & python37-pillow in that order
depieri.carlo commented on 2020-07-12 10:29 (UTC)
If it's an upgrade issue, my current package version could help: 3.7.5-2.
mjakob422 commented on 2020-07-12 09:50 (UTC)
I got the permission denied error reported below when attempting an upgrade but removing the package and reinstalling it worked for me.
depieri.carlo commented on 2020-07-11 17:36 (UTC) (edited on 2020-07-11 17:38 (UTC) by depieri.carlo)
The build is failing for me too, both with
makepkgand withyay. Can confirm that--with-ensurepipis already there in the PKGBUILD.seyn6gk commented on 2020-07-11 05:36 (UTC)
@5long I've got the PKGBUILD per
pikaur -Gcommand, this could be the issue.5long commented on 2020-07-11 04:22 (UTC)
@seyn6gk: this is confusing. I've already enabled the
--with-ensurepipoption back in 2019.12. The patch you've attached shouldn't even apply.seyn6gk commented on 2020-07-10 17:46 (UTC) (edited on 2020-07-10 17:46 (UTC) by seyn6gk)
apply the patch to make python37 to build again, get the idea from python38 package.
seyn6gk commented on 2020-07-10 17:09 (UTC)
build with
makepkgandpikaurboth failed with the error5long commented on 2020-07-10 14:51 (UTC)
Hi, @maln0ir and @javi12
I've verified building this package on my machine with:
1) Raw
makepkg -siwithout any AUR helpers 2)yay -S python37. Yay is a pretty popular AUR helper that I've used before. 3)aur sync -c python37. Theaurcommand is part of aurutils which is what I'm using right now.all of which works fine. I did notice the
object 'libfakeroot.so' from LD_PRELOAD cannot be preloadederror message but the build process eventually finished without problems.I've noticed that @javi12 is building with pamac, which is an AUR helper that I'm not familiar with at all. You might want to ask for help elsewhere.
javi12 commented on 2020-07-10 10:32 (UTC) (edited on 2020-07-10 10:47 (UTC) by javi12)
Build also fails here...
{{... PYTHONPATH=/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7 LD_LIBRARY_PATH=/var/tmp/pamac-build-momo/python37/src/Python-3.7.8 \ ./python -E -Wi /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/compileall.py \ -d /usr/lib/python3.7/site-packages -f \ -x badsyntax /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/site-packages ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. Listing '/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/site-packages'... PYTHONPATH=/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7 LD_LIBRARY_PATH=/var/tmp/pamac-build-momo/python37/src/Python-3.7.8 \ ./python -E -Wi -O /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/compileall.py \ -d /usr/lib/python3.7/site-packages -f \ -x badsyntax /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/site-packages ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. Listing '/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/site-packages'... PYTHONPATH=/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7 LD_LIBRARY_PATH=/var/tmp/pamac-build-momo/python37/src/Python-3.7.8 \ ./python -E -Wi -OO /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/compileall.py \ -d /usr/lib/python3.7/site-packages -f \ -x badsyntax /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/site-packages ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. Listing '/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/site-packages'... PYTHONPATH=/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7 LD_LIBRARY_PATH=/var/tmp/pamac-build-momo/python37/src/Python-3.7.8 \ ./python -E -m lib2to3.pgen2.driver /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/lib2to3/Grammar.txt ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. Generating grammar tables from /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/lib2to3/Grammar.txt Writing grammar tables to /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/lib2to3/Grammar3.7.8.final.0.pickle PYTHONPATH=/var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7 LD_LIBRARY_PATH=/var/tmp/pamac-build-momo/python37/src/Python-3.7.8 \ ./python -E -m lib2to3.pgen2.driver /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/lib2to3/PatternGrammar.txt ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. Generating grammar tables from /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/lib2to3/PatternGrammar.txt Writing grammar tables to /var/tmp/pamac-build-momo/python37/pkg/python37/usr/lib/python3.7/lib2to3/PatternGrammar3.7.8.final.0.pickle if test "xupgrade" != "xno" ; then \ case upgrade in \ upgrade) ensurepip="--altinstall --upgrade" ;; \ install|*) ensurepip="--altinstall" ;; \ esac; \ LD_LIBRARY_PATH=/var/tmp/pamac-build-momo/python37/src/Python-3.7.8 ./python -E -m ensurepip \ $ensurepip --root=/var/tmp/pamac-build-momo/python37/pkg/python37/ ; \ fi ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. ERROR: ld.so: object 'libfakeroot.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored. Looking in links: /tmp/tmp2qn8a9bg Processing /tmp/tmp2qn8a9bg/setuptools-47.1.0-py3-none-any.whl Processing /tmp/tmp2qn8a9bg/pip-20.1.1-py2.py3-none-any.whl Installing collected packages: setuptools, pip Attempting uninstall: setuptools Found existing installation: setuptools 41.2.0 Uninstalling setuptools-41.2.0: ERROR: Could not install packages due to an EnvironmentError: [Errno 13] Permission denied: '/usr/bin/easy_install' Consider using the
--useroption or check the permissions.make: *** [Makefile:1152: altinstall] Error 1 ==> ERROR: A failure occurred in package(). Aborting...}}
commented on 2020-07-09 21:42 (UTC)
Build fails for me with following error:
fakerootpackage is installed, solibfakeroot.sois available.kiengcan9999 commented on 2020-05-07 19:38 (UTC)
I failed to install python37 on Manjaro 20.0 Lysia because of error
I fixed it by
wjhandley commented on 2020-05-07 14:18 (UTC) (edited on 2020-05-07 14:51 (UTC) by wjhandley)
Currently my build fails with
make[1]: Leaving directory '...' /usr/bin/xvfb-run: line 186: kill: (107983) - No such process
Could you advise which dependency is missing?
Edit: It seems to work with xorg-server-xvfb-dev -- perhaps this should be added as a dependency?
EreTIk commented on 2020-04-27 15:02 (UTC) (edited on 2020-04-27 15:03 (UTC) by EreTIk)
Please add aarch64 architecture (arm64 support):
Test: Ok
sponrad commented on 2020-04-24 20:14 (UTC) (edited on 2020-04-24 20:20 (UTC) by sponrad)
EDIT I take it back,
yay -S python37 --rebuilddid fix this for me, if anyone else stumbles on this.I'm having a similar issue to @somini ever since libffi was updated on 2020-04-11. Rebuilding as is has not helped me fix this one I think we do need a bump here?
somini commented on 2020-04-17 21:55 (UTC) (edited on 2020-04-17 21:55 (UTC) by somini)
@5long Of course I can (I did), but for people with the old version this will break, and rebuilding doesn't help since the pkgrel doesn't change.
5long commented on 2020-04-14 10:26 (UTC)
@craynic: The PGP key in PKGBUILD is correct. PKGBUILD requires validpgpkeys to be an array of PGP fingerprints, while
2D347EA6AA65421Dis the ID (not fingerprint) of the correct key.@somini: you can just rebuild manually. See: https://wiki.archlinux.org/index.php/Arch_User_Repository#Installing_packages
somini commented on 2020-04-13 15:14 (UTC)
This needs a pkgrel bump since libffi was updated.
craynic commented on 2020-04-11 15:19 (UTC)
The PGP key declared in the PKGBUILD seems invalid. The correct one is
2D347EA6AA65421D. Ref: https://www.python.org/downloads/#pubkeysvit commented on 2020-04-08 17:53 (UTC)
I'm getting failed tests during PGO stage of compilation. Is this normal?
MarcinWieczorek commented on 2020-02-02 14:32 (UTC)
@tyreq I'm trying the fix rn. 3.7's site packages could be left because of use of pip install with sudo. Check if any package owns the files - also not updated python-* packages from AUR.
tyerq commented on 2020-01-31 10:36 (UTC)
@yurikoles @MarcinWieczorek @MangoMan had the same problem. it seems that
--with-ensurepip=upgrade(default) in./configure/Python tried to delete the previously intalled pip to reinstall it's bundled version (via pip itself) and failed, since it's lacking permissions in/usr/lib/python37/Changing
--with-ensurepipin PKGBUILD:79 to--with-ensurepip=installjust keeps the already installed pip and goes on succesfullySome further thoughts: I've had arch's system python upgrade from 3.7 to 3.8, but it kept
/usr/lib/python3.7/site-packagesfor some reason. Don't know if that is intended or not, but it resulted in this behaviour.MangoMan commented on 2020-01-29 21:01 (UTC)
I have a similar issue like MarcinWieczorek
Ivellios commented on 2020-01-10 04:45 (UTC)
@dspati Cool. Thanks!
yurikoles commented on 2020-01-08 10:32 (UTC)
@MarcinWieczorek do you have something similar with other AUR packages? Because it seems to be a local issue.
MarcinWieczorek commented on 2020-01-08 10:22 (UTC)
I'm getting an error:
I have fakeroot installed.
ajkswamy commented on 2020-01-01 13:47 (UTC)
@Ivellios Please have a look at the comment by '5long' at 2019-11-15 15:05, a few pages down.
Ivellios commented on 2019-12-30 04:43 (UTC)
==> Verifying source file signatures with gpg... Python-3.7.6.tar.xz ... FAILED (unknown public key 2D347EA6AA65421D)
5long commented on 2019-12-11 13:57 (UTC)
@yurikoles
Running tests in
check()is to check for correctness. Running tests with./configure --enable-optimizationsis for, literally, optimizations. They are different things. It's impossible to move./configure --enable-optimizationsintocheck()without breaking the build.yurikoles commented on 2019-12-10 20:04 (UTC)
@5long
why don't you use PKGBUILD check function for test? It's straightforward to then
--nocheck5long commented on 2019-12-10 18:44 (UTC)
@yurikoles: I just pushed 3.7.5-3 to AUR. Now the test suite to run is a lot smaller (from 400+ tests to 40) and the building time is cut down to roughly 1/4. Incidently, the potentially hanging test
test_socketis now skipped by default.If you really want to skip the tests, you can remove the
--enable-optimizationsoption from./configure. But beware that this will build a less performant Python. See https://stackoverflow.com/q/41405728 for an explaination.yurikoles commented on 2019-12-10 18:03 (UTC)
How to disable tests during build?
5long commented on 2019-12-08 15:53 (UTC)
@PeteAnderson: I'm not quite sure what are you trying to achieve. Your
sedcommand added inprepare()will fail, sinceMakefileisn't generated yet before running./configureSuppose the
sedcommand will succeed (say, added inbuild()right before runningmake), it'll append a dangling-xin the test command missing a required test module name and make the test command fail instantly, effectively skipping the whole test suite.Besides, the whole point of PGO build is to run the test suite first, watch how the code behaves and optimize accordingly. If we're trying to skip running the tests, it'll no longer be a PGO build and we might as well just remove
--enable-optimizationsaltogether.PeteAnderson commented on 2019-12-06 22:15 (UTC)
I found this in the makefile:
But whoever wrote the comment must have forgotten to actually add
-xto the command. So I added this to PKGBUILD inprepare():This avoids running the tests while still building an optimized build, and gave me a functional 3.7 build in about 10 minutes.
yurikoles commented on 2019-11-27 11:29 (UTC)
sorry for false-flag
5long commented on 2019-11-25 12:10 (UTC)
@Fuzzy: Yes, building with
--enable-optimizationsdoes run tests due to PGO optimization. Maybe I'll just remove this./configureoption in the next version of PKGBUILD (but I haven't decided yet).About the missing
_ctypesmodule: you can try building the package in a clean chroot, which is explained in the wiki: https://wiki.archlinux.org/index.php/DeveloperWiki:Building_in_a_clean_chroot This isn't trivial to setup and it will take up a fair bit of disk space. But it should solve most problems of "package builds differently on different machines".Fuzzy commented on 2019-11-25 04:49 (UTC)
Hi! First, thanks for this package. I also ran into troubles with last arch upgrade... But: I don't see check() section in my PKGBUILD but somehow tests are ran. There are several failed tests but worse - 'test_socket' gets stuck. I had removed --enable-optimizations and yay built and installed it successfully. At first look it works but now I have another error: No module named '_ctypes' (pacman shows that I have libffi installed...)
5long commented on 2019-11-20 15:36 (UTC) (edited on 2019-11-20 16:15 (UTC) by 5long)
I'm seriously considering bumping the pkgrel for a bunch of changes:
--with-threadsis no longer recognized, which should be removed. No questions with that.Enable
--with-ensurepip. For now, I don't see building--without-ensurepipcould yield any benefits. Theensurepipmodule and.whls of setuptools & pip are installed anyway even when building--without-ensurepip. It's just the bundled.whls might become out of date in the future, which doesn't meet Arch's "only run latest software" guideline.Running tests (either in
check()or building with--enable-optimizations) seems to be causing trouble for some people. Building with--enable-optimizationsclaims to boost performance for 10% ~ 20% (1). Personally I don't quite care about Python's performance on my dev machine that much. Another problem is: the test suite could run for as long as 30+ minutes, which is pretty slow comparing with other AUR packages that I use. But I'm guessing that people might expect this package to be "just like python in the official repo", which is built with--enable-optimizationsenabled. So I still haven't decided on this yet.maria.v commented on 2019-11-20 04:29 (UTC)
@leandro: for pip3.7 you can edit PKGBUILD to replace --without-ensurepip with --with-ensurepip and rebuild the package.
5long commented on 2019-11-18 14:45 (UTC)
@MuffinBomber: I'm no expert in building Python. Maybe the test case you are stuck on is a different test case. Not the one removed in https://bugs.python.org/issue34587.
And yes: by removing the line of
--enable-optimizationsfrom PKGBUILD, a clean rebuild (withmakepkg -Csf) no longer runs the tests on my machine.MuffinBomber commented on 2019-11-18 11:23 (UTC) (edited on 2019-11-18 11:37 (UTC) by MuffinBomber)
When building the package 'test_socket' gets stuck. Is there a way to solve this or should I just remove the '--enable-optimizations' line and skip the tests completely (while also dealing with slower executing code)?
EDIT: According to https://bugs.python.org/issue34587 that test was removed from 3.7, but somehow it is still present?
Lucki commented on 2019-11-16 14:55 (UTC)
Thanks for your insights. I'm no expert especially with python.
My package doesn't run with 3.8 anymore and upstream is (nearly) dead. Looks like this will be a bigger task to get it run again.
5long commented on 2019-11-16 13:58 (UTC)
@Lucki: I'm not quite sure if this pacakges should claim
provides=...in its PKGBUILD.Python packages are, in a way, built for a specific version of Python. Most of their files are installed into
/usr/lib/pythonX.Yand they won't run under a different version of Python. If a package only runs under Python 3.7, I'd suggest it claimspython37as its dependency. The Python version is hard-coded in the package name already.Another problem is: only Arch's official
pythonpackage provides/usr/bin/python. Other packages must not do that. Even ifpython37doesprovides=python<3.8, any python scripts starting with#!/usr/bin/pythonwon't run. So theprovides=...just become a false promise. This whole situation may be resolved once pacman implements an alternatives system (which is being discussed here: https://lists.archlinux.org/pipermail/pacman-dev/2019-October/023691.html) in the future.But still, this is my first time writing a "serious" PKGBUILD. And by writing I mean "copying from official repo and
python36". What I've said above may very well be wrong. Feel free to point out any mistakes.Lucki commented on 2019-11-16 12:52 (UTC) (edited on 2019-11-16 12:54 (UTC) by Lucki)
Isn't this providing
python<3.8or similar?https://wiki.archlinux.org/index.php/PKGBUILD#provides
chrismax commented on 2019-11-16 10:21 (UTC)
@5long Thanks for updating the build file. Now it's working for me without any troubles.
lainproliant commented on 2019-11-15 22:53 (UTC)
@nitrobay Please try what @5long did below. You will need to download the PKGBUILD file yourself, then remove the
check()section, then runmakepkg -s. I ran into the issue that some post-build unit tests refused to pass, and removing thecheck()section allowed me to proceed successfully.nitrobay commented on 2019-11-15 15:05 (UTC)
Even with the very latest version of the aur database, tests are runned despite !test in my makpkg.conf ! Very huge deal I cannot work because I've done the very big mistake of updating everything including python so now the tools used for my work does not work. I will never use arch linux again. It is so stupid you cannot downgrade with it !!!!!!!!! Such a basic feature make the PC unusable. Any update can just break your workflow forever and then you cannot work. I hate arch linux.
5long commented on 2019-11-15 14:05 (UTC)
@ashep: you might wanna manually import the GPG key from the upstream maintainer. See the wiki: https://wiki.archlinux.org/index.php/PKGBUILD#validpgpkeys for how to do it.
Also, if you're using some kind of AUR helper (in my case, it's
yay), it should automatically import the GPG key for you.5long commented on 2019-11-15 14:03 (UTC)
Thanks for all your feed back. I just removed the
check()section with 3.7.5-2.At first I thought that it's up to the users who decide if they would run check() or not when building a package. You can skip the check() phase in makepkg.conf(5). What drives me made the decision is: some of the tests just don't pass and I probably can't fix them anyway.
But still, the tests are gonna run in
build()phase since the default build method is "PGO build", which is still kinda slow.chrismax commented on 2019-11-15 13:21 (UTC)
@Nomalag I also noticed this when I switched back to python36.
Nomalag commented on 2019-11-15 12:07 (UTC)
Is the check section really necessary? The python36 doesn't have it and the tests take more than an hour on my machine.
chrismax commented on 2019-11-15 09:59 (UTC)
Hi thanks for submitting this python package :) During the build I encountered this failure:
== Tests result: FAILURE ==
395 tests OK.
1 test failed: test_curses
9 tests skipped: test_devpoll test_kqueue test_msilib test_ossaudiodev test_startfile test_winconsoleio test_winreg test_winsound test_zipfile64
Total duration: 23 min 56 sec Tests result: FAILURE
ashep commented on 2019-11-15 02:40 (UTC) (edited on 2019-11-15 02:40 (UTC) by ashep)
==> Validating source files with sha512sums... Python-3.7.5.tar.xz ... Passed Python-3.7.5.tar.xz.asc ... Skipped dont-make-libpython-readonly.patch ... Passed 0001-compileall-Fix-ddir-when-recursing.patch ... Passed ==> Verifying source file signatures with gpg... Python-3.7.5.tar.xz ... FAILED (unknown public key 2D347EA6AA65421D) ==> ERROR: One or more PGP signatures could not be verified! ==> ERROR: Makepkg was unable to build python37.