The standard math library (libm) may follow IEEE-754 recommendation to
include an implementation of sinPi(), i.e. sinPi(x):=sin(pi*x).
And this triggers a name clash, found by FreeBSD developer
Steve Kargl, who worked on putting sinpi into libm used on FreeBSD
(it has to be named "sinpi", not "sinPi", cf. e.g.
https://en.cppreference.com/w/c/experimental/fpext4).
- python2.7 and > 3.6 are already fixed
PR: 232792
Submitted by: Steve Kargl <sgk@troutmask.apl.washington.edu>, Dima Pasechnik <dimpase+freebsd@gmail.com>
Approved by: python (maintainer timeout)
Obtained from: b545ba0a50
A single close(fd) syscall is cheap, but when MAXFDS (maximum file
descriptor number) is high, the loop calling close(fd) on each file
descriptor can take several milliseconds.
The default value of subprocess.Popen "close_fds" parameter changed to True
in Python 3. Compared to Python 2, close_fds=True can make Popen 10x
slower: see bpo-37790 [1]
The present workaround on FreeBSD to improve performance is to load and
mount the fdescfs kernel module, but this is not enabled by default.
This change adds minimum viable (and upstreamable) closefrom(2) syscall
support to Python's subprocess and posix modules, improving performance
significantly for loads that involve working with many processes, such as
diffoscope, ansible, and many others.
For additional optimizations, upstream recently (3.8) landed posix_spawn(2)
support [3] and has stated that they will adopt close_range(2) after Linux
merges it [4]. Linux/FreeBSD developers are already collaborating on
ensuring compatible implementations, with FreeBSD's implementation pending
in D21627. [5]
Thank you emaste, cem, kevans for providing analysis, input,
clarifications, comms/upstream support and patches.
[1] https://bugs.python.org/issue37790
[2] https://bugs.python.org/issue38061
[3] https://bugs.python.org/issue35537
[4] https://lwn.net/Articles/789023/
[5] https://reviews.freebsd.org/D21627
Additional References:
https://bugs.python.org/issue8052https://bugs.python.org/issue11284https://bugs.python.org/issue13788https://bugs.python.org/issue1663329https://www.python.org/dev/peps/pep-0446/
PR: 242274, 221700
Submitted by: kevans (emaste, cem)
Approved by: koobs (python (maintainer), santa)
Simplify some ports where DragonFlyBSD no longer needs to be special-cased.
Submitted by: rene
Reviewed by: bapt, jbeich
Differential Revision: https://reviews.freebsd.org/D17724
ports r393217 via bug 200622 [1] originally set MAKE_JOBS_UNSAFE=yes due to
incorrect uses of recursive make [2], causing intermittent build failures when
run with multiple jobs (-jN).
Upstream committed a fix for the issue in default (3.6, at the time), 3.5 and
2.7 which are now contained in all released lang/python?? port versions. 3.4 did
not receieve a backport merge.
lang/python3.5+ ports inadvertently inherited MAKE_JOBS_UNSAFE=yes, via
repocopies from lang/python34 on their creation, when they were infact safe to
use with -j.
Remove MAKE_JOBS_UNSAFE in all lang/python?? ports except python34 accordingly.
[1] https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=200622
[2] https://bugs.python.org/issue22359
PR: 232308
Reported by: cem
Reviewed by: cem
Approved by: koobs (python)
MFH: 2018Q4
Differential Revision: D17579
When python3.?-config is symlinked to another location it starts
outputting bogus paths. For example
$ pwd
/home/tobias
$ python3.6-config --includes
-I/usr/local/include/python3.6m -I/usr/local/include/python3.6m
$ ln -s /usr/local/bin/python3.6-config python3-config
$ ./python3-config --includes
-I/home/include/python3.6m -I/home/include/python3.6m
This breaks ports trying to use BINARY_ALIAS together with
python3.?-config. Apply a patch to resolve the symlink first before
trying to find the install prefix.
PR: 229749
Submitted by: tobik
Reviewed by: antoine, miwi
Approved by: python (miwi)
The change from /usr/local/bin/python to /usr/local/bin/python3.6 is already done by USES=shebangfix.
% head -1 /usr/local/lib/python3.6/cgi.py
#!/usr/local/bin/python3.63.6
- Move BROKEN_SSL upward
- Sort USES
- Remove CPE_*: all of them are default values
- Update PLIST_FILES: do not use %%
- Update http:// links in Makefile comments and patch files
Ports using USE_PYTHON=distutils are now flavored. They will
automatically get flavors (py27, py34, py35, py36) depending on what
versions they support.
There is also a USE_PYTHON=flavors for ports that do not use distutils
but need FLAVORS to be set. A USE_PYTHON=noflavors can be set if
using distutils but flavors are not wanted.
A new USE_PYTHON=optsuffix that will add PYTHON_PKGNAMESUFFIX has been
added to cope with Python ports that did not have the Python
PKGNAMEPREFIX but are flavored.
USES=python now also exports a PY_FLAVOR variable that contains the
current python flavor. It can be used in dependency lines when the
port itself is not python flavored. For example, deskutils/calibre.
By default, all the flavors are generated. To only generate flavors
for the versions in PYTHON2_DEFAULT and PYTHON3_DEFAULT, define
BUILD_DEFAULT_PYTHON_FLAVORS in your make.conf.
In all the ports with Python dependencies, the *_DEPENDS entries MUST
end with the flavor so that the framework knows which to build/use.
This is done by appending '@${PY_FLAVOR}' after the origin (or
@${FLAVOR} if in a Python module with Python flavors, as the content
will be the same). For example:
RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}six>0:devel/py-six@${PY_FLAVOR}
PR: 223071
Reviewed by: portmgr, python
Sponsored by: Absolight
Differential Revision: https://reviews.freebsd.org/D12464
this environment in poudriere, CC is not set to the default of /usr/bin/cc and
a cross-compile toolchain is used. We need to hand edit this so that the run
time configuration for python matches what the FreeBSD base system provides.
PR: 208282
Submitted by: manu
Approved by: portmgr (mat)
Use readline from ports (USES= readline:port) and patch
setup.py to ignore readline from base. The patch is necessary for
FreeBSD < 1100000, as after this the readline library became an
INTERNALLIB, see base r268461 [1]
Link devel/readline against termcapw instead of termcap is part of
this change, see ports r444463 [2]
Note that this is the **ports** approach for getting Python curses
module working with Unicode. The other way is splitting libncurses
into separate libncurses and libtinfo in base, for which an open
issue exists [3].
Apart from Python language ports, at least www/rtv and
sysutils/py-ranger ports have been tested to work correctly
(display Unicode) after this change.
[1] https://svnweb.freebsd.org/changeset/base/268461
[2] https://svnweb.freebsd.org/changeset/ports/444463
[3] https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=197317
PR: 171246, 197317
Reported by: Vitaly Magerya <vmagerya gmail com>
Reviewed by: garga, koobs, miwi, sunpoet
Approved by: garga (mentor), sunpoet (python, with hat)
Differential Revision: https://reviews.freebsd.org/D11127
pkg-static: Unable to access file /wrkdirs/usr/ports/lang/python36/work/stage/usr/local/lib/python3.6/lib-dynload/_ssl.so:No such file or directory
Sponsored by: Absolight
Users with a GDB that supports [1] Python extensions will automatically
load the extra debugging extensions when debugging programs that are
linked with libpythonX.Y.so.foo.
This enables extensions like 'py-bt' and 'py-frame' as described in
the Fedora Wiki Article: Easier Python Debugging [2], which can be
useful for debugging Python program state from crashes in C extensions,
for example.
[1] PYTHON option enabled in devel/gdb
[2] https://fedoraproject.org/wiki/Features/EasierPythonDebugging
PR: 203021
Submitted by: cem
Reviewed by: mat, koobs (python)
Approved by: koobs (python)
Differential Revision: D10398
