Update lang/python27 to 2.7.10

[brnrd]

Summary

Proposed commit-message:

lang/python: Update to 2.7.10

The FreeBSD Python team welcomes Python 2.7.10 to the Ports tree!

Announcement:	https://www.python.org/downloads/release/python-2710/
Changelog:	https://hg.python.org/cpython/raw-file/v2.7.10/Misc/NEWS

Important changes for this release:
 - Python 2.7.10 is a bug fix release of the Python 2.7.x series.
 - Fixes issues with LibreSSL 2.2+

PR: 		192511, 201088
Differential_Revision:	D2637
Reviewed_by:        koobs (mentor), antoine
Approved by:        koobs (mentor), antoine

• Patches Modules/_ssl.c to use the new define OPENSSL_NO_EGD from LibreSSL 2.2.0
• ftpmirror was removed from all versions (vulnerable)
• Patches have been re-generated using makepatch with comments and timestamps reverted

Test Plan

Build with base openssl and ports libressl, regression-test

• porttest (poudriere OK) OpenSSL:
  python27-2.7.10.log539 KBDownload
  LibreSSL:
  python27-2.7.10.log532 KBDownload
• portlint -AC (unmodified timestamps as per mat's guidance)

WARN: Makefile: [8]: possible direct use of command "python" found. use ${PYTHON_CMD} instead.
WARN: Makefile: [19]: possible direct use of command "python" found. use ${PYTHON_CMD} instead.
WARN: Makefile: for new port, make $FreeBSD$ tag in comment section empty, to make SVN happy.
WARN: /usr/freebsd-ports/lang/python27/files/patch-Doc__library__fcntl.rst: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' to ensure proper patch format.
WARN: /usr/freebsd-ports/lang/python27/files/patch-Lib__distutils__unixccompiler.py: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' to ensure proper patch format.
WARN: /usr/freebsd-ports/lang/python27/files/patch-Modules-_ctypes-libffi-configure: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' to ensure proper patch format.
WARN: /usr/freebsd-ports/lang/python27/files/patch-Modules__fcntlmodule.c: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' to ensure proper patch format.
0 fatal errors and 7 warnings found.

• make regression-test
  regression-test.out16 KBDownload
• porttest docs (OK): text
  python-doc-text-2.7.10.log11 KBDownload
  html
  python-doc-html-2.7.10.log11 KBDownload
  pdf-a4
  python-doc-pdf-a4-2.7.10.log11 KBDownload
  pdf-letter
  python-doc-pdf-letter-2.7.10.log12 KBDownload
• porttest python-tools
  py27-python-tools-2.7.10.log14 KBDownload

[antoine]

This breaks lang/python-doc-*

[antoine]

lang/python27/Makefile

107

This looks wrong in multiple ways:

• hardcoded /usr/local instead of ${LOCALBASE}
• CPPFLAGS / LDFLAGS / LIBS are already passed in environment of the configure script by bsd.port.mk
• LIBS should be preferred over LDFLAGS

[brnrd]

Regenerate patch-files with makepatch, merge comments from existing patches, replace /usr/local

[brnrd]

$ portlint -AC
WARN: Makefile: [8]: possible direct use of command "python" found. use ${PYTHON_CMD} instead.
WARN: Makefile: [19]: possible direct use of command "python" found. use ${PYTHON_CMD} instead.
WARN: Makefile: for new port, make $FreeBSD$ tag in comment section empty, to make SVN happy.
0 fatal errors and 3 warnings found.

[brnrd]

lang/python27/Makefile

107

• Fixed the /usr/local after inspecting portling -AC output.
• Without passing LDFLAGS and CPPFLAGS to configure it picks up libcrypto from base

configure:8550: checking for RAND_egd in -lcrypto
configure:8575: cc -o conftest -O2 -fno-strict-aliasing -pipe -march=native -fstack-protector -lpthread -Wl,-rpath
,/usr/local/lib -Wl,-rpath=/usr/lib:/usr/local/lib -fstack-protector conftest.c -lcrypto >&5
configure:8575: $? = 0

[antoine]

This needs an exp-run, PYTHON_REL is now 2710 which is greater than 3XX

[brnrd]

python27-2.7.10.log539 KBDownload

Ouput of poudriere testport

[antoine]

This needs :

• a patch to unbreak lang/python-doc-*
• my comments about LDFLAGS / CPPFLAGS are not addressed, you should just have to do CPPFLAGS+= and LIBS+= , no need to mess with CONFIGURE_ARGS/ENV
• a patch to fix PYTHON_REL in python.mk and/or ports

[mva]

The commit message (after @antoine's mentioned exp-run) should also include a link to the upstream changelog and highlight the most important changes for this release

lang/python27/files/patch-Doc__library__fcntl.rst

5

Avoid timestamp changes

lang/python27/files/patch-Lib__distutils__unixccompiler.py

8

Avoid timestamp changes

lang/python27/files/patch-Modules-_ctypes-libffi-configure

5

Avoid timestamp changes

lang/python27/files/patch-Modules__fcntlmodule.c

5

Avoid timestamp changes

[brnrd]

Replace CONFIGURE_ARGS

Replace CONFIGURE_ARGS in the WITH_OPENSSL_PORT test with
corresponding CPPFLAGS and LIBS as per /usr/ports/CHANGES
20110923

Tested with make regression-test, no additional issues
introduced with this change

[brnrd]

Revert timestamp changes

Timestamps where changed by make makepatch. Is there an easier
way to reroll patches with makepatch and retain the timestamps?

[brnrd]

Revert timestamp changes

[mat]

In D2637#49291, @brnrd wrote:

Revert timestamp changes

Timestamps where changed by make makepatch. Is there an easier
way to reroll patches with makepatch and retain the timestamps?

Yes, svn revert the files that only have the timestamp changes.

[antoine]

Exp-run results:

http://package18.nyi.freebsd.org/jail.html?mastername=101amd64-default-python2710

5 new failures:

+ {"origin"=>"lang/python-doc-html", "pkgname"=>"python-doc-html-2.7.10", "phase"=>"fetch", "errortype"=>"distinfo_update"}
+ {"origin"=>"lang/python-doc-pdf-a4", "pkgname"=>"python-doc-pdf-a4-2.7.10", "phase"=>"fetch", "errortype"=>"distinfo_update"}
+ {"origin"=>"lang/python-doc-pdf-letter", "pkgname"=>"python-doc-pdf-letter-2.7.10", "phase"=>"fetch", "errortype"=>"distinfo_update"}
+ {"origin"=>"lang/python-doc-text", "pkgname"=>"python-doc-text-2.7.10", "phase"=>"fetch", "errortype"=>"distinfo_update"}
+ {"origin"=>"lang/python-tools", "pkgname"=>"py27-python-tools-2.7.10", "phase"=>"patch", "errortype"=>"???"}

Failure logs:

http://package18.nyi.freebsd.org/data/101amd64-default-python2710/2015-05-26_08h42m01s/logs/errors/python-doc-html-2.7.10.log
http://package18.nyi.freebsd.org/data/101amd64-default-python2710/2015-05-26_08h42m01s/logs/errors/python-doc-pdf-a4-2.7.10.log
http://package18.nyi.freebsd.org/data/101amd64-default-python2710/2015-05-26_08h42m01s/logs/errors/python-doc-pdf-letter-2.7.10.log
http://package18.nyi.freebsd.org/data/101amd64-default-python2710/2015-05-26_08h42m01s/logs/errors/python-doc-text-2.7.10.log
http://package18.nyi.freebsd.org/data/101amd64-default-python2710/2015-05-26_08h42m01s/logs/errors/py27-python-tools-2.7.10.log

[brnrd]

Add lang/python-doc-* and lang/python-tools changes

[brnrd]

Silence portlint a bit

[xmj]

What's left to do? How do I help?

-J

[brnrd]

Re-upload patches

[brnrd]

Avoid timestamp changes

[brnrd]

Replace HAVE_RAND_EGD by OPENSSL_NO_EGD

[koobs]

Pending inline patch references as breadcrumbs after @brnrd creates an upstream issue. I didn't mean to accept the review, but overall its I'm much more confident after removal of the global FLAGS block after our conversation on IRC

[brnrd]

Add reference to upstream OPENSSL_NO_EGD change

[brnrd]

Remove ambiguity

[koobs]

SHIP IT BERNARD! Very nice work

[koobs]

Committed by @brnrd in rP391238, rP391239 and rP391240

@brnrd, can you close?

[brnrd]

And learned another thing... Can't close when there's open reviewers. Removed antoine and portmgr and can now close.