build process adds CWD (null entry) to PYTHONPATH if PYTHONPATH is set in the build environment

[bitdancer]

BPO	12282
Nosy	@ned-deily, @merwok, @bitdancer

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2011-06-08.03:10:57.678>
labels = ['build']
title = 'build process adds CWD (null entry) to PYTHONPATH if PYTHONPATH is set in the build environment'
updated_at = <Date 2011-06-09.13:12:03.817>
user = 'https://github.com/bitdancer'

bugs.python.org fields:

activity = <Date 2011-06-09.13:12:03.817>
actor = 'eric.araujo'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Build']
creation = <Date 2011-06-08.03:10:57.678>
creator = 'r.david.murray'
dependencies = []
files = []
hgrepos = []
issue_num = 12282
keywords = []
message_count = 4.0
messages = ['137894', '137897', '137909', '137911']
nosy_count = 3.0
nosy_names = ['ned.deily', 'eric.araujo', 'r.david.murray']
pr_nums = []
priority = 'normal'
resolution = None
stage = 'needs patch'
status = 'open'
superseder = None
type = None
url = 'https://bugs.python.org/issue12282'
versions = ['Python 2.7', 'Python 3.2', 'Python 3.3']

[bitdancer]

After a make distclean I'm seeing the following output during a build of default. It is possible my 2.7 install is borked, but I don't understand why python is apparently being called during a gcc step. The build seems to succeed.

gcc -pthread -c  -g -O0 -Wall -Wstrict-prototypes    -I. -IInclude -I./Include    -DPy_BUILD_CORE \
      -DHGVERSION="\"`LC_ALL=C hg id -i .`\"" \
      -DHGTAG="\"`LC_ALL=C hg id -t .`\"" \
      -DHGBRANCH="\"`LC_ALL=C hg id -b .`\"" \
      -o Modules/getbuildinfo.o ./Modules/getbuildinfo.c
Traceback (most recent call last):
  File "/usr/lib/python2.7/site.py", line 62, in <module>
    import os
  File "/usr/lib/python2.7/os.py", line 398, in <module>
    import UserDict
  File "/usr/lib/python2.7/UserDict.py", line 83, in <module>
    import _abcoll
  File "/usr/lib/python2.7/_abcoll.py", line 11, in <module>
    from abc import ABCMeta, abstractmethod
ImportError: cannot import name ABCMeta
Traceback (most recent call last):
  File "/usr/lib/python2.7/site.py", line 62, in <module>
    import os
  File "/usr/lib/python2.7/os.py", line 398, in <module>
    import UserDict
  File "/usr/lib/python2.7/UserDict.py", line 83, in <module>
    import _abcoll
  File "/usr/lib/python2.7/_abcoll.py", line 11, in <module>
    from abc import ABCMeta, abstractmethod
ImportError: cannot import name ABCMeta
Traceback (most recent call last):
  File "/usr/lib/python2.7/site.py", line 62, in <module>
    import os
  File "/usr/lib/python2.7/os.py", line 398, in <module>
    import UserDict
  File "/usr/lib/python2.7/UserDict.py", line 83, in <module>
    import _abcoll
  File "/usr/lib/python2.7/_abcoll.py", line 11, in <module>
    from abc import ABCMeta, abstractmethod
ImportError: cannot import name ABCMeta

[ned-deily]

It's being called for hg:

-DHGVERSION=""LC_ALL=C hg id -i ."" ...

[bitdancer]

As far as I can tell my python2.7 installation is fine and bug free. hg is working fine when called from the command line.

[bitdancer]

It turns out that I had a stray abc.py file in my current working directory as the result of some previous tests of module-load-order rules. That by itself wouldn't have triggered the problem, but in addition, I have PYTHONPATH set in my environment because I have the hg whitespace hook installed. Apparently the build process turns a non-empty PYTHONPATH into a null entry in the final computed PYTHONPATH it uses when hg is called, and this results in anything located in the CWD being loaded.

I don't think this rises quite to the level of a security issue, since it applies only at build time, but it certainly seems like a bug.