i686-pc-linux-gnu-gcc -pthread -shared -Wl,-O1 -march=i686 -O2 -pipe build/temp.linux-i686-2.6/recodemodule.o -L/usr/lib -lrecode -lpython2.6 -o build/lib.linux-i686-2.6/_recode.so >>> Source compiled. >>> Test phase [not enabled]: dev-python/python-bibtex-1.2.4 >>> Install python-bibtex-1.2.4 into /var/tmp/portage/dev-python/python-bibtex-1.2.4/image/ category dev-python running install running check running build running build_ext error: tests/simple-2.bib: line 1: unexpected result: error: tests/simple-2.bib: line 1: obtained ('entry', ('gobry03', 'article', 0, 1, {'journal': (0, 0, 'My diary'), 'title': (2, 0, 'A simple example'), 'year': (3, 0, '2003', 2003, 0, 0), 'author': (1, 0, 'Fr\xe9d\xe9ric Gobry', [(None, 'Fr\xe9d\xe9ric', 'Gobry', None)])})) error: tests/simple-2.bib: line 1: expected ('entry', ('gobry03', 'article', 0, 1, {'journal': (0, 0, 'My diary'), 'year': (3, 0, '2003', 2003, 0, 0), 'author': (1, 0, 'Fr\xe9d\xe9ric Gobry', [(None, 'Fr\xe9d\xe9ric', 'Gobry', None)]), 'title': (2, 0, 'A simple example')})) error: tests/simple.bib: line 1: unexpected result: error: tests/simple.bib: line 1: obtained ('gobry03', 'article', 0, 1, {'journal': (0, 0, 'My diary'), 'title': (2, 0, 'A simple example'), 'year': (3, 0, '2003', 2003, 0, 0), 'author': (1, 0, 'Fr\xe9d\xe9ric Gobry', [(None, 'Fr\xe9d\xe9ric', 'Gobry', None)])}) error: tests/simple.bib: line 1: expected ('gobry03', 'article', 0, 1, {'journal': (0, 0, 'My diary'), 'year': (3, 0, '2003', 2003, 0, 0), 'author': (1, 0, 'Fr\xe9d\xe9ric Gobry', [(None, 'Fr\xe9d\xe9ric', 'Gobry', None)]), 'title': (2, 0, 'A simple example')}) testsuite: 12 checks, 3 failures error: check failed. * * ERROR: dev-python/python-bibtex-1.2.4 failed. * Call stack: * ebuild.sh, line 48: Called src_install * environment, line 2464: Called distutils_src_install * environment, line 717: Called die * The specific snippet of code: * ${python} setup.py install --root="${D}" --no-compile "$@" || die "python setup.py install failed"; * The die message: * python setup.py install failed * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/log/portage/dev-python:python-bibtex-1.2.4:20090331-205354.log'. * The ebuild environment file is located at '/var/tmp/portage/dev-python/python-bibtex-1.2.4/temp/environment'. * This is with python2.6 but same error happens with python2.5
This seems to be a upstream bug due to dirty test programming. Note the different order of the BibTeX fields in python dictionary of the failing test cases. For example, the obtained "('entry', ('gobry03', 'article', 0, 1, {'journal': (0, 0, 'My diary'), 'title': (2, 0, 'A simple example'), 'year': (3, 0, '2003', 2003, 0, 0), 'author': (1, 0, 'Fr\xe9d\xe9ric Gobry', [(None, 'Fr\xe9d\xe9ric', 'Gobry', None)])}))" has just a permuted order of the dictionary elements with respect to "('entry', ('gobry03', 'article', 0, 1, {'journal': (0, 0, 'My diary'), 'year': (3, 0, '2003', 2003, 0, 0), 'author': (1, 0, 'Fr\xe9d\xe9ric Gobry', [(None, 'Fr\xe9d\xe9ric', 'Gobry', None)]), 'title': (2, 0, 'A simple example')}))". Thus, I guess in python2.6/2.5 the ordering algorithm for dictionaries has changed. Right?
Upstream bug already filed, see http://sourceforge.net/tracker/?func=detail&aid=2474782&group_id=4825&atid=104825
Created attachment 198981 [details] Working version
(In reply to comment #3) > Created an attachment (id=198981) [edit] > Working version > I bodged the test case by telling it to eval() the results (Evil, I know.). I needed it working.
Should not it version 1.2.4-r1 be marked as unstable as long the upstream bug is not fixed? I also want use it (with pybliographer) but it's confusing if a stable marked package does not install with current version of its base library (python) Tests are nice as long they work...thank you Peter for the fix my2c
Looks pretty dead upstream as the frontpage indicates that the project is looking for a new maintainer (see http://pybliographer.org/ ).
Hello? If the emerge fails for everybody, then shouldn't this be ~x86 or masked? Either that or the -r1 version is taken as an acceptable workaround.
Solution by Peter Warner confirmed working here on amd64. Any chance we can get this into portage?
Fixed.
