Created attachment 351822 [details, diff] patch Point 1. the tarball from pypi had done it again, left some content source out. plse substitute SRC_URI="mirror://pypi/j/jedi/jedi-${MY_PV}.tar.gz" with SRC_URI="https://github.com/davidhalter/${PN}/archive/v${PV}.tar.gz -> ${P}.tar.gz" Point 2. You can now run the test suite, like so; test/test_regression.py ......................................................................................................................................................................... test/test_regression.py ............................................................................................................................................................................... test/test_regression.py ............................................................................... test/test_regression.py .............................................. ================== 778 passed in 45.04 seconds =============================== ........... ================== 778 passed in 45.63 seconds =============================== ... ================== 778 passed in 45.90 seconds =============================== ... ================== 778 passed in 46.89 seconds =============================== granted it's a bit hard when half the test folder source is missing.
Comment on attachment 351822 [details, diff] patch >+python_test() { >+ PYTHONPATH="${PYTHONPATH}${S}"/test/ py.test test || die "Tests failed under ${EPYTHON}" >+} When appending to PYTHONPATH, make sure to include a colon; there is no guarantee that the existing PYTHONPATH value will end with a colon. Example: PYTHONPATH="${PYTHONPATH}:${S}/test/" or, if you are being extra careful: PYTHONPATH="${PYTHONPATH}${PYTHONPATH+:}${S}/test"
er yep, ack. oh and Point 3. doc build for good measure
Can I commit as is? If you've taken the path I skirted, that of awaiting upstream to top up the tarball, please say so and change to IN_PROGRESS. timeout period 5 days, then python team will bump it.
if there is no response within a week on the github issue, we can switch to the github tarballs the attached patch does not incorporate #c1
righty
(In reply to Ian Delaney from comment #5) > righty Please don't, see the github issue. It's being dealt with.
I submitted a pull request upstream; hopefully that will grease the wheels a bit.
cool
yes ditto. @tkf @dbrgn When we're doing this, we should probably also add sith.py and .coveragerc, what do you think? ugh
upstream will include the fixes in the next release if you care A LOT about it, then create a tarball with the missing files and push it to the gentoo mirrors, otherwise this is going to be RESOLVED UPSTREAM
will be fixed on version bump
(In reply to Julian Ospald (hasufell) from comment #11) > will be fixed on version bump that works fine
+*jedi-0.7.0 (18 Aug 2013) + + 18 Aug 2013; Julian Ospald <hasufell@gentoo.org> +jedi-0.7.0.ebuild: + version bump, fixes #474628
