Some testcases fail on py3

[robertwb]

I am fairly certain that these are only problems in the test files and not with Cython, but it should be fixed to be sure.

Update:

• extsetslice -- should probably be disabled in Python 3, as it tests py2-only functionality
• extdelslice -- ditto
• cdivision_CEP_516 -- There's a problem with how warnings etc are reported, + sys.maxint isn't available in Py3

Otherwise it looks fine.

Migrated from http://trac.cython.org/ticket/319

[robertwb]

@dagss changed description from

I am fairly certain that these are only problems in the test files and not with Cython, but it should be fixed to be sure.

to

I am fairly certain that these are only problems in the test files and not with Cython, but it should be fixed to be sure.

Update:

• extsetslice -- should probably be disabled in Python 3, as it tests py2-only functionality
• extdelslice -- ditto
• cdivision_CEP_516 -- There's a problem with how warnings etc are reported, + sys.maxint isn't available in Py3

Otherwise it looks fine.
milestone from 0.11.2 to 0.11.3
commented

[robertwb]

scoder commented

also see http://trac.cython.org/ticket/215 regarding the slice functions

[robertwb]

scoder changed milestone from 0.11.3 to 0.12
commented

I disabled the slicing tests in Py3. All three tests no longer fail.

Remaining test failures on cython-unstable:

• complex_numbers_T305 - doctests related to complex numbers fail due to changed complex repr() in Py3.1
• cpp_exceptions_T265 - exception value is set up as string, whereas the later compiler code expects an ExprNode or PyrexType

[robertwb]

scoder commented

Update: the only remaining tests that fail are the "complex" tests, i.e. complex_numbers_*, because repr(complex) changed in Py3.1 and now (correctly) returns -(2j) as (-0-2j)

[robertwb]

@robertwb changed milestone from 0.12 to 0.12.1
commented

[robertwb]

@robertwb changed milestone from 0.12.1 to 0.13
commented

[robertwb]

scoder changed resolution to fixed
status from new to closed
commented

Test cases work in current Py3, so let's close this ancient ticket.