##// END OF EJS Templates
Fix #4777 and #7887...
Fix #4777 and #7887 The function in charge of actually converting cursor offset to CodeMirror line number and character number was actually crashing when the cursor was at the last character (loop until undefined, then access length of variable, which is undefined). This was hiding a bug in which when you would completer to a single completion pressing tab after as-you-type filtering, the completion would be completed twice. The logic that was supposed to detect whether or not all completions had a common prefix was actually faulty as the common prefix used to be a string but was then changed to an object. Hence the logic to check whether or not there was actually a common prefix was always true, even for empty string, leading to the deletion of the line (replace by '') in some cases.

File last commit:

r12600:cfd92dce
r20538:ae7f6d6a
Show More
test_handlers.py
97 lines | 3.2 KiB | text/x-python | PythonLexer
"""Tests for input handlers.
"""
#-----------------------------------------------------------------------------
# Module imports
#-----------------------------------------------------------------------------
# third party
import nose.tools as nt
# our own packages
from IPython.core import autocall
from IPython.testing import tools as tt
from IPython.testing.globalipapp import get_ipython
from IPython.utils import py3compat
#-----------------------------------------------------------------------------
# Globals
#-----------------------------------------------------------------------------
# Get the public instance of IPython
ip = get_ipython()
failures = []
num_tests = 0
#-----------------------------------------------------------------------------
# Test functions
#-----------------------------------------------------------------------------
class CallableIndexable(object):
def __getitem__(self, idx): return True
def __call__(self, *args, **kws): return True
class Autocallable(autocall.IPyAutocall):
def __call__(self):
return "called"
def run(tests):
"""Loop through a list of (pre, post) inputs, where pre is the string
handed to ipython, and post is how that string looks after it's been
transformed (i.e. ipython's notion of _i)"""
tt.check_pairs(ip.prefilter_manager.prefilter_lines, tests)
def test_handlers():
call_idx = CallableIndexable()
ip.user_ns['call_idx'] = call_idx
# For many of the below, we're also checking that leading whitespace
# turns off the esc char, which it should unless there is a continuation
# line.
run([(i,py3compat.u_format(o)) for i,o in \
[('"no change"', '"no change"'), # normal
(u"lsmagic", "get_ipython().magic({u}'lsmagic ')"), # magic
#("a = b # PYTHON-MODE", '_i'), # emacs -- avoids _in cache
]])
# Objects which are instances of IPyAutocall are *always* autocalled
autocallable = Autocallable()
ip.user_ns['autocallable'] = autocallable
# auto
ip.magic('autocall 0')
# Only explicit escapes or instances of IPyAutocallable should get
# expanded
run([
('len "abc"', 'len "abc"'),
('autocallable', 'autocallable()'),
# Don't add extra brackets (gh-1117)
('autocallable()', 'autocallable()'),
])
ip.magic('autocall 1')
run([
('len "abc"', 'len("abc")'),
('len "abc";', 'len("abc");'), # ; is special -- moves out of parens
# Autocall is turned off if first arg is [] and the object
# is both callable and indexable. Like so:
('len [1,2]', 'len([1,2])'), # len doesn't support __getitem__...
('call_idx [1]', 'call_idx [1]'), # call_idx *does*..
('call_idx 1', 'call_idx(1)'),
('len', 'len'), # only at 2 does it auto-call on single args
])
ip.magic('autocall 2')
run([
('len "abc"', 'len("abc")'),
('len "abc";', 'len("abc");'),
('len [1,2]', 'len([1,2])'),
('call_idx [1]', 'call_idx [1]'),
('call_idx 1', 'call_idx(1)'),
# This is what's different:
('len', 'len()'), # only at 2 does it auto-call on single args
])
ip.magic('autocall 1')
nt.assert_equal(failures, [])