Add a unit test for input completion

incomplete, but will help catch obvious wrong behaviors if we change
anything in the completions.

test/test_completion.py

+"""
+Test the completions methods on an altered input object.
+"""
+
+import string
+import pytest
+import random
+import sys
+import os
+
+sys.path.append('src')
+
+class ConfigShim(object):
+    def get(self, *args, **kwargs):
+        return ''
+
+import config
+config.config = ConfigShim()
+
+from windows import Input
+
+@pytest.fixture(scope="function")
+def input_obj():
+    obj = Input()
+    obj.reset_completion()
+    obj.resize = lambda: None
+    obj.rewrite_text = lambda: None
+    obj.refresh = lambda: None
+    return obj
+
+@pytest.fixture(scope="module")
+def random_unquoted_words():
+    letters = string.ascii_lowercase + ((len(string.ascii_lowercase)//4)*' ')
+    acc = [random.choice(letters) for _ in range(200)]
+    words = ''.join(acc).split()
+    return words
+
+@pytest.fixture(scope="module")
+def quoted_words():
+    words = []
+    letters = string.ascii_lowercase + ((len(string.ascii_lowercase)//4)*' ')
+    words_by_letter = {}
+    for start_letter in string.ascii_lowercase:
+        words_by_letter[start_letter] = []
+        for _ in range(5):
+            size = random.randint(0, 15)
+            word = start_letter + ''.join(random.choice(letters) for i in range(size))
+            words.append(word)
+            words_by_letter[start_letter].append(word)
+    return (words, words_by_letter)
+
+
+def test_new_completion_1_unquoted(input_obj):
+
+    input_obj.text = '/example '
+    input_obj.pos = len(input_obj.text) - 1
+
+    input_obj.new_completion(['toto', 'titi'], 1, quotify=False)
+    assert input_obj.text == '/example toto'
+
+    input_obj.new_completion(['toto', 'titi'], 1, quotify=False)
+    assert input_obj.text == '/example titi'
+
+    input_obj.new_completion(['toto', 'titi'], 1, quotify=False)
+    assert input_obj.text == '/example toto'
+
+
+def test_new_completion_1_quoted_spaces(input_obj):
+    input_obj.text = '/example '
+    input_obj.pos = len(input_obj.text) - 1
+
+    input_obj.new_completion(['toto toto', 'titi titi'], 1, quotify=True)
+    assert input_obj.text == '/example "toto toto"'
+
+    input_obj.new_completion(['toto toto', 'titi titi'], 1, quotify=True)
+    assert input_obj.text == '/example "titi titi"'
+
+    input_obj.new_completion(['toto toto', 'titi titi'], 1, quotify=True)
+    assert input_obj.text == '/example "toto toto"'
+
+    input_obj.text = '/example '
+    input_obj.pos = len(input_obj.text) - 1
+    input_obj.reset_completion()
+
+    input_obj.new_completion(['toto toto', 'tata', 'titi titi'], 1, quotify=True)
+    assert input_obj.text == '/example "toto toto"'
+
+    input_obj.new_completion(['toto toto', 'tata', 'titi titi'], 1, quotify=True)
+    assert input_obj.text == '/example tata'
+
+    input_obj.new_completion(['toto toto', 'tata', 'titi titi'], 1, quotify=True)
+    assert input_obj.text == '/example "titi titi"'
+
+def test_new_completion_unquoted_random_override(input_obj, random_unquoted_words):
+    """
+    Complete completely random words and ensure that the input is
+    changed adequately.
+    """
+    words = random_unquoted_words
+
+    # try the completion on the middle element without affecting the others
+    input_obj.text = '/example %s %s %s' % (words[0], words[1], words[2])
+    base = len(input_obj.text) - len(words[2]) - 1
+    input_obj.pos = base
+    def f(n):
+        return '/example %s %s' % (words[0], words[n])
+
+    for i in range(len(words)):
+        pos = input_obj.get_argument_position(False)
+        input_obj.new_completion(words[:], pos, quotify=False, override=True)
+        assert f(i) + " " + words[2] == input_obj.text
+        assert len(f(i)) == input_obj.pos
+
+    assert input_obj.text == '/example %s %s %s' % (words[0], words[-1], words[2])
+
+    pos = input_obj.get_argument_position(False)
+    input_obj.new_completion(words[:], pos, quotify=False, override=True)
+    assert input_obj.text == '/example %s %s %s' % (words[0], words[0], words[2])
+
+    input_obj.reset_completion()
+
+    # try the completion on the final element without affecting the others
+    input_obj.text = '/example %s %s %s' % (words[0], words[1], words[2])
+    base = len(input_obj.text)
+    input_obj.pos = base
+    def f2(n):
+        return '/example %s %s %s' % (words[0], words[1], words[n])
+
+    print(words)
+    for i in range(len(words)):
+        pos = input_obj.get_argument_position(False)
+        input_obj.new_completion(words[:], pos, quotify=False, override=True)
+        assert f2(i) == input_obj.text
+        assert len(f2(i)) == input_obj.pos
+
+    assert input_obj.text == '/example %s %s %s' % (words[0], words[1], words[-1])
+
+
+def test_new_completion_quoted_random(input_obj, quoted_words):
+    """
+    Complete (possibly) quoted words starting with a specific letter.
+    And make sure that the quotes only appear when necessary.
+    """
+    words = quoted_words[0]
+    words_l = quoted_words[1]
+
+    letters = ('', 'a', 'b', 'c')
+
+    # generate the text which is supposed to be present in the input
+    def f(p, i):
+        rep = words_l[letters[p]][i] if not ' ' in words_l[letters[p]][i] else '"'+words_l[letters[p]][i]+'"'
+        fst = letters[1] if p != 1 else rep
+        snd = letters[2] if p != 2 else rep
+        trd = letters[3] if p != 3 else rep
+        return '/example %s %s %s' % (fst, snd, trd)
+
+    for pos in range(1, 4):
+        input_obj.text = '/example a b c'
+        input_obj.reset_completion()
+        input_obj.pos = len('/example') + pos * 2
+
+        extra = (3 - pos) * 2
+        for i in range(5):
+            input_obj.new_completion(words[:], pos, quotify=True)
+            assert f(pos, i) == input_obj.text
+            assert len(f(pos, i)) - extra == input_obj.pos
+