1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
|
# Licensed under the Apache License: http://www.apache.org/licenses/LICENSE-2.0
# For details: https://github.com/nedbat/coveragepy/blob/master/NOTICE.txt
"""Tests for coverage.py's results analysis."""
import math
import pytest
from coverage.exceptions import ConfigError
from coverage.results import format_lines, Numbers, should_fail_under
from tests.coveragetest import CoverageTest
class NumbersTest(CoverageTest):
"""Tests for coverage.py's numeric measurement summaries."""
run_in_temp_dir = False
def test_basic(self):
n1 = Numbers(n_files=1, n_statements=200, n_missing=20)
assert n1.n_statements == 200
assert n1.n_executed == 180
assert n1.n_missing == 20
assert n1.pc_covered == 90
def test_addition(self):
n1 = Numbers(n_files=1, n_statements=200, n_missing=20)
n2 = Numbers(n_files=1, n_statements=10, n_missing=8)
n3 = n1 + n2
assert n3.n_files == 2
assert n3.n_statements == 210
assert n3.n_executed == 182
assert n3.n_missing == 28
assert math.isclose(n3.pc_covered, 86.666666666)
def test_sum(self):
n1 = Numbers(n_files=1, n_statements=200, n_missing=20)
n2 = Numbers(n_files=1, n_statements=10, n_missing=8)
n3 = sum([n1, n2])
assert n3.n_files == 2
assert n3.n_statements == 210
assert n3.n_executed == 182
assert n3.n_missing == 28
assert math.isclose(n3.pc_covered, 86.666666666)
@pytest.mark.parametrize("kwargs, res", [
(dict(n_files=1, n_statements=1000, n_missing=0), "100"),
(dict(n_files=1, n_statements=1000, n_missing=1), "99"),
(dict(n_files=1, n_statements=1000, n_missing=999), "1"),
(dict(n_files=1, n_statements=1000, n_missing=1000), "0"),
(dict(precision=1, n_files=1, n_statements=10000, n_missing=0), "100.0"),
(dict(precision=1, n_files=1, n_statements=10000, n_missing=1), "99.9"),
(dict(precision=1, n_files=1, n_statements=10000, n_missing=9999), "0.1"),
(dict(precision=1, n_files=1, n_statements=10000, n_missing=10000), "0.0"),
])
def test_pc_covered_str(self, kwargs, res):
assert Numbers(**kwargs).pc_covered_str == res
@pytest.mark.parametrize("prec, pc, res", [
(0, 47.87, "48"),
(1, 47.87, "47.9"),
(0, 99.995, "99"),
(2, 99.99995, "99.99"),
])
def test_display_covered(self, prec, pc, res):
assert Numbers(precision=prec).display_covered(pc) == res
@pytest.mark.parametrize("prec, width", [
(0, 3), # 100
(1, 5), # 100.0
(4, 8), # 100.0000
])
def test_pc_str_width(self, prec, width):
assert Numbers(precision=prec).pc_str_width() == width
def test_covered_ratio(self):
n = Numbers(n_files=1, n_statements=200, n_missing=47)
assert n.ratio_covered == (153, 200)
n = Numbers(
n_files=1, n_statements=200, n_missing=47,
n_branches=10, n_missing_branches=3, n_partial_branches=1000,
)
assert n.ratio_covered == (160, 210)
@pytest.mark.parametrize("total, fail_under, precision, result", [
# fail_under==0 means anything is fine!
(0, 0, 0, False),
(0.001, 0, 0, False),
# very small fail_under is possible to fail.
(0.001, 0.01, 0, True),
# Rounding should work properly.
(42.1, 42, 0, False),
(42.1, 43, 0, True),
(42.857, 42, 0, False),
(42.857, 43, 0, False),
(42.857, 44, 0, True),
(42.857, 42.856, 3, False),
(42.857, 42.858, 3, True),
# If you don't specify precision, your fail-under is rounded.
(42.857, 42.856, 0, False),
# Values near 100 should only be treated as 100 if they are 100.
(99.8, 100, 0, True),
(100.0, 100, 0, False),
(99.8, 99.7, 1, False),
(99.88, 99.90, 2, True),
(99.999, 100, 1, True),
(99.999, 100, 2, True),
(99.999, 100, 3, True),
])
def test_should_fail_under(total, fail_under, precision, result):
assert should_fail_under(float(total), float(fail_under), precision) == result
def test_should_fail_under_invalid_value():
with pytest.raises(ConfigError, match=r"fail_under=101"):
should_fail_under(100.0, 101, 0)
@pytest.mark.parametrize("statements, lines, result", [
({1,2,3,4,5,10,11,12,13,14}, {1,2,5,10,11,13,14}, "1-2, 5-11, 13-14"),
([1,2,3,4,5,10,11,12,13,14,98,99], [1,2,5,10,11,13,14,99], "1-2, 5-11, 13-14, 99"),
([1,2,3,4,98,99,100,101,102,103,104], [1,2,99,102,103,104], "1-2, 99, 102-104"),
([17], [17], "17"),
([90,91,92,93,94,95], [90,91,92,93,94,95], "90-95"),
([1, 2, 3, 4, 5], [], ""),
([1, 2, 3, 4, 5], [4], "4"),
])
def test_format_lines(statements, lines, result):
assert format_lines(statements, lines) == result
@pytest.mark.parametrize("statements, lines, arcs, result", [
(
{1,2,3,4,5,10,11,12,13,14},
{1,2,5,10,11,13,14},
(),
"1-2, 5-11, 13-14"
),
(
[1,2,3,4,5,10,11,12,13,14,98,99],
[1,2,5,10,11,13,14,99],
[(3, [4]), (5, [10, 11]), (98, [100, -1])],
"1-2, 3->4, 5-11, 13-14, 98->100, 98->exit, 99"
),
(
[1,2,3,4,98,99,100,101,102,103,104],
[1,2,99,102,103,104],
[(3, [4]), (104, [-1])],
"1-2, 3->4, 99, 102-104"
),
])
def test_format_lines_with_arcs(statements, lines, arcs, result):
assert format_lines(statements, lines, arcs) == result
|