streambuffer/spotbugs-exclude.xml at main · bernardladenthin/streambuffer · GitHub

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
<?xml version="1.0" encoding="UTF-8"?>
<!--
SPDX-FileCopyrightText: 2014-2026 Bernard Ladenthin <bernard.ladenthin@gmail.com>

SPDX-License-Identifier: Apache-2.0
-->
<FindBugsFilter
    xmlns="https://github.com/spotbugs/filter/3.0.0"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="https://github.com/spotbugs/filter/3.0.0 https://raw.githubusercontent.com/spotbugs/spotbugs/4.8.6/spotbugs/etc/findbugsfilter.xsd">

    <!--
        getOutputStream() / getInputStream() intentionally expose the
        SBOutputStream / SBInputStream fields - those streams ARE the
        public API of StreamBuffer.
    -->
    <Match>
        <Class name="net.ladenthin.streambuffer.StreamBuffer"/>
        <Or>
            <Method name="getOutputStream"/>
            <Method name="getInputStream"/>
        </Or>
        <Bug pattern="EI_EXPOSE_REP"/>
    </Match>

    <!--
        bufferLock is private final - its reference cannot be replaced.
        The synthetic access$N bridge is a Java 8 compiler artifact: the
        private inner classes SBInputStream and SBOutputStream access the
        private field directly, and javac emits a package-private static
        accessor to make that legal at bytecode level. No other class in
        net.ladenthin.streambuffer exists to call that accessor.
        The jcstress ConcurrentWriteRace tests confirm the synchronization
        protocol is sound; this is a false positive.
    -->
    <Match>
        <Class name="net.ladenthin.streambuffer.StreamBuffer"/>
        <Method name="toString"/>
        <Bug pattern="USO_UNSAFE_ACCESSIBLE_OBJECT_SYNCHRONIZATION"/>
    </Match>

    <!--
        trim() drains the buffer into a temporary ArrayDeque before
        consolidating it back. SpotBugs PSC_PRESIZE_COLLECTIONS suggests
        passing an initial capacity hint, but a tight bound here would
        require ceil(availableBytes / maxAllocationSize) over two volatile
        long fields. That arithmetic produces equivalent mutants (ArrayDeque
        tolerates any non-negative initial capacity, so the capacity hint
        is unobservable through tests) and would break the package's 100 %
        PIT mutation threshold. trim() is also not a hot path - it runs at
        most once per maxBufferElements writes (default 100), so the
        ArrayDeque default growth 16-32-64 is negligible against the
        surrounding I/O.
    -->
    <Match>
        <Class name="net.ladenthin.streambuffer.StreamBuffer"/>
        <Method name="trim"/>
        <Bug pattern="PSC_PRESIZE_COLLECTIONS"/>
    </Match>
</FindBugsFilter>