cpd algorithm: what is a duplicate? - please vote

[guwirth]

Hi everyone, with #941 the CPD sensor is using a new algorithm:

• ignoreLiterals: CPD ignores literal value differences when evaluating a duplicate block. This means that foo=42; and foo=43; will be seen as equivalent.
• ignoreIdentifiers: Similar to ignoreLiterals but for identifiers; i.e., variable names, methods names, and so forth
• like http://pmd.sourceforge.net/pmd-4.3.0/cpd.html

This approach detects much more duplicates - maybe too much? Like to ask you to vote, What's a duplicate in your opinion:

Example 1:

bits unixtime1(bits ld1, bits ex1)
{
    ld1 &= 0xFF;
    ld1 -= 51;
    if (ex1 < 1855547904U) ld1--;
    ex1 -= 1855548004U;
    return ex1 / 100 + 42949673U * ld1 - ld1 / 25;
}

bits unixtime2(bits ld2, bits ex2)
{
    ld2 &= 0xFF;
    ld2 -= 51;
    if (ex2 < 1855547904U) ld2--;
    ex2 -= 1855548004U;
    return ex2 / 100 + 42949673U * ld2 - ld2 / 25;
}

Example 2:

int object_exists1(char *fn)
{
    int ob;
    if (xosfile_read_stamped_no_path(fn, &ob, 0, 0, 0, 0, 0)) return 0;
    switch (ob)
    {
    case osfile_IS_FILE:return 1;
    case osfile_IS_DIR:return 1;
    case osfile_IS_IMAGE:return 1;
    }
    return 0;
}

int object_exists2(char *fn)
{
    int ob;
    if (xosfile_read_stamped_no_path(fn, &ob, 1, 1, 1, 1, 1)) return 1;
    switch (ob)
    {
    case osfile_IS_FILE:return 2;
    case osfile_IS_DIR:return 2;
    case osfile_IS_IMAGE:return 2;
    }
    return 0;
}

Example 3:

char* tostring1(int value)
{
    switch (value)
    {
    case 0: return "zero";
    case 1: return "one";
    }
    return "undefined";
}

char* tostring2(int value)
{
    switch (value)
    {
    case 2: return "two";
    case 3: return "three";
    }
    return "undefined";
}

Example 4:

int toint1(char value)
{
    switch (value)
    {
    case '0:' return 0;
    case '1': return 1;
    }
    return -1;
}

int toint2(char value)
{
    switch (value)
    {
    case '2:' return 2;
    case '3': return 3;
    }
    return -1;
}

Example 5:

typedef enum ABC {
    ABC           = 12,
    DEF           = 13,
} ABC_e;

typedef enum DEF {
    ABC           = 14,
    DEF           = 15,
} DEF_e;

Example 6:

typedef enum ABC1 {
    ABC1           = 12,
    DEF1           = 13,
} ABC1_e;

typedef enum ABC2 {
    ABC2           = 12,
    DEF2           = 13,
} ABC2_e;

Please vote

[jmecosta]

@guwirth why not provide flags to enable/disable all those options as PMD has? i can see some benift of having those in some projects.

[guwirth]

@jmecosta it's no problem to add some additional configuration settings:

sonar.cxx.cpd.ignoreLiterals=true/false
sonar.cxx.cpd.ignoreIdentifiers=true/false

Nevertheless would be interesting to know your opinion about the cases above to set the defaults. In my opinion the default should be:

sonar.cxx.cpd.ignoreLiterals=false
sonar.cxx.cpd.ignoreIdentifiers=true

[jmecosta]

@guwirth i can ask here to a few developers what they think. and let you know

[jmecosta]

So think 1, 6 should be enabled by default, and perhaps 2. So i supose in our case we would use
sonar.cxx.cpd.ignoreLiterals=true in order to ignore example 5?