Analysis of Specimin vs Perses

[kelloggm]

@M3CHR0M4NC3R, we'll need a similar analysis for each bug that we want to discuss in the paper (this analysis doesn't go into the paper directly, but it needs to appear in the supporting material). Can you do this for NA-323? Or, if you're busy working on the other bugs, can you provide me with the Perses NA-323 output so that I can do this myself?

[M3CHR0M4NC3R]

I don't think I can make a commit to this pr, but this is what I added

NA-323

Specimin Output:

package com.cogvio.time;

import java.time.Clock;
import java.time.Instant;
import java.time.ZoneId;

public final class MutableClock extends Clock {

    private Instant instant;

    private ZoneId zone;

    private MutableClock(final Instant instant, final ZoneId zone) {
        this.setZone(zone);
        this.setInstant(instant);
    }

    public void setInstant(final Instant instant) {
        throw new Error();
    }

    public void setZone(final ZoneId zone) {
        throw new Error();
    }

    public ZoneId getZone() {
        throw new Error();
    }

    public MutableClock withZone(final ZoneId newZone) {
        throw new Error();
    }

    public Instant instant() {
        throw new Error();
    }
}

Perses Output:

package com.cogvio.time;

import java.time.Instant;
import java.time.ZoneId;

class MutableClock {
  Instant instant;
  ZoneId zone;

  MutableClock() {}
}

Line counts using same formatting:

• Specimin: 27
• Perses: 8

Differences:

• Perses doesn't include clock
• Perses removes "public final" before "class" in class declaration and removes "extends"
• Perses removes "private" from variable declarations
• Perses removes arguments for MutableClock()
• Specimin replaces the body of the setters and getters with "throw new Error();" while perses removes them entirely

[kelloggm]

I don't think I can make a commit to this pr

Not directly, but you should be able to open a PR of your own that targets this branch. If you're having trouble with that, though, I can commit this for you.

Either way, this is really helpful - I'll add some text to the paper about it.