Initial Halstead implementation for Java

[dburriss]

• Implement necessary functions
• Passing test
• Cleanup

Some references:

• Java lang ref
• https://www.geeksforgeeks.org/software-engineering-halsteads-software-metrics/?msclkid=5e181114abef11ecbb03527e95a34828

[dburriss]

Test exmaple:

public static void main(String args[]) {
  int a, b, c, avg;
  a = 5; b = 5; c = 5;
  avg = (a + b + c) / 3;
  MessageFormat.format(\"{0}\", avg);
}

I will get to this over the weekend and try to set my test up for expected numbers.
@marco-c who can explain to me the difference between u_operators, operators, u_operands, and operands in the Halstead tests?

[Luni-4]

Test exmaple:

public static void main(String args[]) {
  int a, b, c, avg;
  a = 5; b = 5; c = 5;
  avg = (a + b + c) / 3;
  MessageFormat.format(\"{0}\", avg);
}

I will get to this over the weekend and try to set my test up for expected numbers. @marco-c who can explain to me the difference between u_operators, operators, u_operands, and operands in the Halstead tests?

@dburriss
u_operators are all distinct operators present in the code
u_operands are all unique operands present in the code
operators are all occurrences of u_operators in the code
operands are all occurrences of u_operands in the code.

You can find what are operators and operands in these tests https://github.com/mozilla/rust-code-analysis/blob/master/src/ops.rs#L312 for different programming languages. First vec are operators and second one operands.

[dburriss]

@Luni-4 I have run into a weird issue.

When parsing the first line public static void main(String args[]){

unknown: program 
unknown: local_variable_declaration 
unknown: modifiers 
unknown: public
unknown: static
operator: void_type
unknown: variable_declarator
operand: identifier
operator: ;
unknown: ERROR
unknown: formal_parameters
operator: (
unknown: formal_parameter
operator: type_identifier
operand: identifier
unknown: dimensions
operator: [
operator: ]
operator: )
unknown: block

Note the operator: ; and unknown: ERROR ... probably related. There is no ; in the first line so this is throwing the count off.

The example is valid Java... I compiled it.

[calixteman]

It's very likely a matter of adding public class A {and } around your function.

[dburriss]

It's very likely a matter of adding public class A {and } around your function.

Thanks @calixteman
That did the trick. It actually crossed my mind but I had never seen ERROR in the LoC parsing and most of those tests don't have class. Anyway, I learned something. And I am getting the numbers I expect.

@Luni-4 I included a lot of comments and links, etc. in the code explaining what and why I am including different things. I didn't find definitive rules about what is included or excluded.

[codecov-commenter]

Codecov Report

Merging #811 (fa10e44) into master (b5496f8) will decrease coverage by 1.24%.
The diff coverage is 92.30%.

@@            Coverage Diff             @@
##           master     #811      +/-   ##
==========================================
- Coverage   41.24%   40.00%   -1.25%     
==========================================
  Files          52       52              
  Lines        7201     7490     +289     
  Branches     1039     1041       +2     
==========================================
+ Hits         2970     2996      +26     
- Misses       3519     3782     +263     
  Partials      712      712              

Impacted Files	Coverage Δ
src/metrics/halstead.rs	72.77% <83.33%> (-0.22%)	⬇️
src/getter.rs	79.83% <100.00%> (+4.32%)	⬆️
src/languages/language_java.rs	3.14% <0.00%> (-86.86%)	⬇️
src/spaces.rs	72.61% <0.00%> (-1.79%)	⬇️
src/metrics/nargs.rs	75.09% <0.00%> (+0.38%)	⬆️
src/node.rs	86.53% <0.00%> (+15.38%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b5496f8...fa10e44. Read the comment docs.

[dburriss]

@Luni-4

I looked at ops.rs and added a test but having some problems.

Copying the existing tests quickly and filling in from the list above (or the code), I get the following error:

thread 'ops::tests::java_ops' panicked at 'assertion failed: `(left == right)`
  left: `["()", ")", "+", ",", ".", "/", ";", "=", "[]", "]", "integral_type", "method_invocation", "type_identifier", "void_type", "{}", "}"]`,
 right: `["()", "+", ",", "/", ";", "=", "String", "[]", "format", "int", "void", "{}"]`', src\ops.rs:280:9

A few questions questions:

1. Do you know why the list for Java is showing the grammar type instead of the value ie. 'integral_typeinstead ofint` as it does for the other languages
2. Does the order of operators/operands matter in the assert?
3. Are these println supposed to be here? (hard to believe but they are not mine :) ) https://github.com/mozilla/rust-code-analysis/blob/master/src/ops.rs#L92

[dburriss]

@Luni-4 this is failing on the ops.rs file you mentioned. Not something I introduced. The linter has an opinion on something. I can make the change in this PR but I'm not sure how this got in in the first place. New rule maybe?

Log: https://community-tc.services.mozilla.com/tasks/WqxZrlJ2TIKZMbpDibX5Jg/runs/0/logs/public/logs/live_backing.log

[dburriss]

Shuffled things around a bit and moved ops check to #849
Hopefully, it is ok to merge those 2 separately.
@calixteman @Luni-4 ok to review with this scope?

[Luni-4]

Sorry @dburriss for the delay, yeah, it's fine splitting up the two problems.

Just a final request and we can land this PR

[Luni-4]

Thank you @dburriss! Fine for me!