Issue #14506: False positive for UnusedLocalVariable

[Kevin222004]

Issue #14506: False positive for UnusedLocalVariable

Diff Regression config: https://gist.githubusercontent.com/Kevin222004/0dbdfb487f8e9a050ed0e6356f1a35b4/raw/31f7927f400e11e4285e86fd086b373095227848/unusedLocal.xml

[Kevin222004]

Github, generate report

[github-actions]

https://checkstyle-diff-reports.s3.us-east-2.amazonaws.com/3c60d4e_2024062445/reports/diff/index.html

[rdiachenko]

lgtm

[nrmancuso]

Minor item:

[nrmancuso]

Good to merge, @Kevin222004 please make CI happy

[romani]

Items

[romani]

ok to merge

[rnveach]

I am not sure why LITERAL_SWITCH and why not something that makes more sense like EXPR? A switch is not technically anything used with increments/decrements. At least the other 3 make somewhat more sense.

[Kevin222004]

Whenever the token is of type TokenTypes.ELIST, TokenTypes.INDEX_OP, TokenTypes.ASSIGN, in that cases the Variable is been used for example

int h = ++m;

the last will be

|       |       |--VARIABLE_DEF -> VARIABLE_DEF [66:8]
|       |       |   |--MODIFIERS -> MODIFIERS [66:8]
|       |       |   |--TYPE -> TYPE [66:8]
|       |       |   |   `--LITERAL_INT -> int [66:8]
|       |       |   |--IDENT -> h [66:12]
|       |       |   `--ASSIGN -> = [66:14]
|       |       |       `--EXPR -> EXPR [66:16]
|       |       |           `--INC -> ++ [66:16]
|       |       |               `--IDENT -> m [66:18]

this is the one case where expr parent is TokenTypes.ASSIGN

know particular in switch

        switch (i++) {

the ast will be

        |       |--LITERAL_SWITCH -> switch [12:8]
        |       |   |--LPAREN -> ( [12:15]
        |       |   |--EXPR -> EXPR [12:17]
        |       |   |   `--POST_INC -> ++ [12:17]
        |       |   |       `--IDENT -> i [12:16]
        |       |   |--RPAREN -> ) [12:19]

the way to declare the i++ as used I have to put their TokenTypes.LiteralSwitch

[rnveach]

and why not something that makes more sense like EXPR?

I am not seeing this specifically answered.

[Kevin222004]

Specifically in this function we are trying to find out whether the increment or decrement variable is used or not and in case of TokenTypes.ELIST, TokenTypes.INDEX_OP, TokenTypes.ASSIGN if this tokens are the parent of expr then it is used and they are understandable for example again TokenTypes.ASSIGN

int h = ++m;

this case of LiteralSwitch is slightly unique in general as per the check logic if the increment decrement variable is alone then the variable seems to be unused but in this case, it is not true that's why the TokenTypes.LiteralSwitch is used in the sense if it is parent then the variable is used as I have explained in above comment.
and I also don't see any way some how only EXPR can be used and we can declare this as used variable.

[rnveach]

I am not fully understanding and this isn't making sense.

Either way, the javadoc needs to be updated for this method for the new token being examined.

[rnveach]

@Kevin222004 ping

[Kevin222004]

@rnveach sorry for the delay. I will update the pr as soon as possible.

[romani]

@Kevin222004, ping, please find time to fix this PR. We are almost done. Passing of last reviewer is left only.

[Kevin222004]

I am sorry for such long delay will fix with in 2 days

[Kevin222004]

Done