<sectionName>
- p-<paragraphNumber>
  - <neg|pos>
    - <setenceNumber>.<testNumber>.kt (test source code)
    - <setenceNumber>.<testNumber>.txt (descriptors file)

Example test file path: testData/diagnostics/linked/when-expression/p-2/neg/3.1.kt

Positive and negative tests

Positive tests are considered to be tests in which there is no single diagnostics with a ERROR severity. Positive tests can only contain diagnostics with a WARNING or INFO severity (or not contain at all).

In a negative test, there must be at least one diagnostic with ERROR severity.

Tests format

Test description

Each test file must contain meta information in the form of a multi-line comment.

A comment with meta information has the following format:

/*
 KOTLIN SPEC TEST (<POSITIVE|NEGATIVE>)

 SECTIONS: <sectionNames{,}>
 PARAGRAPH: <paragraphNumber>
 SENTENCE: [<setenceNumber>] <setence>
 NUMBER: <testNumber>
 DESCRIPTION: <testDescription>
 */

Example:

/*
 KOTLIN DIAGNOSTICS SPEC TEST (POSITIVE)

 SECTIONS: when-expression
 PARAGRAPH: 2
 SENTENCE: [3] When expression has two different forms: with bound value and without it.
 NUMBER: 1
 DESCRIPTION: Empty 'when' with bound value.
 */

Meta-information should be placed at the beginning of the file after directives (if any).

Case description

The test can contain many cases. Each case (if there are more than one) should be accompanied by a description.

The case description is a single-line comment of the following format:

// CASE DESCRIPTION: <caseDescription>

Example:

// CASE DESCRIPTION: Checking for exhaustive 'when' (all sealed class subtypes and null value are covered).
fun case_1(expr: Expr?): String = when (expr) {
    is Const -> <!DEBUG_INFO_SMARTCAST!>expr<!>.n
    is Sum -> <!DEBUG_INFO_SMARTCAST!>expr<!>.e1 + <!DEBUG_INFO_SMARTCAST!>expr<!>.e2
    is Mul -> <!DEBUG_INFO_SMARTCAST!>expr<!>.m1 + <!DEBUG_INFO_SMARTCAST!>expr<!>.m2
    null -> ""
}

Test validation

Before running the test, the following validation is performed:

check for correspondence to the format of the folders and file names;
check for correspondence to the format of the meta-information in the test file;
check for consistency between the location and name of the test file and the meta-information in it;
checking whether the test is positive or negative using information about diagnostics severity.

If the validation fails, you will receive exception about it.

Test run log

During the test run, the following information is displayed for each test:

DIAGNOSTICS <POSITIVE|NEGATIVE> SPEC TEST
SECTIONS: <sectionName> (paragraph: <paragraphNumber>)
SENTENCE <sentenceNumber>: <sentence>
TEST NUMBER: <testNumber>
TEST CASES: <casesNumber>
DESCRIPTION: <testDescription>
DIAGNOSTICS: <diagnosticSeverities> | <diagnostics>

Example:

DIAGNOSTICS NEGATIVE SPEC TEST
SECTIONS: when-expression (paragraph: 3)
SENTENCE 1: 
TEST NUMBER: 1
NUMBER OF TEST CASES: 3
DESCRIPTION: 'When' without bound value and not allowed break and continue expression (without labels) in the control structure body.
DIAGNOSTICS: {ERROR=2, WARNING=1} | {BREAK_OR_CONTINUE_IN_WHEN=2, UNREACHABLE_CODE=1}

Statistics on specification tests

To see statistics for existing tests you can run gradle task printSpecTestStatistic in the :compiler:tests-spec.

Example output:


==================================================
SPEC TESTS STATISTIC
--------------------------------------------------
PSI: 0 tests
DIAGNOSTICS: 54 tests
  when-expression: 54 tests
    PARAGRAPH 9: 8 tests [ neg: 4 ] [ pos: 4 ]
    PARAGRAPH 7: 16 tests [ neg: 7 ] [ pos: 9 ]
    PARAGRAPH 6: 2 tests [ neg: 1 ] [ pos: 1 ]
    PARAGRAPH 11: 17 tests [ neg: 8 ] [ pos: 9 ]
    PARAGRAPH 3: 7 tests [ neg: 3 ] [ pos: 4 ]
    PARAGRAPH 2: 2 tests [ pos: 2 ]
    PARAGRAPH 5: 2 tests [ neg: 1 ] [ pos: 1 ]
CODEGEN: 0 tests