class StubParameterParserTest extends TestCase (View source)

Methods

void
setUp()

No description

testItReturnsCorrectInstance()

No description

testItCanParseParameterName()

No description

testItParsesTheByReferenceFlagPerParameter()

&$param used to be dropped on the stub side entirely: parseNode() never called setIsPassedByReference() and ParameterNode had no accessor for nikic's Param::$byRef, so all 10,697 cached stub parameters read false while reflection reported true for 169 of them.

testItAssignsTheSignatureIndexAsThePosition()

setPosition() was never called, so every stub parameter kept PHPParameter's constructor default of 0 while the reflection side recorded real indices. Only the second parameter onwards can tell a real position from that default, hence the three-parameter fixture.

Details

protected void setUp()

No description

Return Value

void

testItReturnsCorrectInstance()

No description

testItCanParseParameterName()

No description

testItCanParseMultipleParameters()

No description

testItReturnsEmptyArrayForFunctionWithNoParameters()

No description

testItParsesParametersFromFunctionWithSingleParameter()

No description

testItParsesTheByReferenceFlagPerParameter()

&$param used to be dropped on the stub side entirely: parseNode() never called setIsPassedByReference() and ParameterNode had no accessor for nikic's Param::$byRef, so all 10,697 cached stub parameters read false while reflection reported true for 169 of them.

Asserting per-parameter (not just "at least one is true") is what distinguishes reading the flag from hardcoding it.

testItParsesByReferenceOnAVariadicParameter()

No description

testItAssignsTheSignatureIndexAsThePosition()

setPosition() was never called, so every stub parameter kept PHPParameter's constructor default of 0 while the reflection side recorded real indices. Only the second parameter onwards can tell a real position from that default, hence the three-parameter fixture.