[Gradle] Explain reason for (isCanBeResolved && isCanBeConsumed) check

This commit is contained in:
Anton Lakotka
2023-11-26 18:47:23 +01:00
parent 7673434ccd
commit bb5e219c1d
@@ -22,6 +22,16 @@ internal fun ConfigurationContainer.createResolvable(name: String): Configuratio
}
internal fun ConfigurationContainer.findResolvable(name: String): Configuration? = findByName(name)?.apply {
// In earlier gradle versions (6.4 for example) some default configurations such as "compile" and "runtime"
// are both resolvable and consumable by default, here they get fixed to one specific role
// this prevents from incorrect usage e.g. two consecutive calls findResolvable("foo") findConsumable("foo")
// would fail and indicate the incorrect usage.
// Another reason is that before this role-based Configurations API code in KGP sets manually the roles
// i.e. `findByName("someDefaultConfiguration").isCanBeConsumed = false` now it is just `findResolvable()`
// In earlier versions of gradle this did make sense i.e. "compileClasspath" configuration was both resolvable and consumable
// and code above fixes configuration role. However with recent gradle versions these default configurations already come in
// correct state. But gradle 8.2 reports a warning when you try to "re-set" already set flags on configuration
// This is why setting a flag is required only when (isCanBeResolved && isCanBeConsumed) are both true.
if (isCanBeResolved && isCanBeConsumed) {
isCanBeConsumed = false
} else {