If running incrementally, even if there aren't any Kotlin sources to
compile, there could be stale output files that need to be removed. E.g
stub generation must have generated .java files removed if all Kotlin
sources are removed. Kotlin compile task will be skipped only if there
are no Kotlin sources and task run is non-incremental.
Test: KaptIncrementalIT.testRemoveAllKotlinSources
Remove sample reference from indexOf(Char) overload: both overloads are on
the same page, and showing the sample for indexOf(String) under indexOf(Char)
is redundant and misleading.
This commit changes XorWowRandom.kt documentation to include a HTTPS link instead of a HTTP link, for security purposes.
This commit also changes the same documentation to reflect the real cycles from the `xorwow` algorithm:
Simple and very fast (125 million/sec), the elements in its cycle of 2^192 − 2^32 easily pass all the tests in Diehard.
Marsaglia, G. 2003. Xorshift RNGs. J. Statis. Soft. 8, 14, p. 5
At last, this commit explicits which part of the code is the implementation of the mentioned algorithm.
Closes KT-35175
* KT-20357: Add samples for functions related to associate
* KT-20357: Use same fib function for both samples
* KT-20357: Use examples with duplicate keys and simplify examples with primitives
* KT-20357: Use primitive samples for all arrays
* KT-20357: Use String splitting example to better illustrate the use for associate over associateBy with a valueSelector
* To be able to deserialize metadata we need to include ":core:deserialization".
* Use per-fragment string tables.
* Read file names from strings instead of file names list.
Replaced createTemporaryVariable usages with createTmpVariables.
The latter doesn't rely on KotlinType's. Using KotlinType with
JS_IR backend causes runtime excpetions, as some expected descriptor
API is not implemented, because it avoids descriptors as much as
possible.
Transitive dependencies weren't fetched properly in some cases, because of the two problems:
- missing dependency configuration for the root dependency
- using URLResolver for the maven-based (IBiblio) resolvers
Also adding resolver name for easier debugging.