Drop kotlinx.serialization compiler plugin as it's unnecessary #3899
Conversation
| fun fromJsonObject(obj: JsonObject): DokkaModuleDescriptionKxs = DokkaModuleDescriptionKxs( | ||
| name = obj["name"]!!.jsonPrimitive.content, | ||
| modulePath = obj["modulePath"]!!.jsonPrimitive.content, | ||
| moduleOutputDirName = obj["moduleOutputDirName"]!!.jsonPrimitive.content, | ||
| moduleIncludesDirName = obj["moduleIncludesDirName"]!!.jsonPrimitive.content, | ||
| ) |
There was a problem hiding this comment.
This needs to be updated, because properties with default constructor arguments are optional.
What do you think about adding some tests? To ensure any code changes don't break things in the future.
adam-enko
added
the
runner: Gradle plugin
whyoleg
force-pushed
the
drop-kx.serialization-plugin
branch
from
1b3e65e to
168c743
Compare
whyoleg
force-pushed
the
drop-kx.serialization-plugin
branch
from
168c743 to
b449849
Compare
| fun toJsonObject(module: DokkaModuleDescriptionKxs): JsonObject = buildJsonObject { | ||
| put("name", module.name) | ||
| put("modulePath", module.modulePath) | ||
| put("moduleOutputDirName", module.moduleOutputDirName) | ||
| put("moduleIncludesDirName", module.moduleIncludesDirName) | ||
| } | ||
|
|
||
| fun fromJsonObject(obj: JsonObject): DokkaModuleDescriptionKxs = DokkaModuleDescriptionKxs( | ||
| name = obj.getString("name"), | ||
| modulePath = obj.getString("modulePath"), | ||
| moduleOutputDirName = obj.getString("moduleOutputDirName", defaultModuleOutputDirName), | ||
| moduleIncludesDirName = obj.getString("moduleIncludesDirName", defaultModuleIncludesDirName), | ||
| ) |
There was a problem hiding this comment.
Instead of these methods, what do you think about exposing a hand written KSerializer<DokkaModuleDescriptionKxs>? It'd be more consistent with the current behaviour, reduce our API surface, and not tie us to a specific format.
There was a problem hiding this comment.
I don't really think it's needed here as the implementation of custom serializer will be much more complex than the implementation of those two functions.
I don't see how it will reduce our API surface, as it's internal (overall the whole class should be internal I believe).
The same with specific format - we don't need it now, and as it's an implementation detail we can implement it later if needed.
Extracted from #3898, required to be able to update to KGP 2.1.0