Misleading error message when cloning fails to create a new instance

[marcotama]

Describe the bug
When the FieldAccessingSolutionCloner fails to construct an object, it throws an exception with message "The class $CLAZZ should have a no-arg constructor to create a planning clone.". However, construction can fail for other reasons, e.g. in my case, I initialize some fields based on other fields, and this failed when a certain collection was empty.

Expected behavior
A better error message, e.g. "Failed to create a new instance of class $CLAZZ for the purpose of solution-cloning. This could be due to a missing no-arg constructor or some other error inside the logic of class $CLAZZ triggered by the no-arg constructor."

Actual behavior
Describe clearly and concisely what actually happened.

To Reproduce
N/A

Environment

Timefold Solver Version or Git ref: 1.1.0

Output of java -version: N/A

Output of uname -a or ver: N/A

Additional information

N/A

[triceo]

Thanks for reporting, @marcotama!
Would you have an example source code of a class that failed to clone in this way?

[marcotama]

Hello!

Something like this (Kotlin).
When you create a new TimeBlock using the default constructor, the list of Timeslots is empty and so next() throws a NoSuchElementException.

import java.time.DayOfWeek

class Timeslot(
    @PlanningId
    val id: Int? = null,
    val name: String? = null,
    val dayOfWeek: DayOfWeek? = null,
    val startTime: LocalTime? = null,
    val endTime: LocalTime? = null,
)

class TimeBlock(
    @PlanningId
    val id: Int? = null,
    val name: String? = null,
    val allTimeslots: Set<Timeslot> = mutableSetOf(),
) {
    val day: DayOfWeek? = allTimeslots.map { it.dayOfWeek }.iterator().next()
}

Of course there are problems with my classes, but I think the error message could be improved - I ended up spending 10 minutes figuring out why my no-arg constructor was not getting picked up.

[triceo]

Thank you for the example. You are right, the message could be improved.