[SPARK-50579][SQL] Fix truncatedString

[MaxGekk]

What changes were proposed in this pull request?

In the PR, I propose to respect the maxFields argument of SparkStringUtils.truncatedString in all cases independently from the size of the another argument seq. If size of seq is grater or equal to maxFields, output exact maxFields elements of seq.

Also, if the number of printed elements is zero, don't print the redundant comma , like [, ... 100 more fields], and don't overflow when maxFields is Int.MinValue.

Why are the changes needed?

To make output consistent for the same maxFields. For example, if maxFields is 2:

truncatedString(Seq(1, 2), "[", ", ", "]", maxFields = 2) -> "[1, 2]"

but for more elements in the input sequence, it prints only one element:

truncatedString(Seq(1, 2, 3), "[", ", ", "]", maxFields = 2) -> "[1, ... 2 more fields]"

though the expected output is "[1, 2, ... 1 more fields]".

Does this PR introduce any user-facing change?

No.

How was this patch tested?

By existing tests from UtilSuite that are moved to StringUtilsSuite and additional checks:

$ build/sbt "test:testOnly *StringUtilsSuite"

Was this patch authored or co-authored using generative AI tooling?

No.

[HyukjinKwon]

Merged to master.