[SPARK-58443][SQL] Fix wrong results in instr and substring_index for an Int.MinValue position or count

[jiwen624]

What changes were proposed in this pull request?

Four backward-search paths negate a user-supplied int and use the result as a number of characters or delimiters to skip. -Integer.MIN_VALUE overflows back to Integer.MIN_VALUE, so the value stays negative: the search loop never runs and the out-of-range check that should have fired is bypassed.

Each site now returns the not-found result up front. A string of n bytes holds at most n characters and at most n non-overlapping delimiters, so any position or count with a larger magnitude is out of range regardless of overflow. The guard is therefore a range check rather than a special case for a single value, and it also short-circuits the backward scan for every other out-of-range negative input.

Site	Function	Collations	Behavior before
UTF8String.indexOf(pattern, start, occurrence)	instr	UTF8_BINARY, UTF8_LCASE (all-ASCII fast path)	position of the last match instead of 0
UTF8String.subStringIndex	substring_index	UTF8_BINARY	NegativeArraySizeException
CollationAwareUTF8String.lowercaseSubStringIndex	substring_index	UTF8_LCASE	empty string
CollationAwareUTF8String.subStringIndex	substring_index	ICU	AssertionError when assertions are enabled; correct without them only by falling through assert(count >= 0)

Why are the changes needed?

SELECT instr('abcabc', 'abc', -2147483648, 1);      -- returns 4, expected 0
SELECT substring_index('a.b.c', '.', -2147483648);  -- NegativeArraySizeException, expected a.b.c
SELECT substring_index(collate('a.b.c', 'UTF8_LCASE'), '.', -2147483648);  -- returns '', expected a.b.c

Every other out-of-range position or count is handled correctly -- instr('abcabc', 'abc', -7, 1) returns 0 and substring_index('a.b.c', '.', -9) returns a.b.c -- so Int.MinValue is the one input where these functions diverge, and they diverge differently per collation, making the same query return different answers under UTF8_BINARY, UTF8_LCASE and ICU.

The substring_index defects affect released versions. instr with a start and occurrence is new in master (SPARK-57436) and has not been released.

Does this PR introduce any user-facing change?

Yes - bug fix

How was this patch tested?

New test cases added

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

Yes