gudusoft.gsqlparser.pp2.token

Class TokenCoverage

Object

gudusoft.gsqlparser.pp2.token.TokenCoverage

public final class TokenCoverage
extends Object

Shared helpers for reasoning about a TSourceToken's position coverage on the input.

Several pp2 stages walk a TSourceTokenList in offset order and need to skip tokens that overlap an already-emitted span. The pattern shows up in:

• Pp2TokenStreamBuilder — folding whitespace, skipping phantom inner tokens emitted by GSP's lexer for shapes like ${name}.
• SourceSpanLedger — covering every byte exactly once.
• TokenEquivalence — building the comparable-token sequence.
• (planned) S11 statement-boundary detector — locating statement terminators inside a multi-region stream.

Centralizing the logic here avoids the predicate drifting across call sites as Phase 2 lands.

Plan reference: §7.3/S9 (the GSP ${name} probe finding).

Method Summary

Modifier and Type	Method and Description
static long	endOffset(TSourceToken token) Convenience: compute the half-open end offset (offset + text.length()) for a token.
static boolean	isFullyShadowed(TSourceToken token, long priorEnd) True when token's source range is FULLY contained in an already-emitted span ending at priorEnd.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

isFullyShadowed

public static boolean isFullyShadowed(TSourceToken token,
                                      long priorEnd)

True when token's source range is FULLY contained in an already-emitted span ending at priorEnd.

Fully-contained tokens are phantom inner tokens emitted by the GSP lexer (e.g., the "{" "name" "}" that follow an outer "${name}" identifier). Skipping them is safe because the outer token already carries the same bytes.

Partial overlap (start shadowed but end extending past) is NOT fully contained; callers must handle it separately (either throw loudly or continue processing — never silently drop, since that would lose bytes).

Parameters:

token - the candidate token; null is tolerated and returns false so callers can treat it as "not shadowed; handle null separately"

priorEnd - the source offset where the previously emitted span ends; pass -1 when nothing has been emitted yet

Returns:

true iff token is fully shadowed

endOffset

public static long endOffset(TSourceToken token)

Convenience: compute the half-open end offset (offset + text.length()) for a token. Returns the token's offset when the text is null or empty.