Concrete output format. "jsonl" is the "json" document written one
entry per line, each line compact and ending in \n. "text" is lossy
plain-text extraction,
"semantic-tokens" is parser-derived highlighting in the LSP shape (see
semanticTokens), "diagnostics" is every issue the JSON envelope
would carry as LSP diagnostics with encoded positions (see
diagnostics), "lex" is the lexer's token stream as a JSON
array ([{ kind, span, text }], byte spans; debug tooling) and "ast" is
the parser's tree as JSON (debug tooling too, PLAN-209: it shows the runs,
dividers and spans a bitmark patch path addresses; the shape follows
the parser and is not a contract) — all five are output-only (they cannot
be used as an input format), and the last four take bitmark input only. Any config mapping id ("html", "xml", …) is
accepted too.
Concrete output format.
"jsonl"is the"json"document written one entry per line, each line compact and ending in\n."text"is lossy plain-text extraction,"semantic-tokens"is parser-derived highlighting in the LSP shape (see semanticTokens),"diagnostics"is every issue the JSON envelope would carry as LSP diagnostics with encoded positions (see diagnostics),"lex"is the lexer's token stream as a JSON array ([{ kind, span, text }], byte spans; debug tooling) and"ast"is the parser's tree as JSON (debug tooling too, PLAN-209: it shows the runs, dividers and spans a bitmark patch path addresses; the shape follows the parser and is not a contract) — all five are output-only (they cannot be used as an input format), and the last four take bitmark input only. Any config mapping id ("html","xml", …) is accepted too.