pub fn extract_with_tail(
filename: &str,
head: &[u8],
tail: &[u8],
templates: &[ResultTemplate],
) -> Vec<ResultInput>Expand description
extract for a file larger than SCAN_MAX_BYTES: head is its
beginning, tail its end (at most SCAN_MAX_BYTES each, the tail
starting at a line; empty when the head holds the whole file). A
template with pick = last looks at the tail first – a solver writes its
final values at the end – and falls back to the head; table headers
(csv, table) always come from the head. pick = first reads the
head only. A tail holding a NUL (zero padding) is ignored. For a file
below twice SCAN_MAX_BYTES head and tail overlap: a value found in
the overlap is labelled by its line from the end, not its line number.