[ydah]

getAttributes() uses the following token's line as endLine. If the node's final token ends with a newline, this includes the next line:

$stmts = $parser->parse("<?php\necho 1 ?>\nfoo\n");
// Echo:       before 2-3, after 2-2
// InlineHTML: before 3-4, after 3-3

Subtract one when the final token ends with \n, matching the inclusive node range. This also corrects interpolated string parts and their heredoc diagnostic positions.

[nikic]

This is working as intended. If the node includes a trailing newline, then it includes a trailing newline, and positions are reported accordingly.