Select each h1 element (in the node context document), immediately followed by the b element (without another intermediate element, although there may be intermediate text).
Destruction:
Select each h1 element that is a descendant of the node root of the document that contains the node context;
[...]
filter out any of these h1 elements that do not meet the following criteria:
[following-sibling::*[1]...]
so the first following sibling element passes this test:
[self::b]
self is an element of b . Literally, this last test means that when I start with the node context and select self (i.e. the node context) subject to the node test, which filters out everything except elements with the name b , the result is a nonempty set of node.
Larsh
source share