Skip to main content

FullPath

property FullPath: UTF8String read;

Example​

procedure CapturePath(Node: TXMLNode);
var
  Path: UTF8String;
begin
  if Node <> nil then
    Path := Node.FullPath;
end;

Usage​

FullPath returns a slash-prefixed path made from the names of this node and its containing nodes. It does not escape names or include indexes or namespace details.

Additional Technical Info​

FullPath recursively concatenates each ancestor's current Name, prefixing every segment with /. A top-level node named Order returns /Order; its child Line returns /Order/Line.

The value is generated on demand. Renaming this node or any ancestor immediately changes it. A detached/top-level node has no Parent segment, so it still returns /<Name>; an unnamed node returns / at that level.

This is a NativeXML convenience string, not XPath and not a stable identifier. It has:

  • no escaping for / or other characters in names;
  • no sibling indices, so duplicate-name siblings have identical paths;
  • no distinction between an attribute, element or other named node;
  • no namespace URI resolution; prefixes remain literal text; and
  • no document identity, so separate documents can produce the same path.

FindNode treats a search beginning with / as a case-insensitive comparison against this exact string. That means its full-path mode also inherits every ambiguity and does not implement XPath predicates or descendant syntax.

The calculation recursively walks Parent. Cost grows with tree depth and repeated FindNode full-path comparisons rebuild paths many times. A cycle created through unsafe NodeAdd use can cause unbounded recursion/stack exhaustion.

The returned UTF8String is a value; the caller owns no node through it. Non-ASCII segments must remain valid UTF-8 bytes through PascalScript's AnsiString alias.

The example is source-reviewed only; no path was evaluated.

External references

Created 2026-07-15