Parser Reuse

What you'll learn

• Why one long-lived parser beats a parser per parse.

Complete example

var parser = treesitter.Parser.init(allocator);
defer parser.deinit();
try parser.setLanguage(lang);

var t1 = try parser.parseString("1 + 2");
defer t1.deinit();
var t2 = try parser.parseString("3 * 4");
defer t2.deinit();

How it works

The parser keeps its state stack, token scratch space, reuse-candidate list, and included-range list between parses with retained capacity. Each parse clears lengths but keeps buffers, so steady-state editor workloads allocate only for genuinely new tree content. setLanguage also clears the stack (states belong to a grammar). Call parser.reset() to additionally drop retained scratch capacity between unrelated workloads.

API used

• Parser — parseString, parse, reset.

Related guides

• Tree Reuse, Incremental Parsing, Performance