3.4 String Types🔗ℹ

3.4.1 Primitive String Types🔗ℹ

See also _bytes/nul-terminated and _bytes for converting between byte strings and C’s char* type.

For the CS implementation of Racket, the conversion of a Racket string for the foreign side is a copy of the Racket representation, where the copy is managed by the garbage collector.

For the BC implementation of Racket, the conversion of a Racket string for the foreign side shares memory with the Racket string representation, since UCS-4 is the native representation format for those variants. The foreign pointer corresponds to the mzchar* type in Racket’s C API.

The conversion of a Racket string for the foreign side is a copy of the Racket representation (reencoded), where the copy is managed by the garbage collector.

For the BC implementation of Racket, the conversion of a Racket path for the foreign side shares memory with the Racket path representation. Otherwise (for the CS implementation or for Racket strings), conversion for the foreign side creates a copy that is managed by the garbage collector.

Beware that changing the current directory via current-directory does n<ot change the OS-level current directory as seen by foreign library functions. Paths normally should be converted to absolute form using path->complete-path (which uses the current-directory parameter) before passing them to a foreign function.

For the CS implementation of Racket, the conversion of a Racket symbol for the foreign side is a copy of the Racket representation, where the copy is managed by the garbage collector.

For the BC implementation of Racket, the conversion of a Racket symbol for the foreign side shares memory with the Racket symbol representation, but points to the middle of the symbol’s allocated memory—so the string pointer must not be used across a garbage collection.

3.4.2 Fixed Auto-Converting String Types🔗ℹ

3.4.3 Variable Auto-Converting String Type🔗ℹ

The _string/ucs-4 type is rarely useful when interacting with foreign code, while using _bytes/nul-terminated is somewhat unnatural, since it forces Racket programmers to use byte strings. Using _string/utf-8, etc., meanwhile, may prematurely commit to a particular encoding of strings as bytes. The _string type supports conversion between Racket strings and char* strings using a parameter-determined conversion.

Don’t use _string when you should use _path. Although C APIs typically represent paths as strings, and although the default _string (via default-_string-type) even implicitly converts Racket paths to strings, using _path ensures the proper encoding of strings as paths, which is not always UTF-8. See also _path for a caveat about relative paths.

3.4.4 Other String Types🔗ℹ