Expand description
Encoding / locale probing utilities.
This module exists mainly for debugging + experiments around R’s string encodings. R’s runtime has both:
- per-CHARSXP encoding tags (UTF-8 / Latin-1 / bytes / native)
- global/locale-level settings (native encoding, UTF-8 locale flags)
§Availability
The global signals are non-API (from Defn.h) and require the nonapi
feature. Only the locale flags R’s shared library exports are read
(utf8locale, mbcslocale, known_to_be_latin1); the hidden ones
(known_to_be_utf8, latin1locale, R_nativeEncoding) must not even be
referenced — an eager data relocation against a hidden symbol aborts
dyn.load of the whole package (see sys::nonapi_encoding).
miniextendr_encoding_init() is not called by package_init; it exists
for debugging and for standalone Rust applications embedding R.
A package loads in any locale. String conversions read each CHARSXP’s
encoding mark and translate when needed; in a session whose locale isn’t
UTF-8, the first translation of a native non-ASCII string raises one
warning (NATIVE_TRANSLATION_CLASS).
Re-exports§
use std::sync::OnceLock;🔒
Structs§
- REncoding
Info - Cached snapshot of R’s encoding / locale state at init time.
Constants§
- NATIVE_
TRANSLATION_ CLASS - The class of the warning a package raises the first time it translates a native string to UTF-8.
Statics§
Functions§
- encoding_
info - Return the cached encoding info (if
miniextendr_encoding_init()has run). - miniextendr_
encoding_ init - Initialize / snapshot R’s encoding state.
- warn_
native_ 🔒translation - Queue a warning, once per session for each package, the first time a
native (unmarked) non-ASCII string is translated to UTF-8: the session’s
locale isn’t UTF-8, and bytes its encoding can’t represent came out as
<xx>escapes.