Optional Unicode 3.2.0 but enabled by default

Useful for smaller builds, like WebAssembly. The UCD 3.2.0 path is only directly used in a handful of places in Python. Assisted-by: GPT-5.6-Luna
joshuamegnauth54 · Oct 8, 2026 · Oct 9, 2026 · 4b3e66064d45586822ac733def65c8794305da32 · bschoenmaeckers · Oct 9, 2026
commit 4b3e66064d45586822ac733def65c8794305da32
diff --git a/.github/workflows/ci.yaml b/.github/workflows/ci.yaml
@@ -19,8 +19,8 @@ concurrency:
   cancel-in-progress: true
 
 env:
-  CARGO_ARGS: --no-default-features --features stdlib,importlib,stdio,encodings,sqlite,ssl-rustls-aws-lc,host_env,doc
-  CARGO_ARGS_NO_SSL: --no-default-features --features stdlib,importlib,stdio,encodings,sqlite,host_env,doc
+  CARGO_ARGS: --no-default-features --features stdlib,importlib,stdio,encodings,sqlite,ssl-rustls-aws-lc,host_env,doc,ucd_3_2_0
+  CARGO_ARGS_NO_SSL: --no-default-features --features stdlib,importlib,stdio,encodings,sqlite,host_env,doc,ucd_3_2_0
   CARGO_ARGS_MOST_FEATS: --features capi,encodings,jit,sqlite,threading
   # Crates excluded from workspace builds:
   # - rustpython_wasm: requires wasm target
@@ -1072,7 +1072,7 @@ jobs:
 
       - name: build rustpython
         if: env.RUN_WASM_WASI == 'true'
-        run: cargo build --locked --profile wasm-release --target wasm32-wasip1 --no-default-features --features freeze-stdlib,stdlib,stdio,importlib,host_env,doc --verbose
+        run: cargo build --locked --profile wasm-release --target wasm32-wasip1 --no-default-features --features freeze-stdlib,stdlib,stdio,importlib,host_env,doc,ucd_3_2_0 --verbose
       - name: run snippets
         if: env.RUN_WASM_WASI == 'true'
         run: |

diff --git a/.github/workflows/cron-ci.yaml b/.github/workflows/cron-ci.yaml
@@ -14,7 +14,7 @@ on:
       - .github/workflows/cron-ci.yaml
 
 env:
-  CARGO_ARGS: --no-default-features --features stdlib,importlib,stdio,encodings,ssl-rustls-aws-lc,jit,host_env,doc
+  CARGO_ARGS: --no-default-features --features stdlib,importlib,stdio,encodings,ssl-rustls-aws-lc,jit,host_env,doc,ucd_3_2_0
 
 jobs:
   # codecov collects code coverage data from the rust tests, python snippets and python test suite.
@@ -42,7 +42,7 @@ jobs:
       - run: sudo apt-get update && sudo apt-get -y install lcov
 
       - name: Run cargo-llvm-cov with Rust tests.
-        run: cargo llvm-cov --no-report --workspace --exclude rustpython_wasm --exclude rustpython-venvlauncher --exclude rustpython-capi --verbose --no-default-features --features stdlib,importlib,stdio,encodings,ssl-rustls-aws-lc,jit,host_env,doc
+        run: cargo llvm-cov --no-report --workspace --exclude rustpython_wasm --exclude rustpython-venvlauncher --exclude rustpython-capi --verbose --no-default-features --features stdlib,importlib,stdio,encodings,ssl-rustls-aws-lc,jit,host_env,doc,ucd_3_2_0
 
       - name: Run cargo-llvm-cov with Python snippets.
         run: python scripts/cargo-llvm-cov.py

diff --git a/Cargo.toml b/Cargo.toml
@@ -11,7 +11,7 @@ license.workspace = true
 
 [features]
 capi = ["dep:rustpython-capi", "threading"]
-default = ["threading", "stdlib", "stdio", "importlib", "ssl-rustls-aws-lc", "host_env", "mimalloc", "doc"]
+default = ["threading", "stdlib", "stdio", "importlib", "ssl-rustls-aws-lc", "host_env", "mimalloc", "doc", "ucd_3_2_0"]
 doc = ["rustpython-vm/doc", "rustpython-stdlib?/doc"]
 mimalloc = ["dep:mimalloc"]
 host_env = ["rustpython-vm/host_env", "rustpython-stdlib?/host_env"]
@@ -32,6 +32,7 @@ ssl-rustls-aws-lc-fips = ["ssl-rustls-aws-lc", "rustls/fips"]
 ssl-openssl = ["ssl", "rustpython-stdlib/ssl-openssl"]
 ssl-openssl-vendor = ["ssl-openssl", "rustpython-stdlib/ssl-openssl-vendor"]
 tkinter = ["rustpython-stdlib/tkinter"]
+ucd_3_2_0 = ["rustpython-stdlib?/ucd_3_2_0", "rustpython-capi?/ucd_3_2_0"]
 
 [build-dependencies]
 winresource = "0.1"
@@ -169,7 +170,7 @@ repository = "https://github.com/RustPython/RustPython"
 license = "MIT"
 
 [workspace.dependencies]
-rustpython-capi = { path = "crates/capi", version = "0.6.1" }
+rustpython-capi = { path = "crates/capi", version = "0.6.1", default-features = false }
 rustpython-compiler-core = { path = "crates/compiler-core", version = "0.6.1" }
 rustpython-compiler = { path = "crates/compiler", version = "0.6.1" }
 rustpython-codegen = { path = "crates/codegen", version = "0.6.1" }
@@ -183,7 +184,7 @@ rustpython-vm = { path = "crates/vm", default-features = false, version = "0.6.1
 rustpython-pylib = { path = "crates/pylib", version = "0.6.1" }
 rustpython-stdlib = { path = "crates/stdlib", default-features = false, version = "0.6.1" }
 rustpython-sre_engine = { path = "crates/sre_engine", version = "0.6.1" }
-rustpython-unicode = { path = "crates/unicode", version = "0.6.1" }
+rustpython-unicode = { path = "crates/unicode", default-features = false, version = "0.6.1" }
 rustpython-wtf8 = { path = "crates/wtf8", version = "0.6.1" }
 rustpython-doc = { path = "crates/doc", version = "0.6.1" }
 

diff --git a/crates/capi/Cargo.toml b/crates/capi/Cargo.toml
@@ -11,6 +11,10 @@ license.workspace = true
 [lib]
 crate-type = ["cdylib", "rlib"]
 
+[features]
+default = ["ucd_3_2_0"]
+ucd_3_2_0 = ["rustpython-stdlib/ucd_3_2_0"]
+
 [dependencies]
 bitflags = { workspace = true }
 itertools = { workspace = true }

diff --git a/crates/common/src/lock.rs b/crates/common/src/lock.rs
@@ -8,14 +8,16 @@ use lock_api::{
 
 cfg_select! {
     feature = "threading" => {
+        pub use core::cell::LazyCell;
         pub use detaching::{BlockingWaitHook, set_blocking_wait_hook, set_world_stopped};
         pub use parking_lot::{RawMutex, RawRwLock, RawThreadId};
         pub use std::sync::OnceLock as OnceCell;
-        pub use core::cell::LazyCell;
     }
     _ => {
         mod cell_lock;
-        pub use cell_lock::{RawCellMutex as RawMutex, RawCellRwLock as RawRwLock, SingleThreadId as RawThreadId};
+        pub use cell_lock::{
+            RawCellMutex as RawMutex, RawCellRwLock as RawRwLock, SingleThreadId as RawThreadId,
+        };
 
         pub use core::cell::{LazyCell, OnceCell};
     }
@@ -37,13 +39,19 @@ cfg_select! {
         unsafe impl<T, F> Sync for LazyLock<T, F> {}
 
         impl<T, F: FnOnce() -> T> LazyLock<T, F> {
-            pub const fn new(f: F) -> Self { Self(core::cell::LazyCell::new(f)) }
-            pub fn force(this: &Self) -> &T { core::cell::LazyCell::force(&this.0) }
+            pub const fn new(f: F) -> Self {
+                Self(core::cell::LazyCell::new(f))
+            }
+            pub fn force(this: &Self) -> &T {
+                core::cell::LazyCell::force(&this.0)
+            }
         }
 
         impl<T, F: FnOnce() -> T> core::ops::Deref for LazyLock<T, F> {
             type Target = T;
-            fn deref(&self) -> &T { &self.0 }
+            fn deref(&self) -> &T {
+                &self.0
+            }
         }
     }
 }

diff --git a/crates/stdlib/Cargo.toml b/crates/stdlib/Cargo.toml
@@ -11,7 +11,8 @@
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [features]
-default = ["compiler", "host_env", "doc"]
+default = ["compiler", "host_env", "doc", "ucd_3_2_0"]
+ucd_3_2_0 = ["rustpython-unicode/ucd_3_2_0"]
 doc = ["rustpython-vm/doc"]
 host_env = ["rustpython-vm/host_env"]
 compiler = ["rustpython-vm/compiler"]
@@ -73,7 +74,7 @@

 # tkinter
 tk-sys = { workspace = true, optional = true }
 tcl-sys = { workspace = true, optional = true }
 widestring = { workspace = true, optional = true }

 # uuid

diff --git a/crates/stdlib/src/_queue.rs b/crates/stdlib/src/_queue.rs
@@ -21,7 +21,7 @@ mod _queue {
             use parking_lot::{Condvar, Mutex, MutexGuard};
 
             type Buf = Mutex<BufInner>;
-        },
+        }
         _ => {
             use crate::common::lock::PyMutex;
 

diff --git a/crates/stdlib/src/faulthandler.rs b/crates/stdlib/src/faulthandler.rs
@@ -679,10 +679,9 @@ mod decl {
                             // frames are still alive because the target thread
                             // is executing (inside a blocking call).
                             for (tid, slot) in &thread_frame_slots {
-                                let iframe_ptr = slot
-                                    .top_iframe
-                                    .load(core::sync::atomic::Ordering::Relaxed)
-                                    as *const rustpython_vm::frame::InterpreterFrame;
+                                let iframe_ptr =
+                                    slot.top_iframe.load(core::sync::atomic::Ordering::Relaxed)
+                                        as *const rustpython_vm::frame::InterpreterFrame;
                                 write_thread_id(fd, *tid, false);
                                 if iframe_ptr.is_null() {
                                     puts(fd, "  <no Python frame>\n");
@@ -703,10 +702,9 @@ mod decl {
                         }
                         all(not(unix), feature = "threading") => {
                             for (tid, slot) in &thread_frame_slots {
-                                let iframe_ptr = slot
-                                    .top_iframe
-                                    .load(core::sync::atomic::Ordering::Relaxed)
-                                    as *const rustpython_vm::frame::InterpreterFrame;
+                                let iframe_ptr =
+                                    slot.top_iframe.load(core::sync::atomic::Ordering::Relaxed)
+                                        as *const rustpython_vm::frame::InterpreterFrame;
                                 write_thread_id(fd, *tid, false);
                                 if iframe_ptr.is_null() {
                                     puts(fd, "  <no Python frame>\n");

diff --git a/crates/stdlib/src/openssl.rs b/crates/stdlib/src/openssl.rs
@@ -1338,7 +1338,8 @@ mod _ssl {
             cfg_select! {
                 ossl110 => {
                     let ctx = zelf.builder();
-                    let result = unsafe { sys::SSL_CTX_set_num_tickets(ctx.as_ptr(), value as usize) };
+                    let result =
+                        unsafe { sys::SSL_CTX_set_num_tickets(ctx.as_ptr(), value as usize) };
                     if result != 1 {
                         return Err(vm.new_value_error("failed to set num tickets."));
                     }
@@ -1360,11 +1361,10 @@ mod _ssl {
                         .load_verify_locations(Some(cert_file.as_path()), Some(cert_dir.as_path()))
                         .map_err(|e| convert_openssl_error(vm, e))
                 }
-                _ => {
-                    zelf.builder()
-                        .set_default_verify_paths()
-                        .map_err(|e| convert_openssl_error(vm, e))
-                }
+                _ => zelf
+                    .builder()
+                    .set_default_verify_paths()
+                    .map_err(|e| convert_openssl_error(vm, e)),
             }
         }
 
@@ -2828,11 +2828,9 @@ mod _ssl {
                         Ok(())
                     }
                 }
-                _ => {
-                    Err(vm.new_not_implemented_error(
-                        "Post-handshake auth is not supported by your OpenSSL version.",
-                    ))
-                }
+                _ => Err(vm.new_not_implemented_error(
+                    "Post-handshake auth is not supported by your OpenSSL version.",
+                )),
             }
         }
 

diff --git a/crates/stdlib/src/resource.rs b/crates/stdlib/src/resource.rs
@@ -17,9 +17,7 @@ mod resource {
 
     #[cfg_attr(target_os = "android", expect(deprecated))]
     const RLIM_NLIMITS: i32 = cfg_select! {
-        target_os = "android" => {
-            host_resource::RLIM_NLIMITS
-        }
+        target_os = "android" => host_resource::RLIM_NLIMITS,
         _ => {
             // This constant isn't abi-stable across os versions, so we just
             // pick a high number so we don't get false positive ValueErrors and just bubble up the

diff --git a/crates/stdlib/src/unicodedata.rs b/crates/stdlib/src/unicodedata.rs
@@ -33,7 +33,7 @@ impl<'a> TryFromBorrowedObject<'a> for NormalizeFormArg {
 mod unicodedata {
     use super::{NormalizeFormArg, unicode_core};
     use crate::vm::{
-        Py, PyObjectRef, PyPayload, PyRef, PyResult, VirtualMachine,
+        Py, PyObjectRef, PyPayload, PyResult, VirtualMachine,
         builtins::{PyModule, PyStr, PyStrRef},
         function::OptionalArg,
     };
@@ -244,8 +244,9 @@ mod unicodedata {
         }
     }
 
+    #[cfg(feature = "ucd_3_2_0")]
     #[pyattr]
-    fn ucd_3_2_0(vm: &VirtualMachine) -> PyRef<Ucd> {
+    fn ucd_3_2_0(vm: &VirtualMachine) -> crate::vm::PyRef<Ucd> {
         Ucd::new(false).into_ref(&vm.ctx)
     }
 

diff --git a/crates/unicode/Cargo.toml b/crates/unicode/Cargo.toml
@@ -8,6 +8,10 @@ repository = { workspace = true }
 license = { workspace = true }
 rust-version = { workspace = true }
 
+[features]
+default = ["ucd_3_2_0"]
+ucd_3_2_0 = []
+
 [dependencies]
 rustpython-wtf8 = { workspace = true }
 

diff --git a/crates/unicode/build.rs b/crates/unicode/build.rs
@@ -678,10 +678,14 @@ fn generate_numeric_value(
         "(u32, u32, f64)",
         &mut values_latest,
     );
+    #[cfg(feature = "ucd_3_2_0")]
     let mut ucd32_diffs: Vec<_> = ucd32_diffs
         .into_iter()
         .map(|((start, end), value)| (start, end, value))
         .collect();
+    #[cfg(not(feature = "ucd_3_2_0"))]
+    let mut ucd32_diffs = Vec::new();
+
     write_slice_debug(
         &mut writer,
         "NUMERIC_VALUES_DIFF",

diff --git a/crates/unicode/src/data.rs b/crates/unicode/src/data.rs
@@ -19,21 +19,46 @@ use icu_properties::props::{
 };
 use rustpython_wtf8::CodePoint;
 
+// Latest Unicode
 include!(concat!(env!("OUT_DIR"), "/generated/algo_names.rs"));
 include!(concat!(env!("OUT_DIR"), "/generated/name_lookups.rs"));
+include!(concat!(env!("OUT_DIR"), "/generated/decomp.rs"));
+include!(concat!(env!("OUT_DIR"), "/generated/numeric_value_3_2.rs"));
+
+// Unicode 3.2.0
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(env!("OUT_DIR"), "/generated/bidi_class_3_2.rs"));
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(env!("OUT_DIR"), "/generated/binary_props_3_2.rs"));
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(
     env!("OUT_DIR"),
     "/generated/combining_class_3_2.rs"
 ));
-include!(concat!(env!("OUT_DIR"), "/generated/decomp.rs"));
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(env!("OUT_DIR"), "/generated/eaw_3_2.rs"));
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(env!("OUT_DIR"), "/generated/gen_cat_3_2.rs"));
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(env!("OUT_DIR"), "/generated/membership_3_2.rs"));
-include!(concat!(env!("OUT_DIR"), "/generated/numeric_value_3_2.rs"));
+#[cfg(feature = "ucd_3_2_0")]
 include!(concat!(env!("OUT_DIR"), "/generated/num_type_3_2.rs"));
 
+#[cfg(not(feature = "ucd_3_2_0"))]
+const BIDI_CLASS_DIFF: &[(u32, u32, BidiClass)] = &[];
+#[cfg(not(feature = "ucd_3_2_0"))]
+const BIDI_MIRRORED: &[(u32, u32)] = &[];
+#[cfg(not(feature = "ucd_3_2_0"))]
+const COMBINING_CLASS: &[(u32, u32, CanonicalCombiningClass)] = &[];
+#[cfg(not(feature = "ucd_3_2_0"))]
+const EAST_ASIAN_WIDTH: &[(u32, u32, EastAsianWidth)] = &[];
+#[cfg(not(feature = "ucd_3_2_0"))]
+const GENERAL_CATEGORY: &[(u32, u32, GeneralCategory)] = &[];
+#[cfg(not(feature = "ucd_3_2_0"))]
+const MEMBERSHIP_3_2: &[(u32, u32)] = &[];
+#[cfg(not(feature = "ucd_3_2_0"))]
+const NUMERIC_TYPE_DIFF: &[(u32, u32, NumericType)] = &[];
+
 #[derive(Clone, Copy)]
 enum DecompositionType {
     Compat,
@@ -511,12 +536,14 @@ mod tests {
     }
 
     #[test]
+    #[cfg(feature = "ucd_3_2_0")]
     fn ucd_3_2_0_view_differs_from_modern() {
         let legacy = Ucd::new(false);
         assert_eq!(legacy.unidata_version(), "3.2.0");
     }
 
     #[test]
+    #[cfg(feature = "ucd_3_2_0")]
     fn name_aliases_resolve_and_do_not_replace_the_official_name() {
         assert_eq!(
             lookup_character("LATIN CAPITAL LETTER GHA"),
@@ -534,6 +561,7 @@ mod tests {
     }
 
     #[test]
+    #[cfg(feature = "ucd_3_2_0")]
     fn named_sequences_resolve_only_on_the_modern_view() {
         assert_eq!(
             lookup_name("LATIN SMALL LETTER R WITH TILDE"),

diff --git a/crates/vm/build.rs b/crates/vm/build.rs
@@ -162,14 +162,16 @@ fn get_version(cpy_version: &str) -> (String, String) {
     // Windows: include MSC v. for compatibility with ctypes.util.find_library
     // MSC v.1929 = VS 2019, version 14+ makes find_msvcrt() return None
     let msc_info = cfg_select! {
-        windows => {{
-            // Include both RustPython identifier and MSC v. for compatibility
-            if cfg!(target_pointer_width = "64") {
-                " MSC v.1929 64 bit (AMD64)"
-            } else {
-                " MSC v.1929 32 bit (Intel)"
+        windows => {
+            {
+                // Include both RustPython identifier and MSC v. for compatibility
+                if cfg!(target_pointer_width = "64") {
+                    " MSC v.1929 64 bit (AMD64)"
+                } else {
+                    " MSC v.1929 32 bit (Intel)"
+                }
             }
-        }},
+        }
         _ => "",
     };
 

diff --git a/crates/vm/src/compiler.rs b/crates/vm/src/compiler.rs
@@ -11,8 +11,8 @@ cfg_select! {
     _ => {
         pub use ruff_python_parser as parser;
 
-        pub use rustpython_compiler_core::Mode;
         pub use rustpython_compiler_core as core;
+        pub use rustpython_compiler_core::Mode;
     }
 }
 

diff --git a/crates/vm/src/exceptions.rs b/crates/vm/src/exceptions.rs
@@ -2726,7 +2726,9 @@ pub(super) mod types {
                             .filter(|f| !vm.is_none(f));
 
                         let winerror: Option<PyObjectRef> = cfg_select! {
-                            windows => obj.get_attr("winerror", vm).ok().filter(|w| !vm.is_none(w)),
+                            windows => {
+                                obj.get_attr("winerror", vm).ok().filter(|w| !vm.is_none(w))
+                            }
                             _ => None,
                         };
-Original file line number
+Diff line change
@@ Expand Up / @@ -21,7 +21,7 @@ mod _queue { @@
                 use parking_lot::{Condvar, Mutex, MutexGuard};
                 type Buf = Mutex<BufInner>;
-            },
+            }
             _ => {
                 use crate::common::lock::PyMutex;
@@ Expand Down @@