Add tzdb_data crate

Splitting up the generated data makes maintenance easier.

.github/workflows/ci.yml

@@ -69,8 +69,8 @@ jobs:
           RUSTC_BOOTSTRAP: 1
         run: cargo generate-lockfile ${{ matrix.versions }}
 
-      - run: cargo check --package tzdb --all-targets
-      - run: cargo clippy --package tzdb --all-targets -- -D warnings
+      - run: cargo check --workspace --all-targets
+      - run: cargo clippy --workspace --all-targets -- -D warnings
 
   clippy-pedantic:
     runs-on: ubuntu-20.04
@@ -85,7 +85,7 @@ jobs:
         with:
           components: clippy
 
-      - run: cargo clippy --package tzdb --all-targets -- -D clippy::pedantic
+      - run: cargo clippy --workspace --all-targets -- -D clippy::pedantic
 
   test:
     runs-on: ubuntu-20.04
@@ -100,9 +100,9 @@ jobs:
         with:
           components: clippy
 
-      - run: cargo check --package tzdb --all-targets
-      - run: cargo clippy --package tzdb --all-targets -- -D warnings
-      - run: cargo test --package tzdb --all-targets
+      - run: cargo check --workspace --all-targets
+      - run: cargo clippy --workspace --all-targets -- -D warnings
+      - run: cargo test --workspace --all-targets
       - run: cargo run --package current-time
       - run: cd testing && cargo test
 
@@ -122,10 +122,10 @@ jobs:
           components: miri
 
       - name: Test (tzdb)
-        run: cargo miri test --package tzdb --all-targets
+        run: cargo miri test --workspace --all-targets
 
       - name: Test (testing)
-        run: cd testing && cargo miri test --package tzdb --all-targets
+        run: cd testing && cargo miri test --workspace --all-targets
 
   cross-miri:
     strategy:
@@ -153,7 +153,7 @@ jobs:
           target: ${{ matrix.target }}
 
       - name: Test
-        run: cargo miri test --package tzdb --target ${{ matrix.target }}
+        run: cargo miri test --workspace --target ${{ matrix.target }}
 
   doc:
     runs-on: ubuntu-20.04
@@ -169,7 +169,7 @@ jobs:
           toolchain: nightly
           components: rust-docs
 
-      - run: cargo doc --package tzdb --all-features --no-deps
+      - run: cargo doc --workspace --all-features --no-deps
         env:
           RUSTDOCFLAGS: -D warnings --cfg docsrs
 
@@ -189,68 +189,6 @@ jobs:
         with:
           token: ${{ secrets.GITHUB_TOKEN }}
 
-  build-cross:
-    strategy:
-      fail-fast: false
-      matrix:
-        target:
-          - x86_64-unknown-freebsd
-          - x86_64-unknown-illumos
-          - x86_64-unknown-netbsd
-          - x86_64-linux-android
-          - i686-linux-android
-          - arm-linux-androideabi
-          - aarch64-linux-android
-    runs-on: ubuntu-20.04
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: recursive
-
-      - name: Setup Rust
-        uses: ./.github/actions/setup-rust
-        with:
-          toolchain: stable
-          target: ${{ matrix.target }}
-          components: clippy
-
-      - name: Install "cross"
-        run: curl --location --silent --show-error --fail https://github.com/cargo-bins/cargo-quickinstall/releases/download/cross-0.2.5-x86_64-unknown-linux-gnu/cross-0.2.5-x86_64-unknown-linux-gnu.tar.gz | tar -xzvvf - -C $HOME/.cargo/bin
-      - run: cross check --target ${{ matrix.target }} --package tzdb --all-targets
-      - run: cross clippy --target ${{ matrix.target }} --package tzdb --all-targets -- -D warnings
-      - run: cross build --target ${{ matrix.target }} --package tzdb --lib
-      - run: cross build --target ${{ matrix.target }} --examples
-
-  build-cross-ios:
-    strategy:
-      fail-fast: false
-      matrix:
-        target:
-          - aarch64-apple-ios-sim
-          - aarch64-apple-ios
-          - x86_64-apple-ios
-    runs-on: macos-12
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: recursive
-
-      - name: Setup Rust
-        uses: ./.github/actions/setup-rust
-        with:
-          toolchain: stable
-          target: ${{ matrix.target }}
-          components: clippy
-
-      - name: Install "cross"
-        run: curl --location --silent --show-error --fail https://github.com/cargo-bins/cargo-quickinstall/releases/download/cross-0.2.5-x86_64-apple-darwin/cross-0.2.5-x86_64-apple-darwin.tar.gz | tar -xzvvf - -C $HOME/.cargo/bin
-      - run: cross check --target ${{ matrix.target }} --package tzdb --all-targets
-      - run: cross clippy --target ${{ matrix.target }} --package tzdb --all-targets -- -D warnings
-      - run: cross build --target ${{ matrix.target }} --package tzdb --lib
-      - run: cross build --target ${{ matrix.target }} --examples
-
   package:
     runs-on: ubuntu-20.04
     steps:
@@ -264,7 +202,7 @@ jobs:
         with:
           toolchain: stable
 
-      - run: cargo package --package tzdb
+      - run: cargo package --workspace
 
   devskim:
     name: DevSkim

.gitmodules

@@ -1,4 +1,4 @@
 [submodule "src/generated"]
-	path = src/generated
+	path = tzdb_data/src/generated
 	url = ./
 	branch = generated

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tzdb"
-version = "0.6.0"
+version = "0.6.1-pre.1"
 edition = "2018"
 authors = ["René Kijewski <crates.io@k6i.de>"]
 repository = "https://github.com/Kijewski/tzdb"
@@ -12,8 +12,9 @@ readme = "README.md"
 # rust-version = "1.56"
 
 [dependencies]
-tz-rs = { version = "^0.6.14", default-features = false, features = ["const", "std"] }
 iana-time-zone = { version = "^0.1.50", default-features = false, features = ["fallback"], optional = true }
+tzdb_data = { version = "0.1.0-pre.1", default-features = false, path = "tzdb_data" }
+tz-rs = { version = "^0.6.14", default-features = false, features = ["const", "std"] }
 
 [features]
 default = ["local"]
@@ -25,4 +26,8 @@ all-features = true
 rustdoc-args = ["--cfg", "docsrs"]
 
 [workspace]
-members = [".", "examples/current-time"]
+members = [
+    ".",
+    "examples/current-time",
+    "tzdb_data",
+]

Makefile

@@ -1,10 +1,16 @@
-.DELETE_ON_ERROR:
+#.DELETE_ON_ERROR:
 
 TZDB_VERSION := tzdb-2023d
 
-src/generated/mod.rs: tmp/${TZDB_VERSION}/usr/share/zoneinfo/ tzdb.tar.lz.sha
+tzdb_data/src/generated/mod.rs: tmp/${TZDB_VERSION}/usr/share/zoneinfo/ tzdb.tar.lz.sha
 	cd make-tzdb && cargo r -- ../$(@D) ../$< ../tzdb.tar.lz.sha
-	cargo +nightly fmt -- $(@D)/mod.rs $(@D)/by_name.rs $(@D)/raw_tzdata.rs $(@D)/test_all_names.rs
+	cargo +nightly fmt -- \
+	    $(@D)/by_name.rs \
+	    $(@D)/mod.rs \
+	    $(@D)/raw_tzdata.rs \
+	    $(@D)/test_all_names.rs \
+	    $(@D)/time_zone.rs \
+	    $(@D)/tzdata.rs
 
 tmp/${TZDB_VERSION}/usr/share/zoneinfo/: tmp/${TZDB_VERSION}/
 	cd tmp/${TZDB_VERSION}/ && make PACKRATDATA=backzone PACKRATLIST=zone.tab TOPDIR="." install

benchmarks/Cargo.toml

@@ -10,7 +10,7 @@ license = "Apache-2.0"
 publish = false
 
 [dependencies]
-tzdb = { version = "0.6.0", path = ".." }
+tzdb_data = { path = "../tzdb_data" }
 
 [dev-dependencies]
 criterion = { version = "0.5", default-features = false }

benchmarks/benches/by-name.rs

@@ -4,14 +4,14 @@ use std::time::{Duration, Instant};
 use rand::seq::{IteratorRandom, SliceRandom};
 use rand::SeedableRng;
 use rand_xoshiro::Xoroshiro128PlusPlus;
-use tzdb::{raw_tz_by_name, TZ_NAMES};
+use tzdb_data::{find_raw, TZ_NAMES};
 
 fn benchmark_by_name(c: &mut criterion::Criterion) {
     // collect all names with "random" capitalization
     let mut names: Vec<(String, usize)> = TZ_NAMES
         .iter()
         .flat_map(|&name| {
-            let raw_len = raw_tz_by_name(name).unwrap().len();
+            let raw_len = find_raw(name.as_bytes()).unwrap().len();
             let upper = name.to_uppercase();
             let lower = name.to_lowercase();
             let inverted = name
@@ -75,12 +75,12 @@ fn benchmark_by_name(c: &mut criterion::Criterion) {
         let city = std::str::from_utf8(city).unwrap();
 
         let raw_name = format!("{}/{}", continent, city);
-        let raw_len = crate::raw_tz_by_name(&raw_name).unwrap_or_default().len();
+        let raw_len = crate::find_raw(raw_name.as_bytes()).unwrap_or_default().len();
         names.push((raw_name, raw_len));
     }
 
     // benchmark per name lookup time
-    c.bench_function("tzdb::raw_tz_by_name", |b| {
+    c.bench_function("tzdb::find_raw", |b| {
         b.iter_custom(|iters| {
             let mut nanos = 0;
             for i in 0..iters {
@@ -91,7 +91,7 @@ fn benchmark_by_name(c: &mut criterion::Criterion) {
                 for &(ref name, raw_len) in names {
                     assert_eq!(
                         raw_len,
-                        crate::raw_tz_by_name(name).unwrap_or_default().len(),
+                        crate::find_raw(name.as_bytes()).unwrap_or_default().len(),
                     );
                 }
                 nanos += start.elapsed().as_nanos();

examples/current-time/src/main.rs

@@ -1,10 +1,12 @@
+#![allow(clippy::pedantic)]
+
 use std::env::args;
 use std::process::exit;
 
 use tzdb::{local_tz, now, time_zone, tz_by_name, TZ_NAMES};
 
 pub fn main() -> Result<(), now::NowError> {
-    let mut args = args().into_iter().fuse();
+    let mut args = args().fuse();
     let exe = args.next();
     let exe = exe.as_deref().unwrap_or("current-time");
     let argument = args.next();

fuzz/Cargo.toml

@@ -12,7 +12,7 @@ edition = "2018"
 cargo-fuzz = true
 
 [dependencies]
-tzdb = { version = "0.6.0", path = ".." }
+tzdb_data = { path = "../tzdb_data" }
 
 afl = { version = "0.15", optional = true }
 honggfuzz = { version = "0.5", optional = true }

fuzz/bin/afl.rs

@@ -1,5 +1,5 @@
 fn main() {
     afl::fuzz!(|name: &[u8]| {
-        let _ = tzdb::tz_by_name(name);
+        let _ = tzdb_data::find_tz(name);
     });
 }

fuzz/bin/honggfuzz.rs

@@ -1,7 +1,7 @@
 fn main() {
     loop {
         honggfuzz::fuzz!(|name: &[u8]| {
-            let _ = tzdb::tz_by_name(name);
+            let _ = tzdb_data::find_tz(name);
         });
     }
 }

fuzz/bin/libfuzzer.rs

@@ -1,5 +1,5 @@
 #![no_main]
 
 libfuzzer_sys::fuzz_target!(|name: &[u8]| {
-    let _ = tzdb::tz_by_name(name);
+    let _ = tzdb_data::find_tz(name);
 });

make-tzdb/generate_lookup_table.py

@@ -137,13 +137,14 @@ def convert(stdin, stdout):
 
     print('use tz::TimeZoneRef;', file=stdout)
     print(file=stdout)
+    print('use crate::eq_ignore_ascii_case;', file=stdout)
     print('use super::raw_tzdata;', file=stdout)
     print('use super::tzdata;', file=stdout)
     print(file=stdout)
 
     print('#[derive(Clone, Copy)]', file=stdout)
     print('#[repr(u16)]', file=stdout)
-    print('enum Index {', file=stdout)
+    print('pub(crate) enum Index {', file=stdout)
     idx = 0
     for entry in table:
         match entry:
@@ -174,15 +175,15 @@ def convert(stdin, stdout):
     print('];', file=stdout)
     print(file=stdout)
 
-    print(f'const TIME_ZONES: [&TimeZoneRef<\'static>; {entry_count}] = [', file=stdout)
+    print(f'pub(crate) const TIME_ZONES: [&TimeZoneRef<\'static>; {entry_count}] = [', file=stdout)
     for entry in table:
         match entry:
             case (name, canon):
                 print(f'    &tzdata::{canon},', file=stdout)
     print('];', file=stdout)
     print(file=stdout)
 
-    print(f'const RAW_TIME_ZONES: [&[u8]; {entry_count}] = [', file=stdout)
+    print(f'pub(crate) const RAW_TIME_ZONES: [&[u8]; {entry_count}] = [', file=stdout)
     for entry in table:
         match entry:
             case (name, canon):
@@ -197,7 +198,7 @@ def convert(stdin, stdout):
     print(f'{max_hash_value + 1}];', file=stdout)
     print(file=stdout)
 
-    print('fn find_key(s: &[u8]) -> Option<Index> {', file=stdout)
+    print('pub(crate) const fn find_key(s: &[u8]) -> Option<Index> {', file=stdout)
     print('    let len = s.len();', file=stdout)
     print(f'    if !matches!(len, {min_word_length}..={max_word_length}) {{', file=stdout)
     print('        return None;', file=stdout)
@@ -239,26 +240,18 @@ def hash_add(idx, offs):
     print(f'    if key > {max_hash_value} {{', file=stdout)
     print('        return None;', file=stdout)
     print('    }', file=stdout)
-    print('    let key = WORDLIST[key]?;', file=stdout)
-    print('    if !NAMES[key as u16 as usize].eq_ignore_ascii_case(s) {', file=stdout)
+    print('    let key = match WORDLIST[key] {', file=stdout)
+    print('        Some(key) => key,', file=stdout)
+    print('        None => return None,', file=stdout)
+    print('    };', file=stdout)
+    print('    if !eq_ignore_ascii_case(s, NAMES[key as u16 as usize]) {', file=stdout)
     print('        return None;', file=stdout)
     print('    }', file=stdout)
     print(file=stdout)
     print('    Some(key)', file=stdout)
     print('}', file=stdout)
     print(file=stdout)
 
-    print('#[inline]')
-    print('pub(crate) fn find_tz(s: &[u8]) -> Option<TimeZoneRef<\'static>> {', file=stdout)
-    print('    Some(*TIME_ZONES[find_key(s)? as u16 as usize])', file=stdout)
-    print('}', file=stdout)
-    print(file=stdout)
-
-    print('#[inline]')
-    print('pub(crate) fn find_raw(s: &[u8]) -> Option<&\'static [u8]> {', file=stdout)
-    print('    Some(RAW_TIME_ZONES[find_key(s)? as u16 as usize])', file=stdout)
-    print('}', file=stdout)
-
 
 if __name__ == '__main__':
     basicConfig(level=INFO)