Merge branch 'main' into gabriel/mnc-mcc

rubdos · web-flow · commit 4b07e85a7585 · 2023-12-14T15:20:31.000+01:00
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
@@ -24,9 +24,14 @@ jobs:
       matrix:
         toolchain: ["stable", "beta"]
         coverage: [false]
+        tests: [true]
         include:
           - toolchain: "nightly"
             coverage: true
+            tests: true
+          - toolchain: "1.58.0"
+            coverage: false
+            tests: false
     steps:
       - name: Checkout repository
         uses: actions/checkout@v2
@@ -39,22 +44,30 @@ jobs:
       - name: Configure CI cache
         uses: Swatinem/rust-cache@v2
 
-      - name: Build
+      - name: Build all targets
         uses: actions-rs/cargo@v1
+        if: ${{ matrix.tests }}
         with:
           command: build
           args: --all-targets
 
+      - name: Build
+        uses: actions-rs/cargo@v1
+        if: ${{ !matrix.tests }}
+        with:
+          command: build
+          args: --lib --all-features
+
       - name: Run tests
         uses: actions-rs/cargo@v1
-        if: ${{ !matrix.coverage }}
+        if: ${{ !matrix.coverage && matrix.tests }}
         with:
           command: test
           args: --all-targets --no-fail-fast
 
       - name: Run tests
         uses: actions-rs/cargo@v1
-        if: ${{ matrix.coverage }}
+        if: ${{ matrix.coverage && matrix.tests }}
         with:
           command: test
           args: --all-targets --no-fail-fast
diff --git a/Cargo.toml b/Cargo.toml
@@ -16,27 +16,32 @@ readme = "README.md"
 bincode = "1.3"
 either = "1.8"
 fnv = "1.0"
-itertools = ">=0.10, <=0.11"
 lazy_static = "1.4"
+itertools = ">=0.10, <= 0.12"
 nom = "7.1"
-quick-xml = "0.28"
+quick-xml = ">=0.28, <= 0.31"
 regex = "1.7"
 regex-cache = "0.2"
 serde = "1.0"
 serde_derive = "1.0"
-strum = { version = "0.24", features = ["derive"] }
+strum = { version = ">=0.24, <=0.25", features = ["derive"] }
 thiserror = "1.0"
 
 [build-dependencies]
 bincode = "1.3"
-quick-xml = "0.28"
+quick-xml = ">=0.28, <=0.31"
 regex = "1.7"
 serde = "1.0"
 serde_derive = "1.0"
 thiserror = "1.0"
 
 [dev-dependencies]
 doc-comment = "0.3"
-rstest = "0.17"
-rstest_reuse = "0.5"
+rstest = ">= 0.13, <=0.18"
+rstest_reuse = "0.6"
 anyhow = "1"
+criterion = ">=0.4, <=0.5"
+
+[[bench]]
+name = "parsing"
+harness = false
diff --git a/README.md b/README.md
@@ -6,6 +6,7 @@
 [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
 
 Rust version of [libphonenumber](https://github.com/googlei18n/libphonenumber).
+We currently require 1.58.0 as minimum supported Rust version (MSRV).
 
 ## Usage
 
@@ -20,8 +21,6 @@ phonenumber = "0.3"
 The following example parses, validates and formats the given phone number.
 
 ```rust,no_run
-extern crate phonenumber;
-
 use phonenumber::Mode;
 use std::env;
 
diff --git a/benches/parsing.rs b/benches/parsing.rs
@@ -0,0 +1,29 @@
+use criterion::{black_box, criterion_group, criterion_main, BenchmarkId, Criterion};
+
+pub fn criterion_benchmark(c: &mut Criterion) {
+    let cases = [
+        "+80012340000",
+        "+61406823897",
+        "+611900123456",
+        "+32474091150",
+        "+34666777888",
+        "+34612345678",
+        "+441212345678",
+        "+13459492311",
+        "+16137827274",
+        "+1 520 878 2491",
+        "+1-520-878-2491",
+    ];
+
+    for case in cases {
+        c.bench_with_input(BenchmarkId::new("parse", case), &case, |b, case| {
+            b.iter(|| {
+                let pn = black_box(case);
+                phonenumber::parse(None, pn)
+            })
+        });
+    }
+}
+
+criterion_group!(benches, criterion_benchmark);
+criterion_main!(benches);
diff --git a/build.rs b/build.rs
@@ -3,15 +3,6 @@ use std::fs::File;
 use std::io::{BufReader, BufWriter};
 use std::path::Path;
 
-extern crate quick_xml as xml;
-extern crate regex;
-extern crate thiserror;
-
-extern crate serde;
-#[macro_use]
-extern crate serde_derive;
-extern crate bincode;
-
 use bincode::Options;
 
 #[path = "src/metadata/loader.rs"]
@@ -21,10 +12,12 @@ mod loader;
 mod error;
 
 fn main() {
+    let pnm_path = "assets/PhoneNumberMetadata.xml";
     let metadata = loader::load(BufReader::new(
-        File::open("assets/PhoneNumberMetadata.xml").expect("could not open metadata file"),
+        File::open(pnm_path).expect("could not open metadata file"),
     ))
     .expect("failed to load metadata");
+    println!("cargo:rerun-if-changed={pnm_path}");
 
     let mut out = BufWriter::new(
         File::create(Path::new(&env::var("OUT_DIR").unwrap()).join("database.bin"))
diff --git a/examples/test.rs b/examples/test.rs
@@ -1,6 +1,5 @@
 use std::env;
 
-extern crate phonenumber;
 use phonenumber::Mode;
 
 fn main() {
diff --git a/src/carrier.rs b/src/carrier.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use serde_derive::{Deserialize, Serialize};
 use std::fmt;
 
 use crate::ParseError;
diff --git a/src/consts.rs b/src/consts.rs
@@ -47,7 +47,7 @@ pub const RFC3966_ISDN_SUBADDRESS: &str = ";isub=";
 
 pub const REGION_CODE_FOR_NON_GEO_ENTITY: &str = "001";
 
-lazy_static! {
+lazy_static::lazy_static! {
     /// Map of country calling codes that use a mobile token before the area code. One example of when
     /// this is relevant is when determining the length of the national destination code, which should
     /// be the length of the area code plus the length of the mobile token.
diff --git a/src/country.rs b/src/country.rs
@@ -14,9 +14,9 @@
 
 //! Country related types.
 
-use strum::{AsRefStr, EnumString};
-
+use serde_derive::{Deserialize, Serialize};
 use std::str;
+use strum::{AsRefStr, EnumString};
 
 #[derive(Copy, Clone, Eq, PartialEq, Serialize, Deserialize, Hash, Debug)]
 pub struct Code {
diff --git a/src/error.rs b/src/error.rs
@@ -95,7 +95,7 @@ pub enum Parse {
 pub enum LoadMetadata {
     /// Parsing XML failed, the XML is malformed.
     #[error("Malformed Metadata XML: {0}")]
-    Xml(#[from] xml::Error),
+    Xml(#[from] quick_xml::Error),
 
     /// Parsing UTF-8 string from XML failed.
     #[error("Non UTF-8 string in Metadata XML: {0}")]
diff --git a/src/extension.rs b/src/extension.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use serde_derive::{Deserialize, Serialize};
 use std::fmt;
 use std::ops::Deref;
 
diff --git a/src/lib.rs b/src/lib.rs
@@ -14,27 +14,8 @@
 
 #![recursion_limit = "1024"]
 
-#[macro_use]
-extern crate lazy_static;
-
-extern crate nom;
-extern crate thiserror;
-
-extern crate either;
-extern crate fnv;
-extern crate itertools;
-extern crate quick_xml as xml;
-extern crate regex;
-extern crate regex_cache;
-
-extern crate serde;
-#[macro_use]
-extern crate serde_derive;
-extern crate bincode;
-
 #[cfg(test)]
-#[macro_use]
-extern crate doc_comment;
+use doc_comment::doctest;
 
 #[cfg(test)]
 use rstest_reuse;
diff --git a/src/metadata/database.rs b/src/metadata/database.rs
@@ -29,7 +29,7 @@ use crate::metadata::loader;
 
 const DATABASE: &[u8] = include_bytes!(concat!(env!("OUT_DIR"), "/database.bin"));
 
-lazy_static! {
+lazy_static::lazy_static! {
     /// The Google provided metadata database, used as default.
     pub static ref DEFAULT: Database =
         Database::from(bincode::options()
diff --git a/src/metadata/loader.rs b/src/metadata/loader.rs
@@ -12,15 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use crate::error;
+use quick_xml::events::attributes::Attribute;
+use quick_xml::events::{self, Event};
+use quick_xml::Reader;
+use serde_derive::{Deserialize, Serialize};
 use std::io::BufRead;
 use std::str;
 
-use crate::xml::events::attributes::Attribute;
-use crate::xml::events::{self, Event};
-use crate::xml::Reader;
-
-use crate::error;
-
 /// Temporary defaults for `Format` and `Descriptor`.
 #[derive(Clone, Default, Serialize, Deserialize, Debug)]
 pub struct Defaults {
@@ -176,7 +175,7 @@ fn territory<R: BufRead>(
     let mut meta = Metadata::default();
 
     for attr in e.attributes() {
-        let Attribute { key, value } = attr.map_err(xml::Error::InvalidAttr)?;
+        let Attribute { key, value } = attr.map_err(quick_xml::Error::InvalidAttr)?;
 
         match (str::from_utf8(key.into_inner())?, str::from_utf8(&value)?) {
             ("id", value) => meta.id = Some(value.into()),
@@ -361,7 +360,8 @@ fn descriptor<R: BufRead>(
             Event::Empty(ref e) => match e.name().into_inner() {
                 b"possibleLengths" => {
                     for attr in e.attributes() {
-                        let Attribute { key, value } = attr.map_err(xml::Error::InvalidAttr)?;
+                        let Attribute { key, value } =
+                            attr.map_err(quick_xml::Error::InvalidAttr)?;
 
                         match (str::from_utf8(key.into_inner())?, str::from_utf8(&value)?) {
                             ("national", value) => descriptor.possible_length = lengths(value)?,
@@ -482,7 +482,7 @@ fn format<R: BufRead>(
     let mut international = None;
 
     for attr in e.attributes() {
-        let Attribute { key, value } = attr.map_err(xml::Error::InvalidAttr)?;
+        let Attribute { key, value } = attr.map_err(quick_xml::Error::InvalidAttr)?;
 
         match (str::from_utf8(key.into_inner())?, str::from_utf8(&value)?) {
             ("pattern", value) => format.pattern = Some(value.into()),
diff --git a/src/national_number.rs b/src/national_number.rs
@@ -12,13 +12,32 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use serde_derive::{Deserialize, Serialize};
 use std::fmt;
 
 /// The national number part of a phone number.
 #[derive(Copy, Clone, Eq, PartialEq, Serialize, Deserialize, Hash, Debug)]
 pub struct NationalNumber {
     pub(crate) value: u64,
+}
+
+impl NationalNumber {
+    pub fn new(value: u64, zeros: u8) -> Self {
+        // E.164 specifies a maximum of 15 decimals, which corresponds to slightly over 48.9 bits.
+        // 56 bits ought to cut it here.
+        assert!(value < (1 << 56), "number too long");
+        Self {
+            value: ((zeros as u64) << 56) | value,
+        }
+    }
 
+    /// The number without any leading zeroes.
+    pub fn value(&self) -> u64 {
+        self.value & 0x00ffffffffffffff
+    }
+
+    /// The number of leading zeroes.
+    ///
     /// In some countries, the national (significant) number starts with one or
     /// more "0"s without this being a national prefix or trunk code of some
     /// kind.  For example, the leading zero in the national (significant) number
@@ -35,18 +54,8 @@ pub struct NationalNumber {
     ///
     /// Clients who use the parsing or conversion functionality of the i18n phone
     /// number libraries will have these fields set if necessary automatically.
-    pub(crate) zeros: u8,
-}
-
-impl NationalNumber {
-    /// The number without any leading zeroes.
-    pub fn value(&self) -> u64 {
-        self.value
-    }
-
-    /// The number of leading zeroes.
     pub fn zeros(&self) -> u8 {
-        self.zeros
+        (self.value >> 56) as u8
     }
 }
 
@@ -58,10 +67,10 @@ impl From<NationalNumber> for u64 {
 
 impl fmt::Display for NationalNumber {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        for _ in 0..self.zeros {
+        for _ in 0..self.zeros() {
             write!(f, "0")?;
         }
 
-        write!(f, "{}", self.value)
+        write!(f, "{}", self.value())
     }
 }
diff --git a/src/parser/mod.rs b/src/parser/mod.rs
diff --git a/src/parser/rfc3966.rs b/src/parser/rfc3966.rs
diff --git a/src/phone_number.rs b/src/phone_number.rs
