Merge pull request #7 from Marcondiro/master

Add CI and fix clippy warnings

.github/workflows/rust.yml

@@ -0,0 +1,75 @@
+name: Rust
+
+on:
+  push:
+    branches: [ master ]
+  pull_request:
+    branches: [ master ]
+
+env:
+  CARGO_INCREMENTAL: 0
+  CARGO_TERM_COLOR: always
+  RUST_BACKTRACE: 1
+  RUSTFLAGS: -D warnings
+  RUSTDOCFLAGS: -D warnings --cfg docsrs
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        rust:
+          - stable
+          - beta
+          - nightly
+    steps:
+    - uses: actions/checkout@v2
+    - name: Install toolchain
+      uses: actions-rs/toolchain@v1
+      with:
+          toolchain: ${{ matrix.rust }}
+          override: true
+          components: rustfmt, clippy
+    - name: Build
+      run: cargo build --verbose
+    - name: Run tests with all features
+      run: cargo test --all-features --verbose
+    - name: Run tests without features
+      run: cargo test --lib --no-default-features --verbose
+    - name: Package
+      run: cargo package
+    - name: Test package
+      run: cd $(find target/package/ -maxdepth 1 -mindepth 1 -type d) && cargo test
+    - name: Test package without features
+      run: cd $(find target/package/ -maxdepth 1 -mindepth 1 -type d) && cargo test --lib --no-default-features
+    - name: Build docs
+      if: matrix.rust == 'nightly'
+      run: cargo doc --all-features --verbose
+    - name: Check formatting
+      if: matrix.rust == 'stable'
+      run: cargo fmt --all --check
+    - name: Check clippy
+      if: matrix.rust == 'stable'
+      run: cargo clippy --all-features --lib --tests --examples --verbose
+    - name: Check benchmarks with clippy
+      if: matrix.rust == 'nightly'
+      run: cargo clippy --all-features --benches --verbose
+  msrv:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - name: Install msrv toolchain
+      uses: actions-rs/toolchain@v1
+      with:
+          toolchain: 1.56
+          override: true
+    - name: Build
+      run: cargo build --verbose --all-features
+  regen:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+    - name: Regen
+      run: cd scripts && python3 unicode.py
+    - name: Diff tables
+      run: diff src/tables.rs scripts/tables.rs

scripts/unicode.py

@@ -275,9 +275,7 @@ def emit_general_category_module(f):
 
     #[inline]
     pub(crate) fn general_category_of_char(c: char) -> GeneralCategory {
-        match c as usize {
-            _ => super::util::bsearch_range_value_table(c, GENERAL_CATEGORY).unwrap_or(GeneralCategory::Unassigned)
-        }
+        super::util::bsearch_range_value_table(c, GENERAL_CATEGORY).unwrap_or(GeneralCategory::Unassigned)
     }
 
     #[inline]
@@ -369,7 +367,7 @@ def emit_general_category_module(f):
                 general_category_char_table[input_idx][1], general_category_group_table[existing_group_count - 1][2])
         else:
             general_category_group_table.append(general_category_char_table[input_idx])
-    emit_table(f, "GENERAL_CATEGORY", general_category_group_table, "&'static [(char, char, GeneralCategory)]", is_pub=False,
+    emit_table(f, "GENERAL_CATEGORY", general_category_group_table, "&[(char, char, GeneralCategory)]", is_pub=False,
             pfun=lambda x: "(%s,%s,%s)" % (escape_char(x[0]), escape_char(x[1]), gc_variants[x[2]]))
     f.write("}\n\n")
 
@@ -405,9 +403,7 @@ def emit_emoji_module(f):
     #[inline]
     pub(crate) fn emoji_status(c: char) -> EmojiStatus {
         // FIXME: do we want to special case ASCII here?
-        match c as usize {
-            _ => super::util::bsearch_range_value_table(c, EMOJI_STATUS).unwrap()
-        }
+        super::util::bsearch_range_value_table(c, EMOJI_STATUS).unwrap()
     }
     #[inline]
     pub(crate) fn is_emoji_status_for_emoji_char_or_emoji_component(s: EmojiStatus) -> bool {
@@ -491,7 +487,7 @@ def group_text(s):
                     emoji_prop_list_pos[prop_list_idx] += 1
         cur_group_first = cur_group_last + 1
 
-    emit_table(f, "EMOJI_STATUS", emoji_table, "&'static [(char, char, EmojiStatus)]", is_pub=False,
+    emit_table(f, "EMOJI_STATUS", emoji_table, "&[(char, char, EmojiStatus)]", is_pub=False,
             pfun=lambda x: "(%s,%s,%s)" % (escape_char(x[0]), escape_char(x[1]), x[2]))
     f.write("}\n\n")
 

src/lib.rs

@@ -21,14 +21,12 @@
 //! use unicode_properties::UnicodeEmoji;
 //! use unicode_properties::UnicodeGeneralCategory;
 //!
-//! fn main() {
-//!     let ch = '🦀'; // U+1F980 CRAB
-//!     let is_emoji = ch.is_emoji_char();
-//!     let group = ch.general_category_group();
-//!     println!("{}({:?})", ch, group);
-//!     println!("The above char {} for use as emoji char.",
-//!              if is_emoji { "is recommended" } else { "is not recommended" });
-//! }
+//! let ch = '🦀'; // U+1F980 CRAB
+//! let is_emoji = ch.is_emoji_char();
+//! let group = ch.general_category_group();
+//! println!("{}({:?})", ch, group);
+//! println!("The above char {} for use as emoji char.",
+//!          if is_emoji { "is recommended" } else { "is not recommended" });
 //! ```
 //!
 //! # Features
@@ -59,17 +57,20 @@ pub mod emoji {
         fn emoji_status(self) -> EmojiStatus;
 
         /// Checks whether this character is recommended for use as emoji, i.e. `Emoji=YES`.
+        #[allow(clippy::wrong_self_convention)]
         fn is_emoji_char(self) -> bool {
             crate::tables::emoji::is_emoji_status_for_emoji_char(self.emoji_status())
         }
 
         /// Checks whether this character are used in emoji sequences where they're not
         /// intended for independent, direct input, i.e. `Emoji_Component=YES`.
+        #[allow(clippy::wrong_self_convention)]
         fn is_emoji_component(self) -> bool {
             crate::tables::emoji::is_emoji_status_for_emoji_component(self.emoji_status())
         }
 
         /// Checks whether this character occurs in emoji sequences, i.e. `Emoji=YES | Emoji_Component=YES`
+        #[allow(clippy::wrong_self_convention)]
         fn is_emoji_char_or_emoji_component(self) -> bool {
             crate::tables::emoji::is_emoji_status_for_emoji_char_or_emoji_component(
                 self.emoji_status(),
@@ -144,6 +145,7 @@ pub mod general_category {
         ///
         /// The `LetterCased` group includes `LetterUppercase`, `LetterLowercase`, and `LetterTitlecase`
         /// categories, and is a subset of the `Letter` group.
+        #[allow(clippy::wrong_self_convention)]
         fn is_letter_cased(self) -> bool {
             crate::tables::general_category::general_category_is_letter_cased(
                 self.general_category(),

src/tables.rs

@@ -127,9 +127,7 @@ pub mod general_category {
 
     #[inline]
     pub(crate) fn general_category_of_char(c: char) -> GeneralCategory {
-        match c as usize {
-            _ => super::util::bsearch_range_value_table(c, GENERAL_CATEGORY).unwrap_or(GeneralCategory::Unassigned)
-        }
+        super::util::bsearch_range_value_table(c, GENERAL_CATEGORY).unwrap_or(GeneralCategory::Unassigned)
     }
 
     #[inline]
@@ -173,7 +171,7 @@ pub mod general_category {
         }
     }
     // General category table:
-    const GENERAL_CATEGORY: &'static [(char, char, GeneralCategory)] = &[
+    const GENERAL_CATEGORY: &[(char, char, GeneralCategory)] = &[
         ('\u{0}', '\u{1f}', GeneralCategory::Control), ('\u{20}', '\u{20}',
         GeneralCategory::SpaceSeparator), ('\u{21}', '\u{23}', GeneralCategory::OtherPunctuation),
         ('\u{24}', '\u{24}', GeneralCategory::CurrencySymbol), ('\u{25}', '\u{27}',
@@ -2743,9 +2741,7 @@ pub mod emoji {
     #[inline]
     pub(crate) fn emoji_status(c: char) -> EmojiStatus {
         // FIXME: do we want to special case ASCII here?
-        match c as usize {
-            _ => super::util::bsearch_range_value_table(c, EMOJI_STATUS).unwrap()
-        }
+        super::util::bsearch_range_value_table(c, EMOJI_STATUS).unwrap()
     }
     #[inline]
     pub(crate) fn is_emoji_status_for_emoji_char_or_emoji_component(s: EmojiStatus) -> bool {
@@ -2762,7 +2758,7 @@ pub mod emoji {
             EmojiStatus::EmojiOtherAndEmojiComponent)
     }
     // Emoji status table:
-    const EMOJI_STATUS: &'static [(char, char, EmojiStatus)] = &[
+    const EMOJI_STATUS: &[(char, char, EmojiStatus)] = &[
         ('\u{0}', '\u{22}', EmojiStatus::NonEmoji), ('\u{23}', '\u{23}',
         EmojiStatus::EmojiOtherAndEmojiComponent), ('\u{24}', '\u{29}', EmojiStatus::NonEmoji),
         ('\u{2a}', '\u{2a}', EmojiStatus::EmojiOtherAndEmojiComponent), ('\u{2b}', '\u{2f}',