Merge pull request #9 from zbraniecki/fix_alphanumeric

Test for numbers in is_ascii_alphanumeric
zbraniecki · Aug 23, 2019 · 9b1e294 · 9b1e294
2 parents 5046730 + 72bbca7
commit 9b1e294
Show file tree

Hide file tree

Showing 4 changed files with 111 additions and 6 deletions.
diff --git a/src/tinystr16.rs b/src/tinystr16.rs
@@ -43,15 +43,35 @@ impl TinyStr16 {
         unsafe { Self(NonZeroU128::new_unchecked(result)) }
     }
 
+    pub fn is_ascii_alphabetic(self) -> bool {
+        let word = self.0.get();
+        let mask =
+            (word + 0x7f7f7f7f_7f7f7f7f_7f7f7f7f_7f7f7f7f) & 0x80808080_80808080_80808080_80808080;
+        let lower = word | 0x20202020_20202020_20202020_20202020;
+        let alpha = !(lower + 0x1f1f1f1f_1f1f1f1f_1f1f1f1f_1f1f1f1f)
+            | (lower + 0x05050505_05050505_05050505_05050505);
+        (alpha & mask) == 0
+    }
+
     pub fn is_ascii_alphanumeric(self) -> bool {
         let word = self.0.get();
         let mask =
             (word + 0x7f7f7f7f_7f7f7f7f_7f7f7f7f_7f7f7f7f) & 0x80808080_80808080_80808080_80808080;
+        let numeric = !(word + 0x50505050_50505050_50505050_50505050)
+            | (word + 0x46464646_46464646_46464646_46464646);
         let lower = word | 0x20202020_20202020_20202020_20202020;
-        ((!(lower + 0x1f1f1f1f_1f1f1f1f_1f1f1f1f_1f1f1f1f)
-            | (lower + 0x05050505_05050505_05050505_05050505))
-            & mask)
-            == 0
+        let alpha = !(lower + 0x1f1f1f1f_1f1f1f1f_1f1f1f1f_1f1f1f1f)
+            | (lower + 0x05050505_05050505_05050505_05050505);
+        (alpha & numeric & mask) == 0
+    }
+
+    pub fn is_ascii_numeric(self) -> bool {
+        let word = self.0.get();
+        let mask =
+            (word + 0x7f7f7f7f_7f7f7f7f_7f7f7f7f_7f7f7f7f) & 0x80808080_80808080_80808080_80808080;
+        let numeric = !(word + 0x50505050_50505050_50505050_50505050)
+            | (word + 0x46464646_46464646_46464646_46464646);
+        (numeric & mask) == 0
     }
 
     pub fn to_ascii_titlecase(self) -> Self {

diff --git a/src/tinystr4.rs b/src/tinystr4.rs
@@ -35,11 +35,28 @@ impl TinyStr4 {
         unsafe { Self(NonZeroU32::new_unchecked(result)) }
     }
 
+    pub fn is_ascii_alphabetic(self) -> bool {
+        let word = self.0.get();
+        let mask = (word + 0x7f7f_7f7f) & 0x8080_8080;
+        let lower = word | 0x2020_2020;
+        let alpha = !(lower + 0x1f1f_1f1f) | (lower + 0x0505_0505);
+        (alpha & mask) == 0
+    }
+
     pub fn is_ascii_alphanumeric(self) -> bool {
         let word = self.0.get();
         let mask = (word + 0x7f7f_7f7f) & 0x8080_8080;
+        let numeric = !(word + 0x5050_5050) | (word + 0x4646_4646);
         let lower = word | 0x2020_2020;
-        ((!(lower + 0x1f1f_1f1f) | (lower + 0x0505_0505)) & mask) == 0
+        let alpha = !(lower + 0x1f1f_1f1f) | (lower + 0x0505_0505);
+        (alpha & numeric & mask) == 0
+    }
+
+    pub fn is_ascii_numeric(self) -> bool {
+        let word = self.0.get();
+        let mask = (word + 0x7f7f_7f7f) & 0x8080_8080;
+        let numeric = !(word + 0x5050_5050) | (word + 0x4646_4646);
+        (numeric & mask) == 0
     }
 
     /// Makes the string all lowercase except for the first character,

diff --git a/src/tinystr8.rs b/src/tinystr8.rs
@@ -43,11 +43,28 @@ impl TinyStr8 {
         unsafe { Self(NonZeroU64::new_unchecked(result)) }
     }
 
+    pub fn is_ascii_alphabetic(self) -> bool {
+        let word = self.0.get();
+        let mask = (word + 0x7f7f7f7f_7f7f7f7f) & 0x80808080_80808080;
+        let lower = word | 0x20202020_20202020;
+        let alpha = !(lower + 0x1f1f1f1f_1f1f1f1f) | (lower + 0x05050505_05050505);
+        (alpha & mask) == 0
+    }
+
     pub fn is_ascii_alphanumeric(self) -> bool {
         let word = self.0.get();
         let mask = (word + 0x7f7f7f7f_7f7f7f7f) & 0x80808080_80808080;
+        let numeric = !(word + 0x50505050_50505050) | (word + 0x46464646_46464646);
         let lower = word | 0x20202020_20202020;
-        ((!(lower + 0x1f1f1f1f_1f1f1f1f) | (lower + 0x05050505_05050505)) & mask) == 0
+        let alpha = !(lower + 0x1f1f1f1f_1f1f1f1f) | (lower + 0x05050505_05050505);
+        (alpha & numeric & mask) == 0
+    }
+
+    pub fn is_ascii_numeric(self) -> bool {
+        let word = self.0.get();
+        let mask = (word + 0x7f7f7f7f_7f7f7f7f) & 0x80808080_80808080;
+        let numeric = !(word + 0x50505050_50505050) | (word + 0x46464646_46464646);
+        (numeric & mask) == 0
     }
 
     pub fn to_ascii_titlecase(self) -> Self {

diff --git a/tests/main.rs b/tests/main.rs
@@ -41,11 +41,23 @@ fn tiny4_nonascii() {
 #[test]
 fn tiny4_alpha() {
     let s: TinyStr4 = "@aZ[".parse().unwrap();
+    assert!(!s.is_ascii_alphabetic());
     assert!(!s.is_ascii_alphanumeric());
     assert_eq!(s.to_ascii_uppercase().as_str(), "@AZ[");
     assert_eq!(s.to_ascii_lowercase().as_str(), "@az[");
 
+    assert!("abYZ".parse::<TinyStr4>().unwrap().is_ascii_alphabetic());
     assert!("abYZ".parse::<TinyStr4>().unwrap().is_ascii_alphanumeric());
+    assert!("a123".parse::<TinyStr4>().unwrap().is_ascii_alphanumeric());
+    assert!(!"a123".parse::<TinyStr4>().unwrap().is_ascii_alphabetic());
+}
+
+#[test]
+fn tiny4_numeric() {
+    let s: TinyStr4 = "@aZ[".parse().unwrap();
+    assert!(!s.is_ascii_numeric());
+
+    assert!("0123".parse::<TinyStr4>().unwrap().is_ascii_numeric());
 }
 
 #[test]
@@ -168,14 +180,29 @@ fn tiny8_nonascii() {
 #[test]
 fn tiny8_alpha() {
     let s: TinyStr8 = "@abcXYZ[".parse().unwrap();
+    assert!(!s.is_ascii_alphabetic());
     assert!(!s.is_ascii_alphanumeric());
     assert_eq!(s.to_ascii_uppercase().as_str(), "@ABCXYZ[");
     assert_eq!(s.to_ascii_lowercase().as_str(), "@abcxyz[");
 
+    assert!("abcXYZ".parse::<TinyStr8>().unwrap().is_ascii_alphabetic());
     assert!("abcXYZ"
         .parse::<TinyStr8>()
         .unwrap()
         .is_ascii_alphanumeric());
+    assert!(!"abc123".parse::<TinyStr8>().unwrap().is_ascii_alphabetic());
+    assert!("abc123"
+        .parse::<TinyStr8>()
+        .unwrap()
+        .is_ascii_alphanumeric());
+}
+
+#[test]
+fn tiny8_numeric() {
+    let s: TinyStr8 = "@abcXYZ[".parse().unwrap();
+    assert!(!s.is_ascii_numeric());
+
+    assert!("01234567".parse::<TinyStr8>().unwrap().is_ascii_numeric());
 }
 
 #[test]
@@ -303,16 +330,40 @@ fn tiny16_nonascii() {
 #[test]
 fn tiny16_alpha() {
     let s: TinyStr16 = "@abcdefgTUVWXYZ[".parse().unwrap();
+    assert!(!s.is_ascii_alphabetic());
     assert!(!s.is_ascii_alphanumeric());
     assert_eq!(s.to_ascii_uppercase().as_str(), "@ABCDEFGTUVWXYZ[");
     assert_eq!(s.to_ascii_lowercase().as_str(), "@abcdefgtuvwxyz[");
 
+    assert!("abcdefgTUVWXYZ"
+        .parse::<TinyStr16>()
+        .unwrap()
+        .is_ascii_alphabetic());
+    assert!("abcdefgTUVWXYZ"
+        .parse::<TinyStr16>()
+        .unwrap()
+        .is_ascii_alphanumeric());
+    assert!(!"abcdefg0123456"
+        .parse::<TinyStr16>()
+        .unwrap()
+        .is_ascii_alphabetic());
     assert!("abcdefgTUVWXYZ"
         .parse::<TinyStr16>()
         .unwrap()
         .is_ascii_alphanumeric());
 }
 
+#[test]
+fn tiny16_numeric() {
+    let s: TinyStr16 = "@abcdefgTUVWXYZ[".parse().unwrap();
+    assert!(!s.is_ascii_numeric());
+
+    assert!("0123456789"
+        .parse::<TinyStr16>()
+        .unwrap()
+        .is_ascii_numeric());
+}
+
 #[test]
 fn tiny16_titlecase() {
     assert_eq!(