use (packed) array instead of typed array

cometkim · Nov 29, 2024 · 94ed937 · 94ed937
1 parent 5094e7a
commit 94ed937
Show file tree

Hide file tree

Showing 8 changed files with 37 additions and 29 deletions.
diff --git a/.changeset/modern-wasps-grin.md b/.changeset/modern-wasps-grin.md
@@ -0,0 +1,10 @@
+---
+"unicode-segmenter": patch
+---
+
+Improved perf and bundle size a bit
+
+It seems using `TypedArray` isn't helpful,
+and deref many prototypes may cause deopt.
+
+`Array` is good enough while it ensures it's packed.
diff --git a/README.md b/README.md
@@ -254,7 +254,7 @@ Since [Hermes doesn't support the `Intl.Segmenter` API](https://github.com/faceb
 
 | Name                         | Unicode® | ESM? |   Size    | Size (min) | Size (min+gzip) | Size (min+br) |
 |------------------------------|----------|------|----------:|-----------:|----------------:|--------------:|
-| `unicode-segmenter/grapheme` |   16.0.0 |    ✔️ |    17,352 |     12,822 |           5,307 |         4,089 |
+| `unicode-segmenter/grapheme` |   16.0.0 |    ✔️ |    17,313 |     12,783 |           5,285 |         3,946 |
 | `graphemer`                  |   15.0.0 |    ✖️ ️|   410,435 |     95,104 |          15,752 |        10,660 |
 | `grapheme-splitter`          |   10.0.0 |    ✖️ |   122,252 |     23,680 |           7,852 |         4,841 |
 | `@formatjs/intl-segmenter`*  |   15.0.0 |    ✖️ |   491,043 |    318,721 |          54,248 |        34,380 |
@@ -270,7 +270,7 @@ Since [Hermes doesn't support the `Intl.Segmenter` API](https://github.com/faceb
 
 | Name                         | Bytecode size | Bytecode size (gzip)* |
 |------------------------------|--------------:|----------------------:|
-| `unicode-segmenter/grapheme` |        24,538 |                12,788 |
+| `unicode-segmenter/grapheme` |        24,386 |                12,690 |
 | `graphemer`                  |       133,949 |                31,710 |
 | `grapheme-splitter`          |        63,810 |                19,125 |
 

diff --git a/scripts/unicode.js b/scripts/unicode.js
@@ -573,19 +573,19 @@ export const ${typeName} = {
 
   f.write(`
 export const ${name}_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(${breakTable.length * 2}),
+  Array(${breakTable.length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${breakTable.map(x => `${x[0] === 0 ? '' : x[0].toString(36)},${x[1] === 0 ? '' : x[1].toString(36)}`).join(',')}')
 );
 
 export const ${name}_cats = initLookupTableBuffer(
-  new Uint8Array(${breakTable.length}),
+  Array(${breakTable.length}),
   /** @type {LookupTableEncoding} */
   ('${breakTable.map(x => inversed[x[2]].toString(36)).join('')}')
 );
 
 const ${name}_lookup = initLookupTableBuffer(
-  new Uint16Array(${lookupTable.length}),
+  Array(${lookupTable.length}),
   /** @type {LookupTableEncoding} */
   ('${lookupTable.map(x => x === 0 ? '' : x.toString(36)).join(',')}'),
   ','
@@ -637,7 +637,7 @@ import { initUnicodeRangeBuffer } from './core.js';
  * The Unicode \`Indic_Conjunct_Break=Consonant\` derived property table
  */
 export const consonant_buffer = initUnicodeRangeBuffer(
-  new Uint16Array(${table.length * 2}),
+  Array(${table.length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${table.map(x => `${x[0] ? x[0].toString(36) : ''},${x[1] ? x[1].toString(36) : ''}`).join(',')}')
 );
@@ -675,7 +675,7 @@ import { initUnicodeRangeBuffer } from './core.js';
  * @type {UnicodeRangeBuffer}
  */
 export const letter_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(${gencats['L'].length * 2}),
+  Array(${gencats['L'].length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${encodeRanges(gencats['L'])}')
 );
@@ -686,7 +686,7 @@ export const letter_buffer = initUnicodeRangeBuffer(
  * @type {UnicodeRangeBuffer}
  */
 export const numeric_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(${gencats['N'].length * 2}),
+  Array(${gencats['N'].length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${encodeRanges(gencats['N'])}')
 );
@@ -697,7 +697,7 @@ export const numeric_buffer = initUnicodeRangeBuffer(
  * @type {UnicodeRangeBuffer}
  */
 export const alphabetic_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(${derived['Alphabetic'].length * 2}),
+  Array(${derived['Alphabetic'].length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${encodeRanges(derived['Alphabetic'])}')
 );
@@ -727,7 +727,7 @@ import { initUnicodeRangeBuffer } from './core.js';
  * @type {UnicodeRangeBuffer}
  */
 export const emoji_presentation_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(${emojiProps['Emoji_Presentation'].length * 2}),
+  Array(${emojiProps['Emoji_Presentation'].length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${encodeRanges(emojiProps['Emoji_Presentation'])}')
 );
@@ -738,7 +738,7 @@ export const emoji_presentation_buffer = initUnicodeRangeBuffer(
  * @type {UnicodeRangeBuffer}
  */
 export const extended_pictographic_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(${emojiProps['Extended_Pictographic'].length * 2}),
+  Array(${emojiProps['Extended_Pictographic'].length * 2}),
   /** @type {UnicodeRangeEncoding} */
   ('${encodeRanges(emojiProps['Extended_Pictographic'])}')
 );

diff --git a/src/_emoji_data.js b/src/_emoji_data.js
@@ -16,7 +16,7 @@ import { initUnicodeRangeBuffer } from './core.js';
  * @type {UnicodeRangeBuffer}
  */
 export const emoji_presentation_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(160),
+  Array(160),
   /** @type {UnicodeRangeEncoding} */
   ('6xm,1,73d,3,73k,,73n,,7i5,1,7is,1,7k8,b,7lr,,7mb,,7mp,,7my,1,7nh,1,7no,1,7ny,,7o4,,7oq,,7oy,1,7p1,,7p6,,7p9,,7ph,,7pm,1,7qg,,7rg,,7ri,,7rn,2,7rr,,7th,2,7u8,,7un,,8ij,1,8k0,,8k5,,2pz8,,2q4v,,2qa6,,2qa9,9,2qcm,p,2qdd,,2qe2,,2qen,,2qeq,4,2qew,2,2qfk,1,2qkg,w,2qlp,8,2qlz,1x,2qny,l,2qow,16,2qq7,4,2qqo,g,2qr8,,2qrc,1y,2qtc,,2qte,56,2qyn,1q,2r0r,3,2r0w,n,2r22,,2r2t,1,2r38,,2r5n,2c,2r9c,1x,2rbg,,2rbk,2,2rbp,2,2rbw,3,2rcb,1,2rck,8,2rj4,b,2rjk,,2rrg,1a,2rss,9,2rt3,54,2s1c,c,2s1s,9,2s27,1j,2s3y,e,2s4f,a,2s4w,8')
 );
@@ -27,7 +27,7 @@ export const emoji_presentation_buffer = initUnicodeRangeBuffer(
  * @type {UnicodeRangeBuffer}
  */
 export const extended_pictographic_buffer = initUnicodeRangeBuffer(
-  new Uint32Array(156),
+  Array(156),
   /** @type {UnicodeRangeEncoding} */
   ('4p,,4u,,6d8,,6dl,,6jm,,6k9,,6ms,5,6nd,1,6xm,1,6y0,,70o,,72n,,73d,a,73s,2,79e,,7fu,1,7g6,,7gg,,7i3,3,7i8,5,7if,b,7is,35,7m8,39,7pk,a,7pw,,7py,,7q5,,7q9,,7qg,,7qr,1,7r8,,7rb,,7rg,,7ri,,7rn,2,7rr,,7s3,4,7th,2,7tt,,7u8,,7un,,850,1,8hx,2,8ij,1,8k0,,8k5,,9io,,9j1,,9zr,,9zt,,2pz4,73,2q6l,2,2q7j,,2q98,5,2q9q,1,2qa6,,2qa9,9,2qb1,1k,2qdd,e,2qe2,,2qen,,2qeq,8,2qf0,3,2qfd,c1,2qrk,8t,2r0m,7d,2r9c,3j,2rg4,b,2rit,16,2rkc,3,2rm0,7,2rmi,5,2rns,7,2rou,29,2rrg,1a,2rss,9,2rt3,c8,2scg,sd')
 );
diff --git a/src/_general_data.js b/src/_general_data.js