feat(mf2): Allow number-literal as unquoted value

eemeli · eemeli · commit 7ae0e1057e91 · 2024-01-01T13:37:43.000+02:00
diff --git a/packages/mf2-messageformat/src/__fixtures/test-messages.json b/packages/mf2-messageformat/src/__fixtures/test-messages.json
@@ -9,8 +9,8 @@
     "exp": "hello world"
   },
   {
-    "src": "hello {$place:-.}",
-    "params": { "place:-.": "world" },
+    "src": "hello {$place-.}",
+    "params": { "place-.": "world" },
     "exp": "hello world"
   },
   {
@@ -29,7 +29,9 @@
     "params": { "one": 1.3, "two": 4.2 },
     "exp": "1,3 et 4,2"
   },
-  { "src": "hello {|4.2| :number}", "exp": "hello 4.2" },
+  { "src": "hello {4.2 :number}", "exp": "hello 4.2" },
+  { "src": "hello {-4.20 :number}", "exp": "hello -4.2" },
+  { "src": "hello {0.42e+1 :number}", "exp": "hello 4.2" },
   {
     "src": "hello {foo :number}",
     "exp": "hello {|foo|}",
@@ -41,15 +43,15 @@
     "errors": [{ "type": "bad-input" }]
   },
   {
-    "src": "hello {|4.2| :number minimumFractionDigits=2}",
+    "src": "hello {4.2 :number minimumFractionDigits=2}",
     "exp": "hello 4.20"
   },
   {
     "src": "hello {|4.2| :number minimumFractionDigits=|2|}",
     "exp": "hello 4.20"
   },
   {
-    "src": "hello {|4.2| :number minimumFractionDigits=$foo}",
+    "src": "hello {4.2 :number minimumFractionDigits=$foo}",
     "params": { "foo": 2 },
     "exp": "hello 4.20"
   },
diff --git a/packages/mf2-messageformat/src/cst-parser/index.ts b/packages/mf2-messageformat/src/cst-parser/index.ts
@@ -2,4 +2,3 @@ import type * as CST from './cst-types.js';
 export { CST };
 export { asDataModel } from './as-data-model.js';
 export { parseMessage } from './message.js';
-export { isValidUnquotedLiteral } from './names.js';
diff --git a/packages/mf2-messageformat/src/cst-parser/names.ts b/packages/mf2-messageformat/src/cst-parser/names.ts
@@ -22,48 +22,53 @@ const isNameStartCode = (cc: number) =>
   (cc >= 0xfdf0 && cc <= 0xfffd) ||
   (cc >= 0x10000 && cc <= 0xeffff);
 
-// unquoted-start = name-start / DIGIT / "."
-//                / %xB7 / %x300-36F / %x203F-2040
-const isUnquotedStartCharCode = (cc: number) =>
+// name-char = name-start / DIGIT / "-" / "."
+//           / %xB7 / %x300-36F / %x203F-2040
+const isNameCharCode = (cc: number) =>
   isNameStartCode(cc) ||
+  cc === 0x2d || // -
   cc === 0x2e || // .
   (cc >= 0x30 && cc <= 0x39) || // 0-9
   cc === 0xb7 || // ·
   (cc >= 0x300 && cc <= 0x36f) ||
   cc === 0x203f || // ‿
   cc === 0x2040; // ⁀
 
-// name-char = name-start / DIGIT / "-" / "." / ":"
-//           / %xB7 / %x300-36F / %x203F-2040
-const isNameCharCode = (cc: number) =>
-  isUnquotedStartCharCode(cc) || cc === 0x2d || cc === 0x3a; // - :
+// This is sticky so that parsing doesn't need to substring the source
+const numberLiteral = /-?(?:0|[1-9]\d*)(?:\.\d+)?(?:[eE][-+]?\d+)?/y;
+
+export function parseNameValue(src: string, start: number): string {
+  if (!isNameStartCode(src.charCodeAt(start))) return '';
+  let pos = start + 1;
+  while (isNameCharCode(src.charCodeAt(pos))) pos += 1;
+  return src.substring(start, pos);
+}
 
 export function isValidUnquotedLiteral(str: string): boolean {
-  if (!isUnquotedStartCharCode(str.charCodeAt(0))) return false;
+  numberLiteral.lastIndex = 0;
+  const num = numberLiteral.exec(str);
+  if (num && num[0].length === str.length) return true;
+
+  if (!isNameStartCode(str.charCodeAt(0))) return false;
   for (let i = 1; i < str.length; ++i) {
     const cc = str.charCodeAt(i);
     if (!isNameCharCode(cc)) return false;
   }
   return str.length > 0;
 }
 
-// name = name-start *name-char
-export function parseNameValue(src: string, start: number): string {
-  if (!isNameStartCode(src.charCodeAt(start))) return '';
-  let pos = start + 1;
-  while (isNameCharCode(src.charCodeAt(pos))) pos += 1;
-  return src.substring(start, pos);
-}
-
-// unquoted = unquoted-start *name-char
 export function parseUnquotedLiteralValue(
-  ctx: ParseContext,
+  { source }: ParseContext,
   start: number
 ): string {
+  numberLiteral.lastIndex = start;
+  const num = numberLiteral.exec(source);
+  if (num) return num[0];
+
   let pos = start;
-  if (isUnquotedStartCharCode(ctx.source.charCodeAt(pos))) {
+  if (isNameStartCode(source.charCodeAt(pos))) {
     pos += 1;
-    while (isNameCharCode(ctx.source.charCodeAt(pos))) pos += 1;
+    while (isNameCharCode(source.charCodeAt(pos))) pos += 1;
   }
-  return ctx.source.substring(start, pos);
+  return source.substring(start, pos);
 }
diff --git a/packages/mf2-messageformat/src/cst-parser/values.ts b/packages/mf2-messageformat/src/cst-parser/values.ts
@@ -70,13 +70,6 @@ export function parseLiteral(
   return { type: 'literal', quoted: false, start, end, value };
 }
 
-// quoted      = "|" *(quoted-char / quoted-escape) "|"
-// quoted-char = %x0-5B         ; omit \
-//             / %x5D-7B        ; omit |
-//             / %x7D-D7FF      ; omit surrogates
-//             / %xE000-10FFFF
-// quoted-escape   = backslash ( backslash / "|" )
-// backslash       = %x5C ; U+005C REVERSE SOLIDUS "\"
 export function parseQuotedLiteral(
   ctx: ParseContext,
   start: number
diff --git a/packages/mf2-messageformat/src/expression/literal.test.ts b/packages/mf2-messageformat/src/expression/literal.test.ts
@@ -30,3 +30,21 @@ describe('quoted literals', () => {
     expect(() => new MessageFormat('{|quoted \\{iteral|}')).toThrow();
   });
 });
+
+describe('unquoted numbers', () => {
+  for (const value of [
+    '0',
+    '42',
+    '2.5',
+    '-1',
+    '-0.999',
+    '1e3',
+    '0.4E+5',
+    '11.1e-1'
+  ]) {
+    test(value, () => {
+      const res = resolve(`{${value}}`);
+      expect(res).toMatchObject([{ type: 'string', value }]);
+    });
+  }
+});
diff --git a/packages/mf2-messageformat/src/stringifier/message.ts b/packages/mf2-messageformat/src/stringifier/message.ts
@@ -1,4 +1,4 @@
-import { isValidUnquotedLiteral } from '../cst-parser/index.js';
+import { isValidUnquotedLiteral } from '../cst-parser/names.js';
 import {
   Declaration,
   Expression,

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-import { isValidUnquotedLiteral } from '../cst-parser/index.js';`
	`1`	`+import { isValidUnquotedLiteral } from '../cst-parser/names.js';`
`2`	`2`	`import {`
`3`	`3`	`Declaration,`
`4`	`4`	`Expression,`