polytypo 1.3.0 → 1.4.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/chunk-25RYOB5K.js +5828 -0
- package/dist/chunk-25RYOB5K.js.map +1 -0
- package/dist/{chunk-W3XMTX7Q.js → chunk-47QVJOT4.js} +2 -2
- package/dist/{chunk-M7YWQ47K.cjs → chunk-6PQMYL2O.cjs} +3 -3
- package/dist/{chunk-M7YWQ47K.cjs.map → chunk-6PQMYL2O.cjs.map} +1 -1
- package/dist/{chunk-NT2RADO3.js → chunk-BGTMX6VS.js} +2 -2
- package/dist/{chunk-B3I7VL5E.js → chunk-D3QQYGKW.js} +2 -2
- package/dist/{chunk-7S5PZ7T5.cjs → chunk-DCC5OQET.cjs} +16 -16
- package/dist/{chunk-7S5PZ7T5.cjs.map → chunk-DCC5OQET.cjs.map} +1 -1
- package/dist/{chunk-KRY73ZPE.cjs → chunk-FZ6YRCGR.cjs} +17 -17
- package/dist/{chunk-KRY73ZPE.cjs.map → chunk-FZ6YRCGR.cjs.map} +1 -1
- package/dist/{chunk-72GQS2N7.cjs → chunk-GJJNGBRH.cjs} +4 -4
- package/dist/{chunk-72GQS2N7.cjs.map → chunk-GJJNGBRH.cjs.map} +1 -1
- package/dist/{chunk-DAUTLRFG.js → chunk-GPOCFMTF.js} +4 -4
- package/dist/{chunk-PNX4XW5I.js → chunk-HTWCCPWI.js} +4 -4
- package/dist/chunk-JYHDIVON.cjs +33 -0
- package/dist/{chunk-QLV5EWGF.cjs.map → chunk-JYHDIVON.cjs.map} +1 -1
- package/dist/{chunk-A3VCFDVL.cjs → chunk-MOXBMEZB.cjs} +13 -13
- package/dist/{chunk-A3VCFDVL.cjs.map → chunk-MOXBMEZB.cjs.map} +1 -1
- package/dist/{chunk-PLWW5KPC.js → chunk-NLMP3JVF.js} +2 -2
- package/dist/chunk-PJ6HB77G.cjs +5828 -0
- package/dist/chunk-PJ6HB77G.cjs.map +1 -0
- package/dist/chunk-QLIPSKVZ.cjs +31 -0
- package/dist/{chunk-442V6TKI.cjs.map → chunk-QLIPSKVZ.cjs.map} +1 -1
- package/dist/{chunk-7ZJ7YYTQ.js → chunk-UL7KCRXN.js} +3 -3
- package/dist/{errors-CGFOVO3h.d.cts → errors-DMxY1LlK.d.cts} +20 -0
- package/dist/{errors-CGFOVO3h.d.ts → errors-DMxY1LlK.d.ts} +20 -0
- package/dist/html.cjs +10 -10
- package/dist/html.d.cts +2 -2
- package/dist/html.d.ts +2 -2
- package/dist/html.js +5 -5
- package/dist/index.cjs +17 -17
- package/dist/index.d.cts +2 -2
- package/dist/index.d.ts +2 -2
- package/dist/index.js +7 -7
- package/dist/markdown.cjs +10 -10
- package/dist/markdown.d.cts +2 -2
- package/dist/markdown.d.ts +2 -2
- package/dist/markdown.js +5 -5
- package/dist/text.cjs +8 -8
- package/dist/text.d.cts +2 -2
- package/dist/text.d.ts +2 -2
- package/dist/text.js +3 -3
- package/dist/yaml.cjs +9 -9
- package/dist/yaml.d.cts +2 -2
- package/dist/yaml.d.ts +2 -2
- package/dist/yaml.js +4 -4
- package/package.json +1 -1
- package/dist/chunk-442V6TKI.cjs +0 -31
- package/dist/chunk-GLQQOA6T.js +0 -5778
- package/dist/chunk-GLQQOA6T.js.map +0 -1
- package/dist/chunk-QLV5EWGF.cjs +0 -33
- package/dist/chunk-YX6QTWOV.cjs +0 -5778
- package/dist/chunk-YX6QTWOV.cjs.map +0 -1
- /package/dist/{chunk-W3XMTX7Q.js.map → chunk-47QVJOT4.js.map} +0 -0
- /package/dist/{chunk-NT2RADO3.js.map → chunk-BGTMX6VS.js.map} +0 -0
- /package/dist/{chunk-B3I7VL5E.js.map → chunk-D3QQYGKW.js.map} +0 -0
- /package/dist/{chunk-DAUTLRFG.js.map → chunk-GPOCFMTF.js.map} +0 -0
- /package/dist/{chunk-PNX4XW5I.js.map → chunk-HTWCCPWI.js.map} +0 -0
- /package/dist/{chunk-PLWW5KPC.js.map → chunk-NLMP3JVF.js.map} +0 -0
- /package/dist/{chunk-7ZJ7YYTQ.js.map → chunk-UL7KCRXN.js.map} +0 -0
|
@@ -6,7 +6,7 @@ import {
|
|
|
6
6
|
runRules,
|
|
7
7
|
runRulesRecording,
|
|
8
8
|
toCodePoints
|
|
9
|
-
} from "./chunk-
|
|
9
|
+
} from "./chunk-25RYOB5K.js";
|
|
10
10
|
|
|
11
11
|
// src/engine/text-pipeline.ts
|
|
12
12
|
function runTextPipeline(input, options) {
|
|
@@ -28,4 +28,4 @@ export {
|
|
|
28
28
|
runTextPipeline,
|
|
29
29
|
analyzeTextPipeline
|
|
30
30
|
};
|
|
31
|
-
//# sourceMappingURL=chunk-
|
|
31
|
+
//# sourceMappingURL=chunk-47QVJOT4.js.map
|
|
@@ -1,11 +1,11 @@
|
|
|
1
1
|
"use strict";Object.defineProperty(exports, "__esModule", {value: true});
|
|
2
2
|
|
|
3
|
-
var
|
|
3
|
+
var _chunkPJ6HB77Gcjs = require('./chunk-PJ6HB77G.cjs');
|
|
4
4
|
|
|
5
5
|
// src/engine/assert-fixed-mode.ts
|
|
6
6
|
function assertFixedMode(mode, fixed, entryPoint) {
|
|
7
7
|
if (mode === void 0 || mode === fixed) return;
|
|
8
|
-
throw new (0,
|
|
8
|
+
throw new (0, _chunkPJ6HB77Gcjs.PolytypoError)(
|
|
9
9
|
"POLYTYPO_INVALID_MODE",
|
|
10
10
|
`"${entryPoint}" only supports mode "${fixed}". Received "${String(mode)}" \u2014 import from "polytypo" for the other modes, or omit \`mode\` here.`
|
|
11
11
|
);
|
|
@@ -14,4 +14,4 @@ function assertFixedMode(mode, fixed, entryPoint) {
|
|
|
14
14
|
|
|
15
15
|
|
|
16
16
|
exports.assertFixedMode = assertFixedMode;
|
|
17
|
-
//# sourceMappingURL=chunk-
|
|
17
|
+
//# sourceMappingURL=chunk-6PQMYL2O.cjs.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-
|
|
1
|
+
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-6PQMYL2O.cjs","../src/engine/assert-fixed-mode.ts"],"names":[],"mappings":"AAAA;AACE;AACF,wDAA6B;AAC7B;AACA;ACMO,SAAS,eAAA,CAAgB,IAAA,EAAe,KAAA,EAAa,UAAA,EAA0B;AACpF,EAAA,GAAA,CAAI,KAAA,IAAS,KAAA,EAAA,GAAa,KAAA,IAAS,KAAA,EAAO,MAAA;AAC1C,EAAA,MAAM,IAAI,oCAAA;AAAA,IACR,uBAAA;AAAA,IACA,CAAA,CAAA,EAAI,UAAU,CAAA,sBAAA,EAAyB,KAAK,CAAA,aAAA,EAAgB,MAAA,CAAO,IAAI,CAAC,CAAA,2EAAA;AAAA,EAE1E,CAAA;AACF;ADLA;AACA;AACE;AACF,0CAAC","file":"/home/runner/work/polytypo-js/polytypo-js/dist/chunk-6PQMYL2O.cjs","sourcesContent":[null,"import { PolytypoError } from \"../errors.js\";\nimport type { Mode } from \"../types.js\";\n\n/**\n * `polytypo/text`, `polytypo/html` and `polytypo/markdown` each fix `mode` at the type level by\n * omitting it from their Options type (AUDIT_REMEDIATION_AND_RELEASE_PLAN.md 5.1, criterion 6).\n * A plain-JS caller — or a TypeScript caller passing a wider-typed variable, which excess-property\n * checking does not catch — can still supply a conflicting `mode` at runtime. That is rejected\n * explicitly here, never silently ignored or overridden.\n */\nexport function assertFixedMode(mode: unknown, fixed: Mode, entryPoint: string): void {\n if (mode === undefined || mode === fixed) return;\n throw new PolytypoError(\n \"POLYTYPO_INVALID_MODE\",\n `\"${entryPoint}\" only supports mode \"${fixed}\". Received \"${String(mode)}\" — ` +\n `import from \"polytypo\" for the other modes, or omit \\`mode\\` here.`,\n );\n}\n"]}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import {
|
|
2
2
|
PolytypoError
|
|
3
|
-
} from "./chunk-
|
|
3
|
+
} from "./chunk-25RYOB5K.js";
|
|
4
4
|
|
|
5
5
|
// src/modes/html.ts
|
|
6
6
|
import { parse, parseFragment } from "parse5";
|
|
@@ -144,4 +144,4 @@ export {
|
|
|
144
144
|
htmlSpans,
|
|
145
145
|
htmlFragmentSpans
|
|
146
146
|
};
|
|
147
|
-
//# sourceMappingURL=chunk-
|
|
147
|
+
//# sourceMappingURL=chunk-BGTMX6VS.js.map
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import {
|
|
2
2
|
PolytypoError
|
|
3
|
-
} from "./chunk-
|
|
3
|
+
} from "./chunk-25RYOB5K.js";
|
|
4
4
|
|
|
5
5
|
// src/engine/assert-fixed-mode.ts
|
|
6
6
|
function assertFixedMode(mode, fixed, entryPoint) {
|
|
@@ -14,4 +14,4 @@ function assertFixedMode(mode, fixed, entryPoint) {
|
|
|
14
14
|
export {
|
|
15
15
|
assertFixedMode
|
|
16
16
|
};
|
|
17
|
-
//# sourceMappingURL=chunk-
|
|
17
|
+
//# sourceMappingURL=chunk-D3QQYGKW.js.map
|
|
@@ -9,7 +9,7 @@
|
|
|
9
9
|
|
|
10
10
|
|
|
11
11
|
|
|
12
|
-
var
|
|
12
|
+
var _chunkPJ6HB77Gcjs = require('./chunk-PJ6HB77G.cjs');
|
|
13
13
|
|
|
14
14
|
// src/modes/spans.ts
|
|
15
15
|
var SPACE = 32;
|
|
@@ -38,7 +38,7 @@ function normalizeSpans(spans) {
|
|
|
38
38
|
continue;
|
|
39
39
|
}
|
|
40
40
|
if (span.start < last.end) {
|
|
41
|
-
throw new (0,
|
|
41
|
+
throw new (0, _chunkPJ6HB77Gcjs.PolytypoError)(
|
|
42
42
|
"POLYTYPO_RULE_CONTRACT",
|
|
43
43
|
`mode extractor produced overlapping spans (${last.start}, ${last.end}) and (${span.start}, ${span.end})`
|
|
44
44
|
);
|
|
@@ -56,9 +56,9 @@ function concatenateSpans(source, spans) {
|
|
|
56
56
|
let previous;
|
|
57
57
|
for (const span of spans) {
|
|
58
58
|
if (previous !== void 0) {
|
|
59
|
-
cp.push(gapIsLineBoundary(source, previous.end, span.start) ?
|
|
59
|
+
cp.push(gapIsLineBoundary(source, previous.end, span.start) ? _chunkPJ6HB77Gcjs.LINE_MARKER : _chunkPJ6HB77Gcjs.MARKER);
|
|
60
60
|
}
|
|
61
|
-
for (const value of
|
|
61
|
+
for (const value of _chunkPJ6HB77Gcjs.toCodePoints.call(void 0, source.slice(span.start, span.end))) cp.push(value);
|
|
62
62
|
previous = span;
|
|
63
63
|
}
|
|
64
64
|
return cp;
|
|
@@ -67,7 +67,7 @@ function spanRangesOf(cp) {
|
|
|
67
67
|
const ranges = [];
|
|
68
68
|
let first = 0;
|
|
69
69
|
for (let i = 0; i < cp.length; i += 1) {
|
|
70
|
-
if (
|
|
70
|
+
if (_chunkPJ6HB77Gcjs.isMarker.call(void 0, cp[i])) {
|
|
71
71
|
ranges.push({ first, last: i - 1 });
|
|
72
72
|
first = i + 1;
|
|
73
73
|
}
|
|
@@ -86,7 +86,7 @@ function filterBoundaryEdits(cp, edits, ranges) {
|
|
|
86
86
|
for (const edit of edits) {
|
|
87
87
|
let containsMarker = false;
|
|
88
88
|
for (let i = edit.start; i < edit.end; i += 1) {
|
|
89
|
-
if (
|
|
89
|
+
if (_chunkPJ6HB77Gcjs.isMarker.call(void 0, cp[i])) {
|
|
90
90
|
containsMarker = true;
|
|
91
91
|
break;
|
|
92
92
|
}
|
|
@@ -111,14 +111,14 @@ function filterBoundaryEdits(cp, edits, ranges) {
|
|
|
111
111
|
function splitOnMarker(cp, expected) {
|
|
112
112
|
const pieces = [[]];
|
|
113
113
|
for (const value of cp) {
|
|
114
|
-
if (
|
|
114
|
+
if (_chunkPJ6HB77Gcjs.isMarker.call(void 0, value)) {
|
|
115
115
|
pieces.push([]);
|
|
116
116
|
continue;
|
|
117
117
|
}
|
|
118
118
|
pieces[pieces.length - 1].push(value);
|
|
119
119
|
}
|
|
120
120
|
if (pieces.length !== expected) {
|
|
121
|
-
throw new (0,
|
|
121
|
+
throw new (0, _chunkPJ6HB77Gcjs.PolytypoError)(
|
|
122
122
|
"POLYTYPO_RULE_CONTRACT",
|
|
123
123
|
`boundary markers did not survive the pipeline: expected ${expected} spans, found ${pieces.length}`
|
|
124
124
|
);
|
|
@@ -140,9 +140,9 @@ function originOfSpans(source, spans) {
|
|
|
140
140
|
const origin = [];
|
|
141
141
|
let previous;
|
|
142
142
|
for (const span of spans) {
|
|
143
|
-
if (previous !== void 0) origin.push(
|
|
143
|
+
if (previous !== void 0) origin.push(_chunkPJ6HB77Gcjs.NO_ORIGIN);
|
|
144
144
|
const base = codePointIndexOf[span.start];
|
|
145
|
-
const length =
|
|
145
|
+
const length = _chunkPJ6HB77Gcjs.toCodePoints.call(void 0, source.slice(span.start, span.end)).length;
|
|
146
146
|
for (let k = 0; k < length; k += 1) origin.push(base + k);
|
|
147
147
|
previous = span;
|
|
148
148
|
}
|
|
@@ -153,10 +153,10 @@ function originOfSpans(source, spans) {
|
|
|
153
153
|
function runRulesOverSpans(cp, planned, locale, mode, narrowTarget) {
|
|
154
154
|
let current = cp;
|
|
155
155
|
for (const id of planned) {
|
|
156
|
-
const rule =
|
|
156
|
+
const rule = _chunkPJ6HB77Gcjs.RULES[id];
|
|
157
157
|
if (rule === void 0) continue;
|
|
158
158
|
const edits = rule.apply({ cp: current, locale, mode, narrowTarget });
|
|
159
|
-
current =
|
|
159
|
+
current = _chunkPJ6HB77Gcjs.applyEdits.call(void 0, current, filterBoundaryEdits(current, edits, spanRangesOf(current)), id);
|
|
160
160
|
}
|
|
161
161
|
return current;
|
|
162
162
|
}
|
|
@@ -175,7 +175,7 @@ function runOverSpans(source, spans, planned, locale, mode, narrowTarget) {
|
|
|
175
175
|
let cursor = 0;
|
|
176
176
|
for (let i = 0; i < normalized.length; i += 1) {
|
|
177
177
|
const span = normalized[i];
|
|
178
|
-
const replacement =
|
|
178
|
+
const replacement = _chunkPJ6HB77Gcjs.fromCodePoints.call(void 0, pieces[i]);
|
|
179
179
|
const original = source.slice(span.start, span.end);
|
|
180
180
|
out += source.slice(cursor, span.start);
|
|
181
181
|
out += replacement === original ? original : replacement;
|
|
@@ -186,14 +186,14 @@ function runOverSpans(source, spans, planned, locale, mode, narrowTarget) {
|
|
|
186
186
|
function analyzeOverSpans(source, spans, planned, locale, mode, narrowTarget) {
|
|
187
187
|
const normalized = normalizeSpans(spans);
|
|
188
188
|
if (normalized.length === 0) return [];
|
|
189
|
-
return
|
|
189
|
+
return _chunkPJ6HB77Gcjs.runRulesRecording.call(void 0,
|
|
190
190
|
concatenateSpans(source, normalized),
|
|
191
191
|
planned,
|
|
192
192
|
locale,
|
|
193
193
|
mode,
|
|
194
194
|
narrowTarget,
|
|
195
195
|
originOfSpans(source, normalized),
|
|
196
|
-
|
|
196
|
+
_chunkPJ6HB77Gcjs.toCodePoints.call(void 0, source).length,
|
|
197
197
|
(current, edits) => filterBoundaryEdits(current, edits, spanRangesOf(current))
|
|
198
198
|
);
|
|
199
199
|
}
|
|
@@ -202,4 +202,4 @@ function analyzeOverSpans(source, spans, planned, locale, mode, narrowTarget) {
|
|
|
202
202
|
|
|
203
203
|
|
|
204
204
|
exports.runOverSpans = runOverSpans; exports.analyzeOverSpans = analyzeOverSpans;
|
|
205
|
-
//# sourceMappingURL=chunk-
|
|
205
|
+
//# sourceMappingURL=chunk-DCC5OQET.cjs.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-7S5PZ7T5.cjs","../src/modes/spans.ts","../src/engine/span-runner.ts"],"names":[],"mappings":"AAAA;AACE;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACF,wDAA6B;AAC7B;AACA;ACMA,IAAM,MAAA,EAAQ,EAAA;AAEd,IAAM,iBAAA,kBAAwC,IAAI,GAAA,CAAI;AAAA,EACpD,EAAA;AAAA,EAAM,EAAA;AAAA,EAAM,EAAA;AAAA,EAAM,EAAA;AAAA,EAAM,GAAA;AAAA,EAAM,IAAA;AAAA,EAAQ;AACxC,CAAC,CAAA;AAED,SAAS,iBAAA,CAAkB,MAAA,EAAgB,IAAA,EAAc,EAAA,EAAqB;AAC5E,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,IAAA,EAAM,EAAA,EAAI,EAAA,EAAI,EAAA,GAAK,CAAA,EAAG;AACjC,IAAA,GAAA,CAAI,gBAAA,CAAiB,GAAA,CAAI,MAAA,CAAO,UAAA,CAAW,CAAC,CAAC,CAAA,EAAG,OAAO,IAAA;AAAA,EACzD;AACA,EAAA,OAAO,KAAA;AACT;AAuBO,SAAS,cAAA,CAAe,KAAA,EAAgC;AAC7D,EAAA,MAAM,OAAA,EAAS,KAAA,CAAM,MAAA,CAAO,CAAC,CAAA,EAAA,GAAM,CAAA,CAAE,IAAA,EAAM,CAAA,CAAE,KAAK,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,EAAA,GAAM,CAAA,CAAE,MAAA,EAAQ,CAAA,CAAE,KAAK,CAAA;AACpF,EAAA,MAAM,IAAA,EAAc,CAAC,CAAA;AACrB,EAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,MAAA,EAAQ;AACzB,IAAA,MAAM,KAAA,EAAO,GAAA,CAAI,GAAA,CAAI,OAAA,EAAS,CAAC,CAAA;AAC/B,IAAA,GAAA,CAAI,KAAA,IAAS,KAAA,CAAA,EAAW;AACtB,MAAA,GAAA,CAAI,IAAA,CAAK,IAAI,CAAA;AACb,MAAA,QAAA;AAAA,IACF;AACA,IAAA,GAAA,CAAI,IAAA,CAAK,MAAA,EAAQ,IAAA,CAAK,GAAA,EAAK;AACzB,MAAA,MAAM,IAAI,oCAAA;AAAA,QACR,wBAAA;AAAA,QACA,CAAA,2CAAA,EAA8C,IAAA,CAAK,KAAK,CAAA,EAAA,EAAK,IAAA,CAAK,GAAG,CAAA,OAAA,EAAU,IAAA,CAAK,KAAK,CAAA,EAAA,EAAK,IAAA,CAAK,GAAG,CAAA,CAAA;AAAA,MACxG,CAAA;AAAA,IACF;AACA,IAAA,GAAA,CAAI,IAAA,CAAK,MAAA,IAAU,IAAA,CAAK,GAAA,EAAK;AAC3B,MAAA,GAAA,CAAI,GAAA,CAAI,OAAA,EAAS,CAAC,EAAA,EAAI,EAAE,KAAA,EAAO,IAAA,CAAK,KAAA,EAAO,GAAA,EAAK,IAAA,CAAK,IAAI,CAAA;AACzD,MAAA,QAAA;AAAA,IACF;AACA,IAAA,GAAA,CAAI,IAAA,CAAK,IAAI,CAAA;AAAA,EACf;AACA,EAAA,OAAO,GAAA;AACT;AAGO,SAAS,gBAAA,CAAiB,MAAA,EAAgB,KAAA,EAAkC;AACjF,EAAA,MAAM,GAAA,EAAe,CAAC,CAAA;AACtB,EAAA,IAAI,QAAA;AACJ,EAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,KAAA,EAAO;AACxB,IAAA,GAAA,CAAI,SAAA,IAAa,KAAA,CAAA,EAAW;AAC1B,MAAA,EAAA,CAAG,IAAA,CAAK,iBAAA,CAAkB,MAAA,EAAQ,QAAA,CAAS,GAAA,EAAK,IAAA,CAAK,KAAK,EAAA,EAAI,8BAAA,EAAc,wBAAM,CAAA;AAAA,IACpF;AACA,IAAA,IAAA,CAAA,MAAW,MAAA,GAAS,4CAAA,MAAa,CAAO,KAAA,CAAM,IAAA,CAAK,KAAA,EAAO,IAAA,CAAK,GAAG,CAAC,CAAA,EAAG,EAAA,CAAG,IAAA,CAAK,KAAK,CAAA;AACnF,IAAA,SAAA,EAAW,IAAA;AAAA,EACb;AACA,EAAA,OAAO,EAAA;AACT;AAOO,SAAS,YAAA,CAAa,EAAA,EAAoC;AAC/D,EAAA,MAAM,OAAA,EAAsB,CAAC,CAAA;AAC7B,EAAA,IAAI,MAAA,EAAQ,CAAA;AACZ,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,CAAA,EAAG,EAAA,EAAI,EAAA,CAAG,MAAA,EAAQ,EAAA,GAAK,CAAA,EAAG;AACrC,IAAA,GAAA,CAAI,wCAAA,EAAS,CAAG,CAAC,CAAW,CAAA,EAAG;AAC7B,MAAA,MAAA,CAAO,IAAA,CAAK,EAAE,KAAA,EAAO,IAAA,EAAM,EAAA,EAAI,EAAE,CAAC,CAAA;AAClC,MAAA,MAAA,EAAQ,EAAA,EAAI,CAAA;AAAA,IACd;AAAA,EACF;AACA,EAAA,MAAA,CAAO,IAAA,CAAK,EAAE,KAAA,EAAO,IAAA,EAAM,EAAA,CAAG,OAAA,EAAS,EAAE,CAAC,CAAA;AAC1C,EAAA,OAAO,MAAA;AACT;AAEA,SAAS,cAAA,CAAe,MAAA,EAA8B,CAAA,EAAkC;AACtF,EAAA,IAAA,CAAA,MAAW,MAAA,GAAS,MAAA,EAAQ;AAC1B,IAAA,GAAA,CAAI,EAAA,GAAK,KAAA,CAAM,MAAA,GAAS,EAAA,GAAK,KAAA,CAAM,KAAA,EAAO,CAAA,EAAG,OAAO,KAAA;AAAA,EACtD;AACA,EAAA,OAAO,KAAA,CAAA;AACT;AA4BO,SAAS,mBAAA,CACd,EAAA,EACA,KAAA,EACA,MAAA,EACQ;AACR,EAAA,MAAM,IAAA,EAAc,CAAC,CAAA;AACrB,EAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,KAAA,EAAO;AACxB,IAAA,IAAI,eAAA,EAAiB,KAAA;AACrB,IAAA,IAAA,CAAA,IAAS,EAAA,EAAI,IAAA,CAAK,KAAA,EAAO,EAAA,EAAI,IAAA,CAAK,GAAA,EAAK,EAAA,GAAK,CAAA,EAAG;AAC7C,MAAA,GAAA,CAAI,wCAAA,EAAS,CAAG,CAAC,CAAW,CAAA,EAAG;AAC7B,QAAA,eAAA,EAAiB,IAAA;AACjB,QAAA,KAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,GAAA,CAAI,cAAA,EAAgB,QAAA;AAIpB,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,KAAA;AACf,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,IAAA,EAAM,CAAA;AACrB,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,IAAA,EAAM,IAAA,CAAK,KAAA;AAC1B,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,WAAA,CAAY,MAAA;AAC3B,IAAA,MAAM,KAAA,EAAO,cAAA,CAAe,MAAA,EAAQ,CAAC,CAAA;AACrC,IAAA,GAAA,CAAI,KAAA,IAAS,KAAA,EAAA,GAAA,CAAc,EAAA,IAAM,IAAA,CAAK,MAAA,GAAS,EAAA,IAAM,IAAA,CAAK,IAAA,CAAA,EAAO;AAC/D,MAAA,GAAA,CAAI,EAAA,EAAI,CAAA,EAAG,QAAA;AAMX,MAAA,MAAM,MAAA,EAAQ,IAAA,CAAK,WAAA,CAAY,CAAC,CAAA;AAChC,MAAA,MAAM,KAAA,EAAO,IAAA,CAAK,WAAA,CAAY,EAAA,EAAI,CAAC,CAAA;AACnC,MAAA,GAAA,CAAI,EAAA,EAAI,EAAA,GAAK,EAAA,IAAM,IAAA,CAAK,MAAA,GAAS,MAAA,IAAU,MAAA,GAAS,EAAA,CAAG,CAAC,EAAA,IAAM,KAAA,EAAO,QAAA;AACrE,MAAA,GAAA,CAAI,EAAA,EAAI,EAAA,GAAK,EAAA,IAAM,IAAA,CAAK,KAAA,GAAQ,KAAA,IAAS,MAAA,GAAS,EAAA,CAAG,CAAC,EAAA,IAAM,KAAA,EAAO,QAAA;AAAA,IACrE;AAEA,IAAA,GAAA,CAAI,IAAA,CAAK,IAAI,CAAA;AAAA,EACf;AACA,EAAA,OAAO,GAAA;AACT;AAGO,SAAS,aAAA,CAAc,EAAA,EAAuB,QAAA,EAA8B;AACjF,EAAA,MAAM,OAAA,EAAqB,CAAC,CAAC,CAAC,CAAA;AAC9B,EAAA,IAAA,CAAA,MAAW,MAAA,GAAS,EAAA,EAAI;AACtB,IAAA,GAAA,CAAI,wCAAA,KAAc,CAAA,EAAG;AACnB,MAAA,MAAA,CAAO,IAAA,CAAK,CAAC,CAAC,CAAA;AACd,MAAA,QAAA;AAAA,IACF;AACA,IAAC,MAAA,CAAO,MAAA,CAAO,OAAA,EAAS,CAAC,CAAA,CAAe,IAAA,CAAK,KAAK,CAAA;AAAA,EACpD;AACA,EAAA,GAAA,CAAI,MAAA,CAAO,OAAA,IAAW,QAAA,EAAU;AAC9B,IAAA,MAAM,IAAI,oCAAA;AAAA,MACR,wBAAA;AAAA,MACA,CAAA,wDAAA,EAA2D,QAAQ,CAAA,cAAA,EAAiB,MAAA,CAAO,MAAM,CAAA;AAAA,IAAA;AACnG,EAAA;AAEF,EAAA;AACF;AASO;AACL,EAAA;AACA,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAAW,EAAA;AAEb,EAAA;AAEA,EAAA;AACA,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAAW,EAAA;AAEb,EAAA;AACF;ADlFA;AACA;AEjIA;AAOE,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAA4F,EAAA;AAE9F,EAAA;AACF;AAYO;AAQL,EAAA;AACA,EAAA;AAEA,EAAA;AAAoB,IAAA;AACiB,IAAA;AACnC,IAAA;AACA,IAAA;AACA,IAAA;AACA,EAAA;AAEF,EAAA;AAEA,EAAA;AACA,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAAc,EAAA;AAEhB,EAAA;AACF;AAOO;AAQL,EAAA;AACA,EAAA;AACA,EAAA;AAAO,IAAA;AAC8B,IAAA;AACnC,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACgC,IAAA;AACX,IAAA;AACwD,EAAA;AAEjF;AF4FA;AACA;AACA;AACA;AACA","file":"/home/runner/work/polytypo-js/polytypo-js/dist/chunk-7S5PZ7T5.cjs","sourcesContent":[null,"import { toCodePoints } from \"../engine/codepoints.js\";\nimport { isMarker, LINE_MARKER, MARKER } from \"../engine/sentinels.js\";\nimport { PolytypoError } from \"../errors.js\";\nimport type { Edit } from \"../types.js\";\nimport { NO_ORIGIN } from \"../engine/origin.js\";\n\n/**\n * The boundary markers are defined in the engine (src/engine/sentinels.ts), which is where all\n * three non-code-point sentinels live and where their disjointness is maintained. They are\n * re-exported here because the mode layer is what writes them into the array.\n *\n * Which marker separates two spans is decided by the **raw source bytes of the gap between\n * them**, so it is decidable without asking the parser anything and is identical in five\n * runtimes: a gap containing a line terminator gives `LINE_MARKER`, anything else `MARKER`.\n */\nexport { LINE_MARKER, MARKER, isMarker } from \"../engine/sentinels.js\";\n\n/** `BREAK` as the rules define it (`spaces.md` 3.1), tested against the gap's raw source. */\n/** U+0020, the one emitted code point whose meaning is positional (modes.md 3.4, 5 item 2). */\nconst SPACE = 0x20;\n\nconst LINE_TERMINATORS: ReadonlySet<number> = new Set([\n 0x0a, 0x0d, 0x0b, 0x0c, 0x85, 0x2028, 0x2029,\n]);\n\nfunction gapIsLineBoundary(source: string, from: number, to: number): boolean {\n for (let i = from; i < to; i += 1) {\n if (LINE_TERMINATORS.has(source.charCodeAt(i))) return true;\n }\n return false;\n}\n\n/**\n * A processable span, identified by its offsets in the **original source**. Offsets are UTF-16\n * indices into the JS source string, which is what every JS parser reports; they are an\n * implementation detail of this runtime and never reach the rules, which index code points.\n */\nexport interface Span {\n readonly start: number;\n readonly end: number;\n}\n\n/** A span's extent in the concatenated code-point array: `s₀` and `s₁` of modes.md 3.4. */\nexport interface SpanRange {\n readonly first: number;\n readonly last: number;\n}\n\n/**\n * Sort, drop empties, and coalesce spans separated by nothing in the source (modes.md 7.5:\n * a parser that reports one text run as two adjacent nodes must not manufacture a boundary).\n * Overlapping spans are an extractor bug and are rejected rather than silently merged.\n */\nexport function normalizeSpans(spans: readonly Span[]): Span[] {\n const sorted = spans.filter((s) => s.end > s.start).sort((a, b) => a.start - b.start);\n const out: Span[] = [];\n for (const span of sorted) {\n const last = out[out.length - 1];\n if (last === undefined) {\n out.push(span);\n continue;\n }\n if (span.start < last.end) {\n throw new PolytypoError(\n \"POLYTYPO_RULE_CONTRACT\",\n `mode extractor produced overlapping spans (${last.start}, ${last.end}) and (${span.start}, ${span.end})`,\n );\n }\n if (span.start === last.end) {\n out[out.length - 1] = { start: last.start, end: span.end };\n continue;\n }\n out.push(span);\n }\n return out;\n}\n\n/** `S₁ ⌢ [marker] ⌢ S₂ ⌢ … ⌢ Sₘ` (modes.md 3.5 step 2). */\nexport function concatenateSpans(source: string, spans: readonly Span[]): number[] {\n const cp: number[] = [];\n let previous: Span | undefined;\n for (const span of spans) {\n if (previous !== undefined) {\n cp.push(gapIsLineBoundary(source, previous.end, span.start) ? LINE_MARKER : MARKER);\n }\n for (const value of toCodePoints(source.slice(span.start, span.end))) cp.push(value);\n previous = span;\n }\n return cp;\n}\n\n/**\n * The span extents of the array as it stands. Recomputed after every rule, because applying\n * edits shifts every index after the first one — the markers themselves are guaranteed to\n * survive, since no edit may contain one.\n */\nexport function spanRangesOf(cp: readonly number[]): SpanRange[] {\n const ranges: SpanRange[] = [];\n let first = 0;\n for (let i = 0; i < cp.length; i += 1) {\n if (isMarker(cp[i] as number)) {\n ranges.push({ first, last: i - 1 });\n first = i + 1;\n }\n }\n ranges.push({ first, last: cp.length - 1 });\n return ranges;\n}\n\nfunction spanContaining(ranges: readonly SpanRange[], p: number): SpanRange | undefined {\n for (const range of ranges) {\n if (p >= range.first && p <= range.last + 1) return range;\n }\n return undefined;\n}\n\n/**\n * modes.md 3.4, two safety nets, both pure functions of `(p, q, r, s₀, s₁)` — so the verdict is\n * identical on every run, which is what makes redistribution deterministic (modes.md 5, point 3).\n *\n * 1. **No edit may contain a marker.** No rule can produce one; one that does is a bug, and the\n * edit is discarded rather than treated as a redistribution question.\n * 2. **The edge-growth rule.** An edit is discarded if it would place code points at an\n * extremity of its span that were not there before: with `d = q - p + 1` the replaced length\n * and `r` the replacement length, discard when `p = s₀ and r > d`, or `q = s₁ and r > d`.\n *\n * The length test is the whole rule and needs no knowledge of Markdown or HTML syntax. It\n * separates exactly the cases that matter: `\"` → `“` at an edge is 1 → 1 and applies; `--` →\n * `␣–␣` at an edge is 2 → 3 and is discarded, while the same edit interior to a span applies;\n * `(c)` → `©` is 3 → 1 and applies, because shrinking is always safe. An insertion has `d = 0`,\n * so it is discarded exactly when its position coincides with a span edge — the rule this one\n * generalises.\n *\n * **Deletion at an edge is not restricted here**, and must not be: `r > d` is false for a\n * deletion, so this filter never sees one. Deletion at an edge is governed by modes.md 3.3's\n * *Edge tests* clause instead — where a rule asks \"am I at the edge of the text I am allowed to\n * modify\" rather than \"what character is here\", the marker behaves as `NONE`. That clause is the\n * one place a rule may treat a span edge as the end of the text, it exists because `spaces` is\n * the only rule that deletes, and 3.3 states the division: **a rule that deletes must treat a\n * span edge as the end of the text; a rule that replaces or inserts must not, and is governed by\n * this filter.** The single test that claims the clause is `spaces.md` 3.2 step 4.\n */\nexport function filterBoundaryEdits(\n cp: readonly number[],\n edits: readonly Edit[],\n ranges: readonly SpanRange[],\n): Edit[] {\n const out: Edit[] = [];\n for (const edit of edits) {\n let containsMarker = false;\n for (let i = edit.start; i < edit.end; i += 1) {\n if (isMarker(cp[i] as number)) {\n containsMarker = true;\n break;\n }\n }\n if (containsMarker) continue;\n\n // `[start, end)` in the engine's half-open form is `cp[p … q]` with p = start, q = end - 1;\n // an insertion is `q = p - 1`, which falls out of the same expression.\n const p = edit.start;\n const q = edit.end - 1;\n const d = edit.end - edit.start;\n const r = edit.replacement.length;\n const span = spanContaining(ranges, p);\n if (span !== undefined && (p === span.first || q === span.last)) {\n if (r > d) continue;\n // The character clause. `r > d` is not the rule, only a formalisation of it that misses\n // `r === d`: `dashes` P3 admits a run of THREE dashes, so `---` -> `␣–␣` is 3 -> 3 and the\n // length test sees nothing while U+0020 lands on both extremities anyway. Testing the\n // character catches it, and only U+0020 needs testing — it is the one code point any rule\n // emits whose meaning comes from its position rather than from itself (modes.md 5 item 2).\n const first = edit.replacement[0];\n const last = edit.replacement[r - 1];\n if (r > 0 && p === span.first && first === SPACE && cp[p] !== SPACE) continue;\n if (r > 0 && q === span.last && last === SPACE && cp[q] !== SPACE) continue;\n }\n\n out.push(edit);\n }\n return out;\n}\n\n/** Redistribute the transformed array back to one piece per span (modes.md 3.5 step 4). */\nexport function splitOnMarker(cp: readonly number[], expected: number): number[][] {\n const pieces: number[][] = [[]];\n for (const value of cp) {\n if (isMarker(value)) {\n pieces.push([]);\n continue;\n }\n (pieces[pieces.length - 1] as number[]).push(value);\n }\n if (pieces.length !== expected) {\n throw new PolytypoError(\n \"POLYTYPO_RULE_CONTRACT\",\n `boundary markers did not survive the pipeline: expected ${expected} spans, found ${pieces.length}`,\n );\n }\n return pieces;\n}\n\n/**\n * The origin map for `concatenateSpans` (analyze.md §2): for every code point of the joined\n * array, the code-point offset of the character it came from IN THE DOCUMENT, and NO_ORIGIN for\n * the markers, which came from nowhere. `Span` bounds are native string indices, so the\n * conversion to code-point offsets happens here and not in the caller — this is the one place\n * that knows both coordinate systems.\n */\nexport function originOfSpans(source: string, spans: readonly Span[]): number[] {\n const codePointIndexOf = new Array<number>(source.length + 1);\n let cpIndex = 0;\n for (let i = 0; i < source.length;) {\n codePointIndexOf[i] = cpIndex;\n const code = source.codePointAt(i) as number;\n const width = code > 0xffff ? 2 : 1;\n if (width === 2) codePointIndexOf[i + 1] = cpIndex;\n i += width;\n cpIndex += 1;\n }\n codePointIndexOf[source.length] = cpIndex;\n\n const origin: number[] = [];\n let previous: Span | undefined;\n for (const span of spans) {\n if (previous !== undefined) origin.push(NO_ORIGIN);\n const base = codePointIndexOf[span.start] as number;\n const length = toCodePoints(source.slice(span.start, span.end)).length;\n for (let k = 0; k < length; k += 1) origin.push(base + k);\n previous = span;\n }\n return origin;\n}\n","import {\n concatenateSpans,\n filterBoundaryEdits,\n normalizeSpans,\n originOfSpans,\n spanRangesOf,\n splitOnMarker,\n type Span,\n} from \"../modes/spans.js\";\nimport { RULES } from \"../rules/registry.js\";\nimport type { LocaleData, Mode, RuleId } from \"../types.js\";\nimport { applyEdits } from \"./edits.js\";\nimport { fromCodePoints, toCodePoints } from \"./codepoints.js\";\nimport type { Change } from \"./origin.js\";\nimport { runRulesRecording } from \"./rule-runner.js\";\n\n/**\n * The same sequence as `runRules` (`./rule-runner.js`), with the two boundary filters of\n * modes.md 3.4 interposed. The span extents are recomputed after every rule, because applying\n * an edit shifts every index after it; the markers themselves always survive, since no edit may\n * contain one.\n */\nfunction runRulesOverSpans(\n cp: readonly number[],\n planned: readonly RuleId[],\n locale: LocaleData,\n mode: Mode,\n narrowTarget: number,\n): readonly number[] {\n let current = cp;\n for (const id of planned) {\n const rule = RULES[id];\n if (rule === undefined) continue;\n const edits = rule.apply({ cp: current, locale, mode, narrowTarget });\n current = applyEdits(current, filterBoundaryEdits(current, edits, spanRangesOf(current)), id);\n }\n return current;\n}\n\n/**\n * modes.md 3.5. The pipeline runs **once**, over the marker-separated concatenation of every\n * processable span — not per span, which would pair quotation marks in isolation, and not over a\n * naive concatenation, which would manufacture adjacencies the document does not have.\n *\n * The output is the input with a set of disjoint substring replacements applied and nothing else\n * (modes.md 4). A span whose content the rules did not change contributes no replacement, so a\n * document needing no changes comes back byte-identical; the parser located the spans and was\n * then discarded, and the document is never serialised.\n */\nexport function runOverSpans(\n source: string,\n spans: readonly Span[],\n planned: readonly RuleId[],\n locale: LocaleData,\n mode: Mode,\n narrowTarget: number,\n): string {\n const normalized = normalizeSpans(spans);\n if (normalized.length === 0) return source;\n\n const transformed = runRulesOverSpans(\n concatenateSpans(source, normalized),\n planned,\n locale,\n mode,\n narrowTarget,\n );\n const pieces = splitOnMarker(transformed, normalized.length);\n\n let out = \"\";\n let cursor = 0;\n for (let i = 0; i < normalized.length; i += 1) {\n const span = normalized[i] as Span;\n const replacement = fromCodePoints(pieces[i] as number[]);\n const original = source.slice(span.start, span.end);\n out += source.slice(cursor, span.start);\n out += replacement === original ? original : replacement;\n cursor = span.end;\n }\n return out + source.slice(cursor);\n}\n\n/**\n * `runOverSpans`, reporting instead of applying (analyze.md §1). The span table supplies the\n * origin map, so every change comes back in DOCUMENT coordinates — analyze.md §6 names a\n * runtime that reports span-local offsets here as the mistake that passes every text-mode test.\n */\nexport function analyzeOverSpans(\n source: string,\n spans: readonly Span[],\n planned: readonly RuleId[],\n locale: LocaleData,\n mode: Mode,\n narrowTarget: number,\n): Change[] {\n const normalized = normalizeSpans(spans);\n if (normalized.length === 0) return [];\n return runRulesRecording(\n concatenateSpans(source, normalized),\n planned,\n locale,\n mode,\n narrowTarget,\n originOfSpans(source, normalized),\n toCodePoints(source).length,\n (current, edits) => filterBoundaryEdits(current, edits, spanRangesOf(current)),\n );\n}\n"]}
|
|
1
|
+
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-DCC5OQET.cjs","../src/modes/spans.ts","../src/engine/span-runner.ts"],"names":[],"mappings":"AAAA;AACE;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACF,wDAA6B;AAC7B;AACA;ACMA,IAAM,MAAA,EAAQ,EAAA;AAEd,IAAM,iBAAA,kBAAwC,IAAI,GAAA,CAAI;AAAA,EACpD,EAAA;AAAA,EAAM,EAAA;AAAA,EAAM,EAAA;AAAA,EAAM,EAAA;AAAA,EAAM,GAAA;AAAA,EAAM,IAAA;AAAA,EAAQ;AACxC,CAAC,CAAA;AAED,SAAS,iBAAA,CAAkB,MAAA,EAAgB,IAAA,EAAc,EAAA,EAAqB;AAC5E,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,IAAA,EAAM,EAAA,EAAI,EAAA,EAAI,EAAA,GAAK,CAAA,EAAG;AACjC,IAAA,GAAA,CAAI,gBAAA,CAAiB,GAAA,CAAI,MAAA,CAAO,UAAA,CAAW,CAAC,CAAC,CAAA,EAAG,OAAO,IAAA;AAAA,EACzD;AACA,EAAA,OAAO,KAAA;AACT;AAuBO,SAAS,cAAA,CAAe,KAAA,EAAgC;AAC7D,EAAA,MAAM,OAAA,EAAS,KAAA,CAAM,MAAA,CAAO,CAAC,CAAA,EAAA,GAAM,CAAA,CAAE,IAAA,EAAM,CAAA,CAAE,KAAK,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,EAAA,GAAM,CAAA,CAAE,MAAA,EAAQ,CAAA,CAAE,KAAK,CAAA;AACpF,EAAA,MAAM,IAAA,EAAc,CAAC,CAAA;AACrB,EAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,MAAA,EAAQ;AACzB,IAAA,MAAM,KAAA,EAAO,GAAA,CAAI,GAAA,CAAI,OAAA,EAAS,CAAC,CAAA;AAC/B,IAAA,GAAA,CAAI,KAAA,IAAS,KAAA,CAAA,EAAW;AACtB,MAAA,GAAA,CAAI,IAAA,CAAK,IAAI,CAAA;AACb,MAAA,QAAA;AAAA,IACF;AACA,IAAA,GAAA,CAAI,IAAA,CAAK,MAAA,EAAQ,IAAA,CAAK,GAAA,EAAK;AACzB,MAAA,MAAM,IAAI,oCAAA;AAAA,QACR,wBAAA;AAAA,QACA,CAAA,2CAAA,EAA8C,IAAA,CAAK,KAAK,CAAA,EAAA,EAAK,IAAA,CAAK,GAAG,CAAA,OAAA,EAAU,IAAA,CAAK,KAAK,CAAA,EAAA,EAAK,IAAA,CAAK,GAAG,CAAA,CAAA;AAAA,MACxG,CAAA;AAAA,IACF;AACA,IAAA,GAAA,CAAI,IAAA,CAAK,MAAA,IAAU,IAAA,CAAK,GAAA,EAAK;AAC3B,MAAA,GAAA,CAAI,GAAA,CAAI,OAAA,EAAS,CAAC,EAAA,EAAI,EAAE,KAAA,EAAO,IAAA,CAAK,KAAA,EAAO,GAAA,EAAK,IAAA,CAAK,IAAI,CAAA;AACzD,MAAA,QAAA;AAAA,IACF;AACA,IAAA,GAAA,CAAI,IAAA,CAAK,IAAI,CAAA;AAAA,EACf;AACA,EAAA,OAAO,GAAA;AACT;AAGO,SAAS,gBAAA,CAAiB,MAAA,EAAgB,KAAA,EAAkC;AACjF,EAAA,MAAM,GAAA,EAAe,CAAC,CAAA;AACtB,EAAA,IAAI,QAAA;AACJ,EAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,KAAA,EAAO;AACxB,IAAA,GAAA,CAAI,SAAA,IAAa,KAAA,CAAA,EAAW;AAC1B,MAAA,EAAA,CAAG,IAAA,CAAK,iBAAA,CAAkB,MAAA,EAAQ,QAAA,CAAS,GAAA,EAAK,IAAA,CAAK,KAAK,EAAA,EAAI,8BAAA,EAAc,wBAAM,CAAA;AAAA,IACpF;AACA,IAAA,IAAA,CAAA,MAAW,MAAA,GAAS,4CAAA,MAAa,CAAO,KAAA,CAAM,IAAA,CAAK,KAAA,EAAO,IAAA,CAAK,GAAG,CAAC,CAAA,EAAG,EAAA,CAAG,IAAA,CAAK,KAAK,CAAA;AACnF,IAAA,SAAA,EAAW,IAAA;AAAA,EACb;AACA,EAAA,OAAO,EAAA;AACT;AAOO,SAAS,YAAA,CAAa,EAAA,EAAoC;AAC/D,EAAA,MAAM,OAAA,EAAsB,CAAC,CAAA;AAC7B,EAAA,IAAI,MAAA,EAAQ,CAAA;AACZ,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,CAAA,EAAG,EAAA,EAAI,EAAA,CAAG,MAAA,EAAQ,EAAA,GAAK,CAAA,EAAG;AACrC,IAAA,GAAA,CAAI,wCAAA,EAAS,CAAG,CAAC,CAAW,CAAA,EAAG;AAC7B,MAAA,MAAA,CAAO,IAAA,CAAK,EAAE,KAAA,EAAO,IAAA,EAAM,EAAA,EAAI,EAAE,CAAC,CAAA;AAClC,MAAA,MAAA,EAAQ,EAAA,EAAI,CAAA;AAAA,IACd;AAAA,EACF;AACA,EAAA,MAAA,CAAO,IAAA,CAAK,EAAE,KAAA,EAAO,IAAA,EAAM,EAAA,CAAG,OAAA,EAAS,EAAE,CAAC,CAAA;AAC1C,EAAA,OAAO,MAAA;AACT;AAEA,SAAS,cAAA,CAAe,MAAA,EAA8B,CAAA,EAAkC;AACtF,EAAA,IAAA,CAAA,MAAW,MAAA,GAAS,MAAA,EAAQ;AAC1B,IAAA,GAAA,CAAI,EAAA,GAAK,KAAA,CAAM,MAAA,GAAS,EAAA,GAAK,KAAA,CAAM,KAAA,EAAO,CAAA,EAAG,OAAO,KAAA;AAAA,EACtD;AACA,EAAA,OAAO,KAAA,CAAA;AACT;AA4BO,SAAS,mBAAA,CACd,EAAA,EACA,KAAA,EACA,MAAA,EACQ;AACR,EAAA,MAAM,IAAA,EAAc,CAAC,CAAA;AACrB,EAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,KAAA,EAAO;AACxB,IAAA,IAAI,eAAA,EAAiB,KAAA;AACrB,IAAA,IAAA,CAAA,IAAS,EAAA,EAAI,IAAA,CAAK,KAAA,EAAO,EAAA,EAAI,IAAA,CAAK,GAAA,EAAK,EAAA,GAAK,CAAA,EAAG;AAC7C,MAAA,GAAA,CAAI,wCAAA,EAAS,CAAG,CAAC,CAAW,CAAA,EAAG;AAC7B,QAAA,eAAA,EAAiB,IAAA;AACjB,QAAA,KAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,GAAA,CAAI,cAAA,EAAgB,QAAA;AAIpB,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,KAAA;AACf,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,IAAA,EAAM,CAAA;AACrB,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,IAAA,EAAM,IAAA,CAAK,KAAA;AAC1B,IAAA,MAAM,EAAA,EAAI,IAAA,CAAK,WAAA,CAAY,MAAA;AAC3B,IAAA,MAAM,KAAA,EAAO,cAAA,CAAe,MAAA,EAAQ,CAAC,CAAA;AACrC,IAAA,GAAA,CAAI,KAAA,IAAS,KAAA,EAAA,GAAA,CAAc,EAAA,IAAM,IAAA,CAAK,MAAA,GAAS,EAAA,IAAM,IAAA,CAAK,IAAA,CAAA,EAAO;AAC/D,MAAA,GAAA,CAAI,EAAA,EAAI,CAAA,EAAG,QAAA;AAMX,MAAA,MAAM,MAAA,EAAQ,IAAA,CAAK,WAAA,CAAY,CAAC,CAAA;AAChC,MAAA,MAAM,KAAA,EAAO,IAAA,CAAK,WAAA,CAAY,EAAA,EAAI,CAAC,CAAA;AACnC,MAAA,GAAA,CAAI,EAAA,EAAI,EAAA,GAAK,EAAA,IAAM,IAAA,CAAK,MAAA,GAAS,MAAA,IAAU,MAAA,GAAS,EAAA,CAAG,CAAC,EAAA,IAAM,KAAA,EAAO,QAAA;AACrE,MAAA,GAAA,CAAI,EAAA,EAAI,EAAA,GAAK,EAAA,IAAM,IAAA,CAAK,KAAA,GAAQ,KAAA,IAAS,MAAA,GAAS,EAAA,CAAG,CAAC,EAAA,IAAM,KAAA,EAAO,QAAA;AAAA,IACrE;AAEA,IAAA,GAAA,CAAI,IAAA,CAAK,IAAI,CAAA;AAAA,EACf;AACA,EAAA,OAAO,GAAA;AACT;AAGO,SAAS,aAAA,CAAc,EAAA,EAAuB,QAAA,EAA8B;AACjF,EAAA,MAAM,OAAA,EAAqB,CAAC,CAAC,CAAC,CAAA;AAC9B,EAAA,IAAA,CAAA,MAAW,MAAA,GAAS,EAAA,EAAI;AACtB,IAAA,GAAA,CAAI,wCAAA,KAAc,CAAA,EAAG;AACnB,MAAA,MAAA,CAAO,IAAA,CAAK,CAAC,CAAC,CAAA;AACd,MAAA,QAAA;AAAA,IACF;AACA,IAAC,MAAA,CAAO,MAAA,CAAO,OAAA,EAAS,CAAC,CAAA,CAAe,IAAA,CAAK,KAAK,CAAA;AAAA,EACpD;AACA,EAAA,GAAA,CAAI,MAAA,CAAO,OAAA,IAAW,QAAA,EAAU;AAC9B,IAAA,MAAM,IAAI,oCAAA;AAAA,MACR,wBAAA;AAAA,MACA,CAAA,wDAAA,EAA2D,QAAQ,CAAA,cAAA,EAAiB,MAAA,CAAO,MAAM,CAAA;AAAA,IAAA;AACnG,EAAA;AAEF,EAAA;AACF;AASO;AACL,EAAA;AACA,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAAW,EAAA;AAEb,EAAA;AAEA,EAAA;AACA,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAAW,EAAA;AAEb,EAAA;AACF;ADlFA;AACA;AEjIA;AAOE,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAA4F,EAAA;AAE9F,EAAA;AACF;AAYO;AAQL,EAAA;AACA,EAAA;AAEA,EAAA;AAAoB,IAAA;AACiB,IAAA;AACnC,IAAA;AACA,IAAA;AACA,IAAA;AACA,EAAA;AAEF,EAAA;AAEA,EAAA;AACA,EAAA;AACA,EAAA;AACE,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AAAc,EAAA;AAEhB,EAAA;AACF;AAOO;AAQL,EAAA;AACA,EAAA;AACA,EAAA;AAAO,IAAA;AAC8B,IAAA;AACnC,IAAA;AACA,IAAA;AACA,IAAA;AACA,IAAA;AACgC,IAAA;AACX,IAAA;AACwD,EAAA;AAEjF;AF4FA;AACA;AACA;AACA;AACA","file":"/home/runner/work/polytypo-js/polytypo-js/dist/chunk-DCC5OQET.cjs","sourcesContent":[null,"import { toCodePoints } from \"../engine/codepoints.js\";\nimport { isMarker, LINE_MARKER, MARKER } from \"../engine/sentinels.js\";\nimport { PolytypoError } from \"../errors.js\";\nimport type { Edit } from \"../types.js\";\nimport { NO_ORIGIN } from \"../engine/origin.js\";\n\n/**\n * The boundary markers are defined in the engine (src/engine/sentinels.ts), which is where all\n * three non-code-point sentinels live and where their disjointness is maintained. They are\n * re-exported here because the mode layer is what writes them into the array.\n *\n * Which marker separates two spans is decided by the **raw source bytes of the gap between\n * them**, so it is decidable without asking the parser anything and is identical in five\n * runtimes: a gap containing a line terminator gives `LINE_MARKER`, anything else `MARKER`.\n */\nexport { LINE_MARKER, MARKER, isMarker } from \"../engine/sentinels.js\";\n\n/** `BREAK` as the rules define it (`spaces.md` 3.1), tested against the gap's raw source. */\n/** U+0020, the one emitted code point whose meaning is positional (modes.md 3.4, 5 item 2). */\nconst SPACE = 0x20;\n\nconst LINE_TERMINATORS: ReadonlySet<number> = new Set([\n 0x0a, 0x0d, 0x0b, 0x0c, 0x85, 0x2028, 0x2029,\n]);\n\nfunction gapIsLineBoundary(source: string, from: number, to: number): boolean {\n for (let i = from; i < to; i += 1) {\n if (LINE_TERMINATORS.has(source.charCodeAt(i))) return true;\n }\n return false;\n}\n\n/**\n * A processable span, identified by its offsets in the **original source**. Offsets are UTF-16\n * indices into the JS source string, which is what every JS parser reports; they are an\n * implementation detail of this runtime and never reach the rules, which index code points.\n */\nexport interface Span {\n readonly start: number;\n readonly end: number;\n}\n\n/** A span's extent in the concatenated code-point array: `s₀` and `s₁` of modes.md 3.4. */\nexport interface SpanRange {\n readonly first: number;\n readonly last: number;\n}\n\n/**\n * Sort, drop empties, and coalesce spans separated by nothing in the source (modes.md 7.5:\n * a parser that reports one text run as two adjacent nodes must not manufacture a boundary).\n * Overlapping spans are an extractor bug and are rejected rather than silently merged.\n */\nexport function normalizeSpans(spans: readonly Span[]): Span[] {\n const sorted = spans.filter((s) => s.end > s.start).sort((a, b) => a.start - b.start);\n const out: Span[] = [];\n for (const span of sorted) {\n const last = out[out.length - 1];\n if (last === undefined) {\n out.push(span);\n continue;\n }\n if (span.start < last.end) {\n throw new PolytypoError(\n \"POLYTYPO_RULE_CONTRACT\",\n `mode extractor produced overlapping spans (${last.start}, ${last.end}) and (${span.start}, ${span.end})`,\n );\n }\n if (span.start === last.end) {\n out[out.length - 1] = { start: last.start, end: span.end };\n continue;\n }\n out.push(span);\n }\n return out;\n}\n\n/** `S₁ ⌢ [marker] ⌢ S₂ ⌢ … ⌢ Sₘ` (modes.md 3.5 step 2). */\nexport function concatenateSpans(source: string, spans: readonly Span[]): number[] {\n const cp: number[] = [];\n let previous: Span | undefined;\n for (const span of spans) {\n if (previous !== undefined) {\n cp.push(gapIsLineBoundary(source, previous.end, span.start) ? LINE_MARKER : MARKER);\n }\n for (const value of toCodePoints(source.slice(span.start, span.end))) cp.push(value);\n previous = span;\n }\n return cp;\n}\n\n/**\n * The span extents of the array as it stands. Recomputed after every rule, because applying\n * edits shifts every index after the first one — the markers themselves are guaranteed to\n * survive, since no edit may contain one.\n */\nexport function spanRangesOf(cp: readonly number[]): SpanRange[] {\n const ranges: SpanRange[] = [];\n let first = 0;\n for (let i = 0; i < cp.length; i += 1) {\n if (isMarker(cp[i] as number)) {\n ranges.push({ first, last: i - 1 });\n first = i + 1;\n }\n }\n ranges.push({ first, last: cp.length - 1 });\n return ranges;\n}\n\nfunction spanContaining(ranges: readonly SpanRange[], p: number): SpanRange | undefined {\n for (const range of ranges) {\n if (p >= range.first && p <= range.last + 1) return range;\n }\n return undefined;\n}\n\n/**\n * modes.md 3.4, two safety nets, both pure functions of `(p, q, r, s₀, s₁)` — so the verdict is\n * identical on every run, which is what makes redistribution deterministic (modes.md 5, point 3).\n *\n * 1. **No edit may contain a marker.** No rule can produce one; one that does is a bug, and the\n * edit is discarded rather than treated as a redistribution question.\n * 2. **The edge-growth rule.** An edit is discarded if it would place code points at an\n * extremity of its span that were not there before: with `d = q - p + 1` the replaced length\n * and `r` the replacement length, discard when `p = s₀ and r > d`, or `q = s₁ and r > d`.\n *\n * The length test is the whole rule and needs no knowledge of Markdown or HTML syntax. It\n * separates exactly the cases that matter: `\"` → `“` at an edge is 1 → 1 and applies; `--` →\n * `␣–␣` at an edge is 2 → 3 and is discarded, while the same edit interior to a span applies;\n * `(c)` → `©` is 3 → 1 and applies, because shrinking is always safe. An insertion has `d = 0`,\n * so it is discarded exactly when its position coincides with a span edge — the rule this one\n * generalises.\n *\n * **Deletion at an edge is not restricted here**, and must not be: `r > d` is false for a\n * deletion, so this filter never sees one. Deletion at an edge is governed by modes.md 3.3's\n * *Edge tests* clause instead — where a rule asks \"am I at the edge of the text I am allowed to\n * modify\" rather than \"what character is here\", the marker behaves as `NONE`. That clause is the\n * one place a rule may treat a span edge as the end of the text, it exists because `spaces` is\n * the only rule that deletes, and 3.3 states the division: **a rule that deletes must treat a\n * span edge as the end of the text; a rule that replaces or inserts must not, and is governed by\n * this filter.** The single test that claims the clause is `spaces.md` 3.2 step 4.\n */\nexport function filterBoundaryEdits(\n cp: readonly number[],\n edits: readonly Edit[],\n ranges: readonly SpanRange[],\n): Edit[] {\n const out: Edit[] = [];\n for (const edit of edits) {\n let containsMarker = false;\n for (let i = edit.start; i < edit.end; i += 1) {\n if (isMarker(cp[i] as number)) {\n containsMarker = true;\n break;\n }\n }\n if (containsMarker) continue;\n\n // `[start, end)` in the engine's half-open form is `cp[p … q]` with p = start, q = end - 1;\n // an insertion is `q = p - 1`, which falls out of the same expression.\n const p = edit.start;\n const q = edit.end - 1;\n const d = edit.end - edit.start;\n const r = edit.replacement.length;\n const span = spanContaining(ranges, p);\n if (span !== undefined && (p === span.first || q === span.last)) {\n if (r > d) continue;\n // The character clause. `r > d` is not the rule, only a formalisation of it that misses\n // `r === d`: `dashes` P3 admits a run of THREE dashes, so `---` -> `␣–␣` is 3 -> 3 and the\n // length test sees nothing while U+0020 lands on both extremities anyway. Testing the\n // character catches it, and only U+0020 needs testing — it is the one code point any rule\n // emits whose meaning comes from its position rather than from itself (modes.md 5 item 2).\n const first = edit.replacement[0];\n const last = edit.replacement[r - 1];\n if (r > 0 && p === span.first && first === SPACE && cp[p] !== SPACE) continue;\n if (r > 0 && q === span.last && last === SPACE && cp[q] !== SPACE) continue;\n }\n\n out.push(edit);\n }\n return out;\n}\n\n/** Redistribute the transformed array back to one piece per span (modes.md 3.5 step 4). */\nexport function splitOnMarker(cp: readonly number[], expected: number): number[][] {\n const pieces: number[][] = [[]];\n for (const value of cp) {\n if (isMarker(value)) {\n pieces.push([]);\n continue;\n }\n (pieces[pieces.length - 1] as number[]).push(value);\n }\n if (pieces.length !== expected) {\n throw new PolytypoError(\n \"POLYTYPO_RULE_CONTRACT\",\n `boundary markers did not survive the pipeline: expected ${expected} spans, found ${pieces.length}`,\n );\n }\n return pieces;\n}\n\n/**\n * The origin map for `concatenateSpans` (analyze.md §2): for every code point of the joined\n * array, the code-point offset of the character it came from IN THE DOCUMENT, and NO_ORIGIN for\n * the markers, which came from nowhere. `Span` bounds are native string indices, so the\n * conversion to code-point offsets happens here and not in the caller — this is the one place\n * that knows both coordinate systems.\n */\nexport function originOfSpans(source: string, spans: readonly Span[]): number[] {\n const codePointIndexOf = new Array<number>(source.length + 1);\n let cpIndex = 0;\n for (let i = 0; i < source.length;) {\n codePointIndexOf[i] = cpIndex;\n const code = source.codePointAt(i) as number;\n const width = code > 0xffff ? 2 : 1;\n if (width === 2) codePointIndexOf[i + 1] = cpIndex;\n i += width;\n cpIndex += 1;\n }\n codePointIndexOf[source.length] = cpIndex;\n\n const origin: number[] = [];\n let previous: Span | undefined;\n for (const span of spans) {\n if (previous !== undefined) origin.push(NO_ORIGIN);\n const base = codePointIndexOf[span.start] as number;\n const length = toCodePoints(source.slice(span.start, span.end)).length;\n for (let k = 0; k < length; k += 1) origin.push(base + k);\n previous = span;\n }\n return origin;\n}\n","import {\n concatenateSpans,\n filterBoundaryEdits,\n normalizeSpans,\n originOfSpans,\n spanRangesOf,\n splitOnMarker,\n type Span,\n} from \"../modes/spans.js\";\nimport { RULES } from \"../rules/registry.js\";\nimport type { LocaleData, Mode, RuleId } from \"../types.js\";\nimport { applyEdits } from \"./edits.js\";\nimport { fromCodePoints, toCodePoints } from \"./codepoints.js\";\nimport type { Change } from \"./origin.js\";\nimport { runRulesRecording } from \"./rule-runner.js\";\n\n/**\n * The same sequence as `runRules` (`./rule-runner.js`), with the two boundary filters of\n * modes.md 3.4 interposed. The span extents are recomputed after every rule, because applying\n * an edit shifts every index after it; the markers themselves always survive, since no edit may\n * contain one.\n */\nfunction runRulesOverSpans(\n cp: readonly number[],\n planned: readonly RuleId[],\n locale: LocaleData,\n mode: Mode,\n narrowTarget: number,\n): readonly number[] {\n let current = cp;\n for (const id of planned) {\n const rule = RULES[id];\n if (rule === undefined) continue;\n const edits = rule.apply({ cp: current, locale, mode, narrowTarget });\n current = applyEdits(current, filterBoundaryEdits(current, edits, spanRangesOf(current)), id);\n }\n return current;\n}\n\n/**\n * modes.md 3.5. The pipeline runs **once**, over the marker-separated concatenation of every\n * processable span — not per span, which would pair quotation marks in isolation, and not over a\n * naive concatenation, which would manufacture adjacencies the document does not have.\n *\n * The output is the input with a set of disjoint substring replacements applied and nothing else\n * (modes.md 4). A span whose content the rules did not change contributes no replacement, so a\n * document needing no changes comes back byte-identical; the parser located the spans and was\n * then discarded, and the document is never serialised.\n */\nexport function runOverSpans(\n source: string,\n spans: readonly Span[],\n planned: readonly RuleId[],\n locale: LocaleData,\n mode: Mode,\n narrowTarget: number,\n): string {\n const normalized = normalizeSpans(spans);\n if (normalized.length === 0) return source;\n\n const transformed = runRulesOverSpans(\n concatenateSpans(source, normalized),\n planned,\n locale,\n mode,\n narrowTarget,\n );\n const pieces = splitOnMarker(transformed, normalized.length);\n\n let out = \"\";\n let cursor = 0;\n for (let i = 0; i < normalized.length; i += 1) {\n const span = normalized[i] as Span;\n const replacement = fromCodePoints(pieces[i] as number[]);\n const original = source.slice(span.start, span.end);\n out += source.slice(cursor, span.start);\n out += replacement === original ? original : replacement;\n cursor = span.end;\n }\n return out + source.slice(cursor);\n}\n\n/**\n * `runOverSpans`, reporting instead of applying (analyze.md §1). The span table supplies the\n * origin map, so every change comes back in DOCUMENT coordinates — analyze.md §6 names a\n * runtime that reports span-local offsets here as the mistake that passes every text-mode test.\n */\nexport function analyzeOverSpans(\n source: string,\n spans: readonly Span[],\n planned: readonly RuleId[],\n locale: LocaleData,\n mode: Mode,\n narrowTarget: number,\n): Change[] {\n const normalized = normalizeSpans(spans);\n if (normalized.length === 0) return [];\n return runRulesRecording(\n concatenateSpans(source, normalized),\n planned,\n locale,\n mode,\n narrowTarget,\n originOfSpans(source, normalized),\n toCodePoints(source).length,\n (current, edits) => filterBoundaryEdits(current, edits, spanRangesOf(current)),\n );\n}\n"]}
|
|
@@ -2,17 +2,17 @@
|
|
|
2
2
|
|
|
3
3
|
|
|
4
4
|
|
|
5
|
-
var
|
|
5
|
+
var _chunkGJJNGBRHcjs = require('./chunk-GJJNGBRH.cjs');
|
|
6
6
|
|
|
7
7
|
|
|
8
8
|
|
|
9
|
-
var
|
|
9
|
+
var _chunkDCC5OQETcjs = require('./chunk-DCC5OQET.cjs');
|
|
10
10
|
|
|
11
11
|
|
|
12
12
|
|
|
13
13
|
|
|
14
14
|
|
|
15
|
-
var
|
|
15
|
+
var _chunkPJ6HB77Gcjs = require('./chunk-PJ6HB77G.cjs');
|
|
16
16
|
|
|
17
17
|
// src/modes/markdown.ts
|
|
18
18
|
var _micromark = require('micromark');
|
|
@@ -90,7 +90,7 @@ function updateElementStack(stack, tag, caseSensitive) {
|
|
|
90
90
|
if (stack[stack.length - 1] === name) stack.pop();
|
|
91
91
|
return;
|
|
92
92
|
}
|
|
93
|
-
if (!parsed.selfClosing &&
|
|
93
|
+
if (!parsed.selfClosing && _chunkGJJNGBRHcjs.isSkippedElement.call(void 0, name)) stack.push(name);
|
|
94
94
|
}
|
|
95
95
|
function tokenize(source, extensions) {
|
|
96
96
|
return _micromark.postprocess.call(void 0,
|
|
@@ -100,12 +100,12 @@ function tokenize(source, extensions) {
|
|
|
100
100
|
function resolveDialect(dialect) {
|
|
101
101
|
if (dialect === "commonmark" || dialect === "mdx") return dialect;
|
|
102
102
|
if (dialect === void 0) {
|
|
103
|
-
throw new (0,
|
|
103
|
+
throw new (0, _chunkPJ6HB77Gcjs.PolytypoError)(
|
|
104
104
|
"POLYTYPO_INVALID_DIALECT",
|
|
105
105
|
'Mode "markdown" requires a dialect. Expected "commonmark" or "mdx"; there is no default.'
|
|
106
106
|
);
|
|
107
107
|
}
|
|
108
|
-
throw new (0,
|
|
108
|
+
throw new (0, _chunkPJ6HB77Gcjs.PolytypoError)(
|
|
109
109
|
"POLYTYPO_INVALID_DIALECT",
|
|
110
110
|
`Unknown dialect "${String(dialect)}". Expected "commonmark" or "mdx".`
|
|
111
111
|
);
|
|
@@ -116,7 +116,7 @@ function extensionsFor(dialect) {
|
|
|
116
116
|
return dialect === "mdx" ? [fm, _micromarkextensiongfm.gfm.call(void 0, ), _micromarkextensionmdxjs.mdxjs.call(void 0, )] : [fm, _micromarkextensiongfm.gfm.call(void 0, )];
|
|
117
117
|
}
|
|
118
118
|
function markdownSpans(source, dialect) {
|
|
119
|
-
const events =
|
|
119
|
+
const events = _chunkGJJNGBRHcjs.wrapParserErrors.call(void 0, dialect, () => tokenize(source, extensionsFor(dialect)));
|
|
120
120
|
const spans = [];
|
|
121
121
|
const elementStack = [];
|
|
122
122
|
let skipDepth = 0;
|
|
@@ -139,7 +139,7 @@ function markdownSpans(source, dialect) {
|
|
|
139
139
|
}
|
|
140
140
|
if (type === "htmlFlow") {
|
|
141
141
|
if (kind === "enter" && skipDepth === 0 && elementStack.length === 0) {
|
|
142
|
-
for (const span of
|
|
142
|
+
for (const span of _chunkGJJNGBRHcjs.htmlFragmentSpans.call(void 0,
|
|
143
143
|
source.slice(token.start.offset, token.end.offset),
|
|
144
144
|
token.start.offset
|
|
145
145
|
)) {
|
|
@@ -158,19 +158,19 @@ function markdownSpans(source, dialect) {
|
|
|
158
158
|
|
|
159
159
|
// src/engine/markdown-pipeline.ts
|
|
160
160
|
function runMarkdownPipeline(input, options) {
|
|
161
|
-
const narrowTarget =
|
|
162
|
-
const planned =
|
|
163
|
-
const locale =
|
|
161
|
+
const narrowTarget = _chunkPJ6HB77Gcjs.resolveNarrowTarget.call(void 0, options.narrowNbsp);
|
|
162
|
+
const planned = _chunkPJ6HB77Gcjs.planRules.call(void 0, options.rules);
|
|
163
|
+
const locale = _chunkPJ6HB77Gcjs.getLocaleData.call(void 0, options.locale);
|
|
164
164
|
const dialect = resolveDialect(options.dialect);
|
|
165
165
|
const spans = markdownSpans(input, dialect);
|
|
166
|
-
return
|
|
166
|
+
return _chunkDCC5OQETcjs.runOverSpans.call(void 0, input, spans, planned, locale, "markdown", narrowTarget);
|
|
167
167
|
}
|
|
168
168
|
function analyzeMarkdownPipeline(input, options) {
|
|
169
|
-
const narrowTarget =
|
|
170
|
-
const planned =
|
|
171
|
-
const locale =
|
|
169
|
+
const narrowTarget = _chunkPJ6HB77Gcjs.resolveNarrowTarget.call(void 0, options.narrowNbsp);
|
|
170
|
+
const planned = _chunkPJ6HB77Gcjs.planRules.call(void 0, options.rules);
|
|
171
|
+
const locale = _chunkPJ6HB77Gcjs.getLocaleData.call(void 0, options.locale);
|
|
172
172
|
const dialect = resolveDialect(options.dialect);
|
|
173
|
-
return
|
|
173
|
+
return _chunkDCC5OQETcjs.analyzeOverSpans.call(void 0,
|
|
174
174
|
input,
|
|
175
175
|
markdownSpans(input, dialect),
|
|
176
176
|
planned,
|
|
@@ -184,4 +184,4 @@ function analyzeMarkdownPipeline(input, options) {
|
|
|
184
184
|
|
|
185
185
|
|
|
186
186
|
exports.runMarkdownPipeline = runMarkdownPipeline; exports.analyzeMarkdownPipeline = analyzeMarkdownPipeline;
|
|
187
|
-
//# sourceMappingURL=chunk-
|
|
187
|
+
//# sourceMappingURL=chunk-FZ6YRCGR.cjs.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-KRY73ZPE.cjs","../src/modes/markdown.ts","../src/engine/markdown-pipeline.ts"],"names":[],"mappings":"AAAA;AACE;AACA;AACA;AACF,wDAA6B;AAC7B;AACE;AACA;AACF,wDAA6B;AAC7B;AACE;AACA;AACA;AACA;AACF,wDAA6B;AAC7B;AACA;AChBA,sCAA+C;AAC/C,gFAA4B;AAC5B,gEAAoB;AACpB,oEAAsB;AAgBtB,IAAM,oBAAA,kBAA2C,IAAI,GAAA,CAAI;AAAA;AAAA,EAEvD,YAAA;AAAA,EACA,cAAA;AAAA;AAAA,EAEA,UAAA;AAAA;AAAA,EAEA,UAAA;AAAA,EACA,iBAAA;AAAA;AAAA;AAAA,EAGA,UAAA;AAAA,EACA,WAAA;AAAA,EACA,YAAA;AAAA;AAAA;AAAA;AAAA,EAIA,mBAAA;AAAA,EACA,mBAAA;AAAA,EACA,UAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAMA;AACF,CAAC,CAAA;AAGD,IAAM,oBAAA,kBAA2C,IAAI,GAAA,CAAI;AAAA,EACvD,UAAA;AAAA,EACA,eAAA;AAAA,EACA;AACF,CAAC,CAAA;AAED,IAAM,UAAA,EAAY,EAAA;AAClB,IAAM,aAAA,EAAe,EAAA;AACrB,IAAM,MAAA,EAAQ,EAAA;AACd,IAAM,YAAA,EAAc,EAAA;AACpB,IAAM,SAAA,EAAW,EAAA;AAEjB,SAAS,aAAA,CAAc,IAAA,EAAuB;AAC5C,EAAA,OAAQ,KAAA,GAAQ,GAAA,GAAQ,KAAA,GAAQ,IAAA,GAAU,KAAA,GAAQ,GAAA,GAAQ,KAAA,GAAQ,EAAA;AACpE;AAEA,SAAS,aAAA,CAAc,IAAA,EAAuB;AAI5C,EAAA,OACE,KAAA,IAAS,aAAA,GACT,KAAA,IAAS,MAAA,GACT,KAAA,IAAS,GAAA,GACT,KAAA,IAAS,EAAA,GACT,KAAA,IAAS,GAAA,GACT,KAAA,IAAS,EAAA;AAEb;AAGA,SAAS,UAAA,CAAW,KAAA,EAAuB;AACzC,EAAA,IAAI,IAAA,EAAM,EAAA;AACV,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,CAAA,EAAG,EAAA,EAAI,KAAA,CAAM,MAAA,EAAQ,EAAA,GAAK,CAAA,EAAG;AACxC,IAAA,MAAM,KAAA,EAAO,KAAA,CAAM,UAAA,CAAW,CAAC,CAAA;AAC/B,IAAA,IAAA,GAAO,KAAA,GAAQ,GAAA,GAAQ,KAAA,GAAQ,GAAA,EAAO,MAAA,CAAO,YAAA,CAAa,KAAA,EAAO,EAAI,EAAA,EAAI,KAAA,CAAM,CAAC,CAAA;AAAA,EAClF;AACA,EAAA,OAAO,GAAA;AACT;AAQA,SAAS,OAAA,CAAQ,GAAA,EAA4B;AAC3C,EAAA,GAAA,CAAI,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,SAAA,EAAW,OAAO,IAAA;AAC5C,EAAA,IAAI,EAAA,EAAI,CAAA;AACR,EAAA,MAAM,QAAA,EAAU,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,KAAA;AACtC,EAAA,GAAA,CAAI,OAAA,EAAS,EAAA,GAAK,CAAA;AAElB,EAAA,GAAA,CAAI,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,YAAA,GAAe,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,QAAA,EAAU,OAAO,IAAA;AAChF,EAAA,GAAA,CAAI,CAAC,aAAA,CAAc,GAAA,CAAI,UAAA,CAAW,CAAC,CAAC,CAAA,EAAG,OAAO,IAAA;AAC9C,EAAA,MAAM,OAAA,EAAS,CAAA;AACf,EAAA,MAAA,CAAO,EAAA,EAAI,GAAA,CAAI,OAAA,GAAU,aAAA,CAAc,GAAA,CAAI,UAAA,CAAW,CAAC,CAAC,CAAA,EAAG,EAAA,GAAK,CAAA;AAChE,EAAA,MAAM,YAAA,EAAc,GAAA,CAAI,UAAA,CAAW,GAAA,CAAI,OAAA,EAAS,CAAC,EAAA,IAAM,KAAA;AACvD,EAAA,OAAO,EAAE,IAAA,EAAM,GAAA,CAAI,KAAA,CAAM,MAAA,EAAQ,CAAC,CAAA,EAAG,OAAA,EAAS,YAAY,CAAA;AAC5D;AAQA,SAAS,kBAAA,CAAmB,KAAA,EAAiB,GAAA,EAAa,aAAA,EAA8B;AACtF,EAAA,MAAM,OAAA,EAAS,OAAA,CAAQ,GAAG,CAAA;AAC1B,EAAA,GAAA,CAAI,OAAA,IAAW,IAAA,EAAM,MAAA;AACrB,EAAA,MAAM,KAAA,EAAO,cAAA,EAAgB,MAAA,CAAO,KAAA,EAAO,UAAA,CAAW,MAAA,CAAO,IAAI,CAAA;AACjE,EAAA,GAAA,CAAI,MAAA,CAAO,OAAA,EAAS;AAClB,IAAA,GAAA,CAAI,KAAA,CAAM,KAAA,CAAM,OAAA,EAAS,CAAC,EAAA,IAAM,IAAA,EAAM,KAAA,CAAM,GAAA,CAAI,CAAA;AAChD,IAAA,MAAA;AAAA,EACF;AACA,EAAA,GAAA,CAAI,CAAC,MAAA,CAAO,YAAA,GAAe,gDAAA,IAAqB,CAAA,EAAG,KAAA,CAAM,IAAA,CAAK,IAAI,CAAA;AACpE;AAEA,SAAS,QAAA,CAAS,MAAA,EAAgB,UAAA,EAA2C;AAC3E,EAAA,OAAO,oCAAA;AAAA,IACL,8BAAA,EAAQ,UAAA,EAAY,CAAC,GAAG,UAAU,EAAE,CAAC,CAAA,CAClC,QAAA,CAAS,CAAA,CACT,KAAA,CAAM,mCAAA,CAAW,CAAE,MAAA,EAAQ,IAAA,EAAM,IAAI,CAAC;AAAA,EAC3C,CAAA;AACF;AAQO,SAAS,cAAA,CAAe,OAAA,EAAuC;AACpE,EAAA,GAAA,CAAI,QAAA,IAAY,aAAA,GAAgB,QAAA,IAAY,KAAA,EAAO,OAAO,OAAA;AAC1D,EAAA,GAAA,CAAI,QAAA,IAAY,KAAA,CAAA,EAAW;AACzB,IAAA,MAAM,IAAI,oCAAA;AAAA,MACR,0BAAA;AAAA,MACA;AAAA,IACF,CAAA;AAAA,EACF;AACA,EAAA,MAAM,IAAI,oCAAA;AAAA,IACR,0BAAA;AAAA,IACA,CAAA,iBAAA,EAAoB,MAAA,CAAO,OAAO,CAAC,CAAA,kCAAA;AAAA,EACrC,CAAA;AACF;AAKA,IAAM,oBAAA,EAAsB,CAAC,MAAA,EAAQ,MAAM,CAAA;AAE3C,SAAS,aAAA,CAAc,OAAA,EAA+B;AACpD,EAAA,MAAM,GAAA,EAAK,wDAAA,CAAa,GAAG,mBAAmB,CAAC,CAAA;AAC/C,EAAA,OAAO,QAAA,IAAY,MAAA,EAAQ,CAAC,EAAA,EAAI,wCAAA,CAAI,EAAG,4CAAA,CAAO,EAAA,EAAI,CAAC,EAAA,EAAI,wCAAA,CAAK,CAAA;AAC9D;AAEO,SAAS,aAAA,CAAc,MAAA,EAAgB,OAAA,EAA0B;AACtE,EAAA,MAAM,OAAA,EAAS,gDAAA,OAAiB,EAAS,CAAA,EAAA,GAAM,QAAA,CAAS,MAAA,EAAQ,aAAA,CAAc,OAAO,CAAC,CAAC,CAAA;AACvF,EAAA,MAAM,MAAA,EAAgB,CAAC,CAAA;AACvB,EAAA,MAAM,aAAA,EAAyB,CAAC,CAAA;AAChC,EAAA,IAAI,UAAA,EAAY,CAAA;AAEhB,EAAA,IAAA,CAAA,MAAW,CAAC,IAAA,EAAM,KAAK,EAAA,GAAK,MAAA,EAAQ;AAClC,IAAA,MAAM,KAAA,EAAO,KAAA,CAAM,IAAA;AAEnB,IAAA,GAAA,CAAI,mBAAA,CAAoB,GAAA,CAAI,IAAI,CAAA,EAAG;AACjC,MAAA,UAAA,GAAa,KAAA,IAAS,QAAA,EAAU,EAAA,EAAI,CAAA,CAAA;AACpC,MAAA,QAAA;AAAA,IACF;AAEA,IAAA,GAAA,CAAI,mBAAA,CAAoB,GAAA,CAAI,IAAI,CAAA,EAAG;AACjC,MAAA,GAAA,CAAI,KAAA,IAAS,OAAA,EAAS;AACpB,QAAA,kBAAA;AAAA,UACE,YAAA;AAAA,UACA,MAAA,CAAO,KAAA,CAAM,KAAA,CAAM,KAAA,CAAM,MAAA,EAAQ,KAAA,CAAM,GAAA,CAAI,MAAM,CAAA;AAAA,UACjD,KAAA,IAAS;AAAA,QACX,CAAA;AAAA,MACF;AACA,MAAA,UAAA,GAAa,KAAA,IAAS,QAAA,EAAU,EAAA,EAAI,CAAA,CAAA;AACpC,MAAA,QAAA;AAAA,IACF;AAIA,IAAA,GAAA,CAAI,KAAA,IAAS,UAAA,EAAY;AACvB,MAAA,GAAA,CAAI,KAAA,IAAS,QAAA,GAAW,UAAA,IAAc,EAAA,GAAK,YAAA,CAAa,OAAA,IAAW,CAAA,EAAG;AACpE,QAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,iDAAA;AAAA,UACjB,MAAA,CAAO,KAAA,CAAM,KAAA,CAAM,KAAA,CAAM,MAAA,EAAQ,KAAA,CAAM,GAAA,CAAI,MAAM,CAAA;AAAA,UACjD,KAAA,CAAM,KAAA,CAAM;AAAA,QACd,CAAA,EAAG;AACD,UAAA,KAAA,CAAM,IAAA,CAAK,IAAI,CAAA;AAAA,QACjB;AAAA,MACF;AACA,MAAA,UAAA,GAAa,KAAA,IAAS,QAAA,EAAU,EAAA,EAAI,CAAA,CAAA;AACpC,MAAA,QAAA;AAAA,IACF;AAEA,IAAA,GAAA,CAAI,KAAA,IAAS,QAAA,GAAW,KAAA,IAAS,OAAA,GAAU,UAAA,IAAc,EAAA,GAAK,YAAA,CAAa,OAAA,IAAW,CAAA,EAAG;AACvF,MAAA,KAAA,CAAM,IAAA,CAAK,EAAE,KAAA,EAAO,KAAA,CAAM,KAAA,CAAM,MAAA,EAAQ,GAAA,EAAK,KAAA,CAAM,GAAA,CAAI,OAAO,CAAC,CAAA;AAAA,IACjE;AAAA,EACF;AAEA,EAAA,OAAO,KAAA;AACT;ADrDA;AACA;AE5IO,SAAS,mBAAA,CAAoB,KAAA,EAAe,OAAA,EAAmC;AACpF,EAAA,MAAM,aAAA,EAAe,mDAAA,OAAoB,CAAQ,UAAU,CAAA;AAC3D,EAAA,MAAM,QAAA,EAAU,yCAAA,OAAU,CAAQ,KAAK,CAAA;AACvC,EAAA,MAAM,OAAA,EAAS,6CAAA,OAAc,CAAQ,MAAM,CAAA;AAC3C,EAAA,MAAM,QAAA,EAAU,cAAA,CAAe,OAAA,CAAQ,OAAO,CAAA;AAC9C,EAAA,MAAM,MAAA,EAAQ,aAAA,CAAc,KAAA,EAAO,OAAO,CAAA;AAC1C,EAAA,OAAO,4CAAA,KAAa,EAAO,KAAA,EAAO,OAAA,EAAS,MAAA,EAAQ,UAAA,EAAY,YAAY,CAAA;AAC7E;AAKO,SAAS,uBAAA,CAAwB,KAAA,EAAe,OAAA,EAAqC;AAC1F,EAAA,MAAM,aAAA,EAAe,mDAAA,OAAoB,CAAQ,UAAU,CAAA;AAC3D,EAAA,MAAM,QAAA,EAAU,yCAAA,OAAU,CAAQ,KAAK,CAAA;AACvC,EAAA,MAAM,OAAA,EAAS,6CAAA,OAAc,CAAQ,MAAM,CAAA;AAC3C,EAAA,MAAM,QAAA,EAAU,cAAA,CAAe,OAAA,CAAQ,OAAO,CAAA;AAC9C,EAAA,OAAO,gDAAA;AAAA,IACL,KAAA;AAAA,IACA,aAAA,CAAc,KAAA,EAAO,OAAO,CAAA;AAAA,IAC5B,OAAA;AAAA,IACA,MAAA;AAAA,IACA,UAAA;AAAA,IACA;AAAA,EACF,CAAA;AACF;AF0IA;AACA;AACE;AACA;AACF,6GAAC","file":"/home/runner/work/polytypo-js/polytypo-js/dist/chunk-KRY73ZPE.cjs","sourcesContent":[null,"import { parse, postprocess, preprocess } from \"micromark\";\nimport { frontmatter } from \"micromark-extension-frontmatter\";\nimport { gfm } from \"micromark-extension-gfm\";\nimport { mdxjs } from \"micromark-extension-mdxjs\";\nimport type { Event, Extension } from \"micromark-util-types\";\nimport { PolytypoError } from \"../errors.js\";\nimport type { Dialect } from \"../types.js\";\nimport { htmlFragmentSpans, isSkippedElement } from \"./html.js\";\nimport { wrapParserErrors } from \"./parse-error.js\";\nimport type { Span } from \"./spans.js\";\n\n/**\n * spec/rules/modes.md 3.7. Each entry is skipped **whole**, including anything that looks\n * processable inside it. The complement is not enumerated: a span is emitted only for a\n * micromark `data` token, which is the tokenizer's own name for \"literal content\", so list\n * markers, emphasis delimiters, heading sequences, table padding, line endings, hard-break\n * spaces, character escapes and character references are outside every span by construction\n * rather than by a second list that could drift from the first.\n */\nconst SKIPPED_TOKEN_TYPES: ReadonlySet<string> = new Set([\n // fenced code blocks (including the info string and the fences) and indented code blocks\n \"codeFenced\",\n \"codeIndented\",\n // inline code spans, including the backticks\n \"codeText\",\n // autolinks `<https://…>`, and GFM's bare-URL form\n \"autolink\",\n \"literalAutolink\",\n // link and image destinations and titles; the definition line of a reference link. The link\n // *text* is a `labelText`, which is not skipped.\n \"resource\",\n \"reference\",\n \"definition\",\n // MDX: expression containers in full, and every JSX attribute — the whole tag is skipped, so\n // attributes never surface. JSX element *children* sit outside the tag tokens and are\n // processable.\n \"mdxFlowExpression\",\n \"mdxTextExpression\",\n \"mdxjsEsm\",\n // modes.md 3.7.3, and it matters: without it the second `---` of a YAML block reads as a setext\n // underline, `title: Une note` becomes a paragraph, and `fr` puts a narrow no-break space in\n // front of the colon of a machine-read metadata field. Covered by conformance fixtures\n // en-us-markdown-{commonmark,mdx}-frontmatter, fr-markdown-{commonmark,mdx}-frontmatter-nbsp,\n // en-us-markdown-commonmark-frontmatter-toml and -frontmatter-unterminated.\n \"frontmatter\",\n]);\n\n/** Token types whose enter/exit also maintains the skipped-element stack. */\nconst RAW_TAG_TOKEN_TYPES: ReadonlySet<string> = new Set([\n \"htmlText\",\n \"mdxJsxTextTag\",\n \"mdxJsxFlowTag\",\n]);\n\nconst LESS_THAN = 0x3c;\nconst GREATER_THAN = 0x3e;\nconst SLASH = 0x2f;\nconst EXCLAMATION = 0x21;\nconst QUESTION = 0x3f;\n\nfunction isAsciiLetter(unit: number): boolean {\n return (unit >= 0x61 && unit <= 0x7a) || (unit >= 0x41 && unit <= 0x5a);\n}\n\nfunction isTagNameUnit(unit: number): boolean {\n // Deliberately permissive: anything that is not a delimiter belongs to the name. Custom\n // elements (`my-callout`) and member expressions (`Foo.Bar`) must come out whole so that they\n // fail the skip-list membership test rather than being truncated into something that passes.\n return (\n unit !== GREATER_THAN &&\n unit !== SLASH &&\n unit !== 0x20 &&\n unit !== 0x09 &&\n unit !== 0x0a &&\n unit !== 0x0d\n );\n}\n\n/** ASCII-only, because HTML tag names are ASCII case-insensitive and `toLowerCase()` is not. */\nfunction asciiLower(value: string): string {\n let out = \"\";\n for (let i = 0; i < value.length; i += 1) {\n const unit = value.charCodeAt(i);\n out += unit >= 0x41 && unit <= 0x5a ? String.fromCharCode(unit + 0x20) : value[i];\n }\n return out;\n}\n\ninterface RawTag {\n readonly name: string;\n readonly closing: boolean;\n readonly selfClosing: boolean;\n}\n\nfunction readTag(tag: string): RawTag | null {\n if (tag.charCodeAt(0) !== LESS_THAN) return null;\n let i = 1;\n const closing = tag.charCodeAt(i) === SLASH;\n if (closing) i += 1;\n // Comments, declarations and processing instructions carry no element name.\n if (tag.charCodeAt(i) === EXCLAMATION || tag.charCodeAt(i) === QUESTION) return null;\n if (!isAsciiLetter(tag.charCodeAt(i))) return null;\n const nameAt = i;\n while (i < tag.length && isTagNameUnit(tag.charCodeAt(i))) i += 1;\n const selfClosing = tag.charCodeAt(tag.length - 2) === SLASH;\n return { name: tag.slice(nameAt, i), closing, selfClosing };\n}\n\n/**\n * modes.md 3.7: raw HTML is \"handed to the html skip list of 3.6\". Inline raw HTML reaches the\n * tokenizer as isolated tags with ordinary markdown content between them, so honouring the\n * subtree rule means tracking which skipped element is currently open. JSX names are compared\n * case-sensitively (`<Code>` is a component, `<code>` is an element); HTML names are not.\n */\nfunction updateElementStack(stack: string[], tag: string, caseSensitive: boolean): void {\n const parsed = readTag(tag);\n if (parsed === null) return;\n const name = caseSensitive ? parsed.name : asciiLower(parsed.name);\n if (parsed.closing) {\n if (stack[stack.length - 1] === name) stack.pop();\n return;\n }\n if (!parsed.selfClosing && isSkippedElement(name)) stack.push(name);\n}\n\nfunction tokenize(source: string, extensions: readonly Extension[]): Event[] {\n return postprocess(\n parse({ extensions: [...extensions] })\n .document()\n .write(preprocess()(source, null, true)),\n );\n}\n\n/**\n * modes.md 3.7.1. The dialect is the caller's, never detected: `\"commonmark\"` is CommonMark 0.31\n * plus GFM, `\"mdx\"` is the same minus indented code blocks and `<…>` autolinks, plus JSX and\n * `{…}` expression containers. Both enable frontmatter and GFM, which 3.7.2 makes normative —\n * §4's promise that table alignment rows survive is empty unless tables are recognised at all.\n */\nexport function resolveDialect(dialect: Dialect | undefined): Dialect {\n if (dialect === \"commonmark\" || dialect === \"mdx\") return dialect;\n if (dialect === undefined) {\n throw new PolytypoError(\n \"POLYTYPO_INVALID_DIALECT\",\n 'Mode \"markdown\" requires a dialect. Expected \"commonmark\" or \"mdx\"; there is no default.',\n );\n }\n throw new PolytypoError(\n \"POLYTYPO_INVALID_DIALECT\",\n `Unknown dialect \"${String(dialect)}\". Expected \"commonmark\" or \"mdx\".`,\n );\n}\n\n// modes.md 3.7.3 names both frontmatter delimiters, `---` (YAML) and `+++` (TOML), and skips\n// either whole. micromark's default matter is YAML alone, so both must be asked for by name: with\n// the default, a `+++` block parses as prose and the machine-read fields inside it get typeset.\nconst FRONTMATTER_MATTERS = [\"yaml\", \"toml\"] as const;\n\nfunction extensionsFor(dialect: Dialect): Extension[] {\n const fm = frontmatter([...FRONTMATTER_MATTERS]);\n return dialect === \"mdx\" ? [fm, gfm(), mdxjs()] : [fm, gfm()];\n}\n\nexport function markdownSpans(source: string, dialect: Dialect): Span[] {\n const events = wrapParserErrors(dialect, () => tokenize(source, extensionsFor(dialect)));\n const spans: Span[] = [];\n const elementStack: string[] = [];\n let skipDepth = 0;\n\n for (const [kind, token] of events) {\n const type = token.type;\n\n if (SKIPPED_TOKEN_TYPES.has(type)) {\n skipDepth += kind === \"enter\" ? 1 : -1;\n continue;\n }\n\n if (RAW_TAG_TOKEN_TYPES.has(type)) {\n if (kind === \"enter\") {\n updateElementStack(\n elementStack,\n source.slice(token.start.offset, token.end.offset),\n type !== \"htmlText\",\n );\n }\n skipDepth += kind === \"enter\" ? 1 : -1;\n continue;\n }\n\n // An HTML block is handed to the `html` extractor rather than skipped whole, so that the\n // prose inside `<div>…</div>` is typeset while the markup is not.\n if (type === \"htmlFlow\") {\n if (kind === \"enter\" && skipDepth === 0 && elementStack.length === 0) {\n for (const span of htmlFragmentSpans(\n source.slice(token.start.offset, token.end.offset),\n token.start.offset,\n )) {\n spans.push(span);\n }\n }\n skipDepth += kind === \"enter\" ? 1 : -1;\n continue;\n }\n\n if (kind === \"enter\" && type === \"data\" && skipDepth === 0 && elementStack.length === 0) {\n spans.push({ start: token.start.offset, end: token.end.offset });\n }\n }\n\n return spans;\n}\n","import { markdownSpans, resolveDialect } from \"../modes/markdown.js\";\nimport type { Options } from \"../types.js\";\nimport { getLocaleData } from \"./locale.js\";\nimport { resolveNarrowTarget } from \"./narrow-target.js\";\nimport { planRules } from \"./rule-runner.js\";\nimport { analyzeOverSpans, runOverSpans } from \"./span-runner.js\";\nimport type { Change } from \"./origin.js\";\n\n/**\n * `markdown` mode only. Imports the Micromark/MDX stack and `parse5` (via `../modes/markdown.js`\n * importing `../modes/html.js` for normative embedded-HTML handling, modes.md 3.7) — both are\n * legitimately reachable from `polytypo/markdown` (AUDIT_REMEDIATION_AND_RELEASE_PLAN.md 5.1).\n *\n * Validation order — `planRules`, then `getLocaleData`, then `resolveDialect`/parsing — mirrors\n * the pre-Stage-5 aggregate `runPipeline` exactly: an unknown-rule error wins over an\n * unknown-locale error, which wins over a missing/invalid dialect, which wins over a parse\n * failure. All four are public, tested behaviour this refactor was not authorised to change.\n */\nexport function runMarkdownPipeline(input: string, options: Partial<Options>): string {\n const narrowTarget = resolveNarrowTarget(options.narrowNbsp);\n const planned = planRules(options.rules);\n const locale = getLocaleData(options.locale);\n const dialect = resolveDialect(options.dialect);\n const spans = markdownSpans(input, dialect);\n return runOverSpans(input, spans, planned, locale, \"markdown\", narrowTarget);\n}\n\n/** analyze.md §1, `markdown` mode. Dialect validation happens here exactly as it does for\n * `runMarkdownPipeline`, so an absent dialect throws POLYTYPO_INVALID_DIALECT from `analyze`\n * too (analyze.md §4 A1). */\nexport function analyzeMarkdownPipeline(input: string, options: Partial<Options>): Change[] {\n const narrowTarget = resolveNarrowTarget(options.narrowNbsp);\n const planned = planRules(options.rules);\n const locale = getLocaleData(options.locale);\n const dialect = resolveDialect(options.dialect);\n return analyzeOverSpans(\n input,\n markdownSpans(input, dialect),\n planned,\n locale,\n \"markdown\",\n narrowTarget,\n );\n}\n"]}
|
|
1
|
+
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-FZ6YRCGR.cjs","../src/modes/markdown.ts","../src/engine/markdown-pipeline.ts"],"names":[],"mappings":"AAAA;AACE;AACA;AACA;AACF,wDAA6B;AAC7B;AACE;AACA;AACF,wDAA6B;AAC7B;AACE;AACA;AACA;AACA;AACF,wDAA6B;AAC7B;AACA;AChBA,sCAA+C;AAC/C,gFAA4B;AAC5B,gEAAoB;AACpB,oEAAsB;AAgBtB,IAAM,oBAAA,kBAA2C,IAAI,GAAA,CAAI;AAAA;AAAA,EAEvD,YAAA;AAAA,EACA,cAAA;AAAA;AAAA,EAEA,UAAA;AAAA;AAAA,EAEA,UAAA;AAAA,EACA,iBAAA;AAAA;AAAA;AAAA,EAGA,UAAA;AAAA,EACA,WAAA;AAAA,EACA,YAAA;AAAA;AAAA;AAAA;AAAA,EAIA,mBAAA;AAAA,EACA,mBAAA;AAAA,EACA,UAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAMA;AACF,CAAC,CAAA;AAGD,IAAM,oBAAA,kBAA2C,IAAI,GAAA,CAAI;AAAA,EACvD,UAAA;AAAA,EACA,eAAA;AAAA,EACA;AACF,CAAC,CAAA;AAED,IAAM,UAAA,EAAY,EAAA;AAClB,IAAM,aAAA,EAAe,EAAA;AACrB,IAAM,MAAA,EAAQ,EAAA;AACd,IAAM,YAAA,EAAc,EAAA;AACpB,IAAM,SAAA,EAAW,EAAA;AAEjB,SAAS,aAAA,CAAc,IAAA,EAAuB;AAC5C,EAAA,OAAQ,KAAA,GAAQ,GAAA,GAAQ,KAAA,GAAQ,IAAA,GAAU,KAAA,GAAQ,GAAA,GAAQ,KAAA,GAAQ,EAAA;AACpE;AAEA,SAAS,aAAA,CAAc,IAAA,EAAuB;AAI5C,EAAA,OACE,KAAA,IAAS,aAAA,GACT,KAAA,IAAS,MAAA,GACT,KAAA,IAAS,GAAA,GACT,KAAA,IAAS,EAAA,GACT,KAAA,IAAS,GAAA,GACT,KAAA,IAAS,EAAA;AAEb;AAGA,SAAS,UAAA,CAAW,KAAA,EAAuB;AACzC,EAAA,IAAI,IAAA,EAAM,EAAA;AACV,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,CAAA,EAAG,EAAA,EAAI,KAAA,CAAM,MAAA,EAAQ,EAAA,GAAK,CAAA,EAAG;AACxC,IAAA,MAAM,KAAA,EAAO,KAAA,CAAM,UAAA,CAAW,CAAC,CAAA;AAC/B,IAAA,IAAA,GAAO,KAAA,GAAQ,GAAA,GAAQ,KAAA,GAAQ,GAAA,EAAO,MAAA,CAAO,YAAA,CAAa,KAAA,EAAO,EAAI,EAAA,EAAI,KAAA,CAAM,CAAC,CAAA;AAAA,EAClF;AACA,EAAA,OAAO,GAAA;AACT;AAQA,SAAS,OAAA,CAAQ,GAAA,EAA4B;AAC3C,EAAA,GAAA,CAAI,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,SAAA,EAAW,OAAO,IAAA;AAC5C,EAAA,IAAI,EAAA,EAAI,CAAA;AACR,EAAA,MAAM,QAAA,EAAU,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,KAAA;AACtC,EAAA,GAAA,CAAI,OAAA,EAAS,EAAA,GAAK,CAAA;AAElB,EAAA,GAAA,CAAI,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,YAAA,GAAe,GAAA,CAAI,UAAA,CAAW,CAAC,EAAA,IAAM,QAAA,EAAU,OAAO,IAAA;AAChF,EAAA,GAAA,CAAI,CAAC,aAAA,CAAc,GAAA,CAAI,UAAA,CAAW,CAAC,CAAC,CAAA,EAAG,OAAO,IAAA;AAC9C,EAAA,MAAM,OAAA,EAAS,CAAA;AACf,EAAA,MAAA,CAAO,EAAA,EAAI,GAAA,CAAI,OAAA,GAAU,aAAA,CAAc,GAAA,CAAI,UAAA,CAAW,CAAC,CAAC,CAAA,EAAG,EAAA,GAAK,CAAA;AAChE,EAAA,MAAM,YAAA,EAAc,GAAA,CAAI,UAAA,CAAW,GAAA,CAAI,OAAA,EAAS,CAAC,EAAA,IAAM,KAAA;AACvD,EAAA,OAAO,EAAE,IAAA,EAAM,GAAA,CAAI,KAAA,CAAM,MAAA,EAAQ,CAAC,CAAA,EAAG,OAAA,EAAS,YAAY,CAAA;AAC5D;AAQA,SAAS,kBAAA,CAAmB,KAAA,EAAiB,GAAA,EAAa,aAAA,EAA8B;AACtF,EAAA,MAAM,OAAA,EAAS,OAAA,CAAQ,GAAG,CAAA;AAC1B,EAAA,GAAA,CAAI,OAAA,IAAW,IAAA,EAAM,MAAA;AACrB,EAAA,MAAM,KAAA,EAAO,cAAA,EAAgB,MAAA,CAAO,KAAA,EAAO,UAAA,CAAW,MAAA,CAAO,IAAI,CAAA;AACjE,EAAA,GAAA,CAAI,MAAA,CAAO,OAAA,EAAS;AAClB,IAAA,GAAA,CAAI,KAAA,CAAM,KAAA,CAAM,OAAA,EAAS,CAAC,EAAA,IAAM,IAAA,EAAM,KAAA,CAAM,GAAA,CAAI,CAAA;AAChD,IAAA,MAAA;AAAA,EACF;AACA,EAAA,GAAA,CAAI,CAAC,MAAA,CAAO,YAAA,GAAe,gDAAA,IAAqB,CAAA,EAAG,KAAA,CAAM,IAAA,CAAK,IAAI,CAAA;AACpE;AAEA,SAAS,QAAA,CAAS,MAAA,EAAgB,UAAA,EAA2C;AAC3E,EAAA,OAAO,oCAAA;AAAA,IACL,8BAAA,EAAQ,UAAA,EAAY,CAAC,GAAG,UAAU,EAAE,CAAC,CAAA,CAClC,QAAA,CAAS,CAAA,CACT,KAAA,CAAM,mCAAA,CAAW,CAAE,MAAA,EAAQ,IAAA,EAAM,IAAI,CAAC;AAAA,EAC3C,CAAA;AACF;AAQO,SAAS,cAAA,CAAe,OAAA,EAAuC;AACpE,EAAA,GAAA,CAAI,QAAA,IAAY,aAAA,GAAgB,QAAA,IAAY,KAAA,EAAO,OAAO,OAAA;AAC1D,EAAA,GAAA,CAAI,QAAA,IAAY,KAAA,CAAA,EAAW;AACzB,IAAA,MAAM,IAAI,oCAAA;AAAA,MACR,0BAAA;AAAA,MACA;AAAA,IACF,CAAA;AAAA,EACF;AACA,EAAA,MAAM,IAAI,oCAAA;AAAA,IACR,0BAAA;AAAA,IACA,CAAA,iBAAA,EAAoB,MAAA,CAAO,OAAO,CAAC,CAAA,kCAAA;AAAA,EACrC,CAAA;AACF;AAKA,IAAM,oBAAA,EAAsB,CAAC,MAAA,EAAQ,MAAM,CAAA;AAE3C,SAAS,aAAA,CAAc,OAAA,EAA+B;AACpD,EAAA,MAAM,GAAA,EAAK,wDAAA,CAAa,GAAG,mBAAmB,CAAC,CAAA;AAC/C,EAAA,OAAO,QAAA,IAAY,MAAA,EAAQ,CAAC,EAAA,EAAI,wCAAA,CAAI,EAAG,4CAAA,CAAO,EAAA,EAAI,CAAC,EAAA,EAAI,wCAAA,CAAK,CAAA;AAC9D;AAEO,SAAS,aAAA,CAAc,MAAA,EAAgB,OAAA,EAA0B;AACtE,EAAA,MAAM,OAAA,EAAS,gDAAA,OAAiB,EAAS,CAAA,EAAA,GAAM,QAAA,CAAS,MAAA,EAAQ,aAAA,CAAc,OAAO,CAAC,CAAC,CAAA;AACvF,EAAA,MAAM,MAAA,EAAgB,CAAC,CAAA;AACvB,EAAA,MAAM,aAAA,EAAyB,CAAC,CAAA;AAChC,EAAA,IAAI,UAAA,EAAY,CAAA;AAEhB,EAAA,IAAA,CAAA,MAAW,CAAC,IAAA,EAAM,KAAK,EAAA,GAAK,MAAA,EAAQ;AAClC,IAAA,MAAM,KAAA,EAAO,KAAA,CAAM,IAAA;AAEnB,IAAA,GAAA,CAAI,mBAAA,CAAoB,GAAA,CAAI,IAAI,CAAA,EAAG;AACjC,MAAA,UAAA,GAAa,KAAA,IAAS,QAAA,EAAU,EAAA,EAAI,CAAA,CAAA;AACpC,MAAA,QAAA;AAAA,IACF;AAEA,IAAA,GAAA,CAAI,mBAAA,CAAoB,GAAA,CAAI,IAAI,CAAA,EAAG;AACjC,MAAA,GAAA,CAAI,KAAA,IAAS,OAAA,EAAS;AACpB,QAAA,kBAAA;AAAA,UACE,YAAA;AAAA,UACA,MAAA,CAAO,KAAA,CAAM,KAAA,CAAM,KAAA,CAAM,MAAA,EAAQ,KAAA,CAAM,GAAA,CAAI,MAAM,CAAA;AAAA,UACjD,KAAA,IAAS;AAAA,QACX,CAAA;AAAA,MACF;AACA,MAAA,UAAA,GAAa,KAAA,IAAS,QAAA,EAAU,EAAA,EAAI,CAAA,CAAA;AACpC,MAAA,QAAA;AAAA,IACF;AAIA,IAAA,GAAA,CAAI,KAAA,IAAS,UAAA,EAAY;AACvB,MAAA,GAAA,CAAI,KAAA,IAAS,QAAA,GAAW,UAAA,IAAc,EAAA,GAAK,YAAA,CAAa,OAAA,IAAW,CAAA,EAAG;AACpE,QAAA,IAAA,CAAA,MAAW,KAAA,GAAQ,iDAAA;AAAA,UACjB,MAAA,CAAO,KAAA,CAAM,KAAA,CAAM,KAAA,CAAM,MAAA,EAAQ,KAAA,CAAM,GAAA,CAAI,MAAM,CAAA;AAAA,UACjD,KAAA,CAAM,KAAA,CAAM;AAAA,QACd,CAAA,EAAG;AACD,UAAA,KAAA,CAAM,IAAA,CAAK,IAAI,CAAA;AAAA,QACjB;AAAA,MACF;AACA,MAAA,UAAA,GAAa,KAAA,IAAS,QAAA,EAAU,EAAA,EAAI,CAAA,CAAA;AACpC,MAAA,QAAA;AAAA,IACF;AAEA,IAAA,GAAA,CAAI,KAAA,IAAS,QAAA,GAAW,KAAA,IAAS,OAAA,GAAU,UAAA,IAAc,EAAA,GAAK,YAAA,CAAa,OAAA,IAAW,CAAA,EAAG;AACvF,MAAA,KAAA,CAAM,IAAA,CAAK,EAAE,KAAA,EAAO,KAAA,CAAM,KAAA,CAAM,MAAA,EAAQ,GAAA,EAAK,KAAA,CAAM,GAAA,CAAI,OAAO,CAAC,CAAA;AAAA,IACjE;AAAA,EACF;AAEA,EAAA,OAAO,KAAA;AACT;ADrDA;AACA;AE5IO,SAAS,mBAAA,CAAoB,KAAA,EAAe,OAAA,EAAmC;AACpF,EAAA,MAAM,aAAA,EAAe,mDAAA,OAAoB,CAAQ,UAAU,CAAA;AAC3D,EAAA,MAAM,QAAA,EAAU,yCAAA,OAAU,CAAQ,KAAK,CAAA;AACvC,EAAA,MAAM,OAAA,EAAS,6CAAA,OAAc,CAAQ,MAAM,CAAA;AAC3C,EAAA,MAAM,QAAA,EAAU,cAAA,CAAe,OAAA,CAAQ,OAAO,CAAA;AAC9C,EAAA,MAAM,MAAA,EAAQ,aAAA,CAAc,KAAA,EAAO,OAAO,CAAA;AAC1C,EAAA,OAAO,4CAAA,KAAa,EAAO,KAAA,EAAO,OAAA,EAAS,MAAA,EAAQ,UAAA,EAAY,YAAY,CAAA;AAC7E;AAKO,SAAS,uBAAA,CAAwB,KAAA,EAAe,OAAA,EAAqC;AAC1F,EAAA,MAAM,aAAA,EAAe,mDAAA,OAAoB,CAAQ,UAAU,CAAA;AAC3D,EAAA,MAAM,QAAA,EAAU,yCAAA,OAAU,CAAQ,KAAK,CAAA;AACvC,EAAA,MAAM,OAAA,EAAS,6CAAA,OAAc,CAAQ,MAAM,CAAA;AAC3C,EAAA,MAAM,QAAA,EAAU,cAAA,CAAe,OAAA,CAAQ,OAAO,CAAA;AAC9C,EAAA,OAAO,gDAAA;AAAA,IACL,KAAA;AAAA,IACA,aAAA,CAAc,KAAA,EAAO,OAAO,CAAA;AAAA,IAC5B,OAAA;AAAA,IACA,MAAA;AAAA,IACA,UAAA;AAAA,IACA;AAAA,EACF,CAAA;AACF;AF0IA;AACA;AACE;AACA;AACF,6GAAC","file":"/home/runner/work/polytypo-js/polytypo-js/dist/chunk-FZ6YRCGR.cjs","sourcesContent":[null,"import { parse, postprocess, preprocess } from \"micromark\";\nimport { frontmatter } from \"micromark-extension-frontmatter\";\nimport { gfm } from \"micromark-extension-gfm\";\nimport { mdxjs } from \"micromark-extension-mdxjs\";\nimport type { Event, Extension } from \"micromark-util-types\";\nimport { PolytypoError } from \"../errors.js\";\nimport type { Dialect } from \"../types.js\";\nimport { htmlFragmentSpans, isSkippedElement } from \"./html.js\";\nimport { wrapParserErrors } from \"./parse-error.js\";\nimport type { Span } from \"./spans.js\";\n\n/**\n * spec/rules/modes.md 3.7. Each entry is skipped **whole**, including anything that looks\n * processable inside it. The complement is not enumerated: a span is emitted only for a\n * micromark `data` token, which is the tokenizer's own name for \"literal content\", so list\n * markers, emphasis delimiters, heading sequences, table padding, line endings, hard-break\n * spaces, character escapes and character references are outside every span by construction\n * rather than by a second list that could drift from the first.\n */\nconst SKIPPED_TOKEN_TYPES: ReadonlySet<string> = new Set([\n // fenced code blocks (including the info string and the fences) and indented code blocks\n \"codeFenced\",\n \"codeIndented\",\n // inline code spans, including the backticks\n \"codeText\",\n // autolinks `<https://…>`, and GFM's bare-URL form\n \"autolink\",\n \"literalAutolink\",\n // link and image destinations and titles; the definition line of a reference link. The link\n // *text* is a `labelText`, which is not skipped.\n \"resource\",\n \"reference\",\n \"definition\",\n // MDX: expression containers in full, and every JSX attribute — the whole tag is skipped, so\n // attributes never surface. JSX element *children* sit outside the tag tokens and are\n // processable.\n \"mdxFlowExpression\",\n \"mdxTextExpression\",\n \"mdxjsEsm\",\n // modes.md 3.7.3, and it matters: without it the second `---` of a YAML block reads as a setext\n // underline, `title: Une note` becomes a paragraph, and `fr` puts a narrow no-break space in\n // front of the colon of a machine-read metadata field. Covered by conformance fixtures\n // en-us-markdown-{commonmark,mdx}-frontmatter, fr-markdown-{commonmark,mdx}-frontmatter-nbsp,\n // en-us-markdown-commonmark-frontmatter-toml and -frontmatter-unterminated.\n \"frontmatter\",\n]);\n\n/** Token types whose enter/exit also maintains the skipped-element stack. */\nconst RAW_TAG_TOKEN_TYPES: ReadonlySet<string> = new Set([\n \"htmlText\",\n \"mdxJsxTextTag\",\n \"mdxJsxFlowTag\",\n]);\n\nconst LESS_THAN = 0x3c;\nconst GREATER_THAN = 0x3e;\nconst SLASH = 0x2f;\nconst EXCLAMATION = 0x21;\nconst QUESTION = 0x3f;\n\nfunction isAsciiLetter(unit: number): boolean {\n return (unit >= 0x61 && unit <= 0x7a) || (unit >= 0x41 && unit <= 0x5a);\n}\n\nfunction isTagNameUnit(unit: number): boolean {\n // Deliberately permissive: anything that is not a delimiter belongs to the name. Custom\n // elements (`my-callout`) and member expressions (`Foo.Bar`) must come out whole so that they\n // fail the skip-list membership test rather than being truncated into something that passes.\n return (\n unit !== GREATER_THAN &&\n unit !== SLASH &&\n unit !== 0x20 &&\n unit !== 0x09 &&\n unit !== 0x0a &&\n unit !== 0x0d\n );\n}\n\n/** ASCII-only, because HTML tag names are ASCII case-insensitive and `toLowerCase()` is not. */\nfunction asciiLower(value: string): string {\n let out = \"\";\n for (let i = 0; i < value.length; i += 1) {\n const unit = value.charCodeAt(i);\n out += unit >= 0x41 && unit <= 0x5a ? String.fromCharCode(unit + 0x20) : value[i];\n }\n return out;\n}\n\ninterface RawTag {\n readonly name: string;\n readonly closing: boolean;\n readonly selfClosing: boolean;\n}\n\nfunction readTag(tag: string): RawTag | null {\n if (tag.charCodeAt(0) !== LESS_THAN) return null;\n let i = 1;\n const closing = tag.charCodeAt(i) === SLASH;\n if (closing) i += 1;\n // Comments, declarations and processing instructions carry no element name.\n if (tag.charCodeAt(i) === EXCLAMATION || tag.charCodeAt(i) === QUESTION) return null;\n if (!isAsciiLetter(tag.charCodeAt(i))) return null;\n const nameAt = i;\n while (i < tag.length && isTagNameUnit(tag.charCodeAt(i))) i += 1;\n const selfClosing = tag.charCodeAt(tag.length - 2) === SLASH;\n return { name: tag.slice(nameAt, i), closing, selfClosing };\n}\n\n/**\n * modes.md 3.7: raw HTML is \"handed to the html skip list of 3.6\". Inline raw HTML reaches the\n * tokenizer as isolated tags with ordinary markdown content between them, so honouring the\n * subtree rule means tracking which skipped element is currently open. JSX names are compared\n * case-sensitively (`<Code>` is a component, `<code>` is an element); HTML names are not.\n */\nfunction updateElementStack(stack: string[], tag: string, caseSensitive: boolean): void {\n const parsed = readTag(tag);\n if (parsed === null) return;\n const name = caseSensitive ? parsed.name : asciiLower(parsed.name);\n if (parsed.closing) {\n if (stack[stack.length - 1] === name) stack.pop();\n return;\n }\n if (!parsed.selfClosing && isSkippedElement(name)) stack.push(name);\n}\n\nfunction tokenize(source: string, extensions: readonly Extension[]): Event[] {\n return postprocess(\n parse({ extensions: [...extensions] })\n .document()\n .write(preprocess()(source, null, true)),\n );\n}\n\n/**\n * modes.md 3.7.1. The dialect is the caller's, never detected: `\"commonmark\"` is CommonMark 0.31\n * plus GFM, `\"mdx\"` is the same minus indented code blocks and `<…>` autolinks, plus JSX and\n * `{…}` expression containers. Both enable frontmatter and GFM, which 3.7.2 makes normative —\n * §4's promise that table alignment rows survive is empty unless tables are recognised at all.\n */\nexport function resolveDialect(dialect: Dialect | undefined): Dialect {\n if (dialect === \"commonmark\" || dialect === \"mdx\") return dialect;\n if (dialect === undefined) {\n throw new PolytypoError(\n \"POLYTYPO_INVALID_DIALECT\",\n 'Mode \"markdown\" requires a dialect. Expected \"commonmark\" or \"mdx\"; there is no default.',\n );\n }\n throw new PolytypoError(\n \"POLYTYPO_INVALID_DIALECT\",\n `Unknown dialect \"${String(dialect)}\". Expected \"commonmark\" or \"mdx\".`,\n );\n}\n\n// modes.md 3.7.3 names both frontmatter delimiters, `---` (YAML) and `+++` (TOML), and skips\n// either whole. micromark's default matter is YAML alone, so both must be asked for by name: with\n// the default, a `+++` block parses as prose and the machine-read fields inside it get typeset.\nconst FRONTMATTER_MATTERS = [\"yaml\", \"toml\"] as const;\n\nfunction extensionsFor(dialect: Dialect): Extension[] {\n const fm = frontmatter([...FRONTMATTER_MATTERS]);\n return dialect === \"mdx\" ? [fm, gfm(), mdxjs()] : [fm, gfm()];\n}\n\nexport function markdownSpans(source: string, dialect: Dialect): Span[] {\n const events = wrapParserErrors(dialect, () => tokenize(source, extensionsFor(dialect)));\n const spans: Span[] = [];\n const elementStack: string[] = [];\n let skipDepth = 0;\n\n for (const [kind, token] of events) {\n const type = token.type;\n\n if (SKIPPED_TOKEN_TYPES.has(type)) {\n skipDepth += kind === \"enter\" ? 1 : -1;\n continue;\n }\n\n if (RAW_TAG_TOKEN_TYPES.has(type)) {\n if (kind === \"enter\") {\n updateElementStack(\n elementStack,\n source.slice(token.start.offset, token.end.offset),\n type !== \"htmlText\",\n );\n }\n skipDepth += kind === \"enter\" ? 1 : -1;\n continue;\n }\n\n // An HTML block is handed to the `html` extractor rather than skipped whole, so that the\n // prose inside `<div>…</div>` is typeset while the markup is not.\n if (type === \"htmlFlow\") {\n if (kind === \"enter\" && skipDepth === 0 && elementStack.length === 0) {\n for (const span of htmlFragmentSpans(\n source.slice(token.start.offset, token.end.offset),\n token.start.offset,\n )) {\n spans.push(span);\n }\n }\n skipDepth += kind === \"enter\" ? 1 : -1;\n continue;\n }\n\n if (kind === \"enter\" && type === \"data\" && skipDepth === 0 && elementStack.length === 0) {\n spans.push({ start: token.start.offset, end: token.end.offset });\n }\n }\n\n return spans;\n}\n","import { markdownSpans, resolveDialect } from \"../modes/markdown.js\";\nimport type { Options } from \"../types.js\";\nimport { getLocaleData } from \"./locale.js\";\nimport { resolveNarrowTarget } from \"./narrow-target.js\";\nimport { planRules } from \"./rule-runner.js\";\nimport { analyzeOverSpans, runOverSpans } from \"./span-runner.js\";\nimport type { Change } from \"./origin.js\";\n\n/**\n * `markdown` mode only. Imports the Micromark/MDX stack and `parse5` (via `../modes/markdown.js`\n * importing `../modes/html.js` for normative embedded-HTML handling, modes.md 3.7) — both are\n * legitimately reachable from `polytypo/markdown` (AUDIT_REMEDIATION_AND_RELEASE_PLAN.md 5.1).\n *\n * Validation order — `planRules`, then `getLocaleData`, then `resolveDialect`/parsing — mirrors\n * the pre-Stage-5 aggregate `runPipeline` exactly: an unknown-rule error wins over an\n * unknown-locale error, which wins over a missing/invalid dialect, which wins over a parse\n * failure. All four are public, tested behaviour this refactor was not authorised to change.\n */\nexport function runMarkdownPipeline(input: string, options: Partial<Options>): string {\n const narrowTarget = resolveNarrowTarget(options.narrowNbsp);\n const planned = planRules(options.rules);\n const locale = getLocaleData(options.locale);\n const dialect = resolveDialect(options.dialect);\n const spans = markdownSpans(input, dialect);\n return runOverSpans(input, spans, planned, locale, \"markdown\", narrowTarget);\n}\n\n/** analyze.md §1, `markdown` mode. Dialect validation happens here exactly as it does for\n * `runMarkdownPipeline`, so an absent dialect throws POLYTYPO_INVALID_DIALECT from `analyze`\n * too (analyze.md §4 A1). */\nexport function analyzeMarkdownPipeline(input: string, options: Partial<Options>): Change[] {\n const narrowTarget = resolveNarrowTarget(options.narrowNbsp);\n const planned = planRules(options.rules);\n const locale = getLocaleData(options.locale);\n const dialect = resolveDialect(options.dialect);\n return analyzeOverSpans(\n input,\n markdownSpans(input, dialect),\n planned,\n locale,\n \"markdown\",\n narrowTarget,\n );\n}\n"]}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
"use strict";Object.defineProperty(exports, "__esModule", {value: true});
|
|
2
2
|
|
|
3
|
-
var
|
|
3
|
+
var _chunkPJ6HB77Gcjs = require('./chunk-PJ6HB77G.cjs');
|
|
4
4
|
|
|
5
5
|
// src/modes/html.ts
|
|
6
6
|
var _parse5 = require('parse5');
|
|
@@ -20,8 +20,8 @@ function wrapParserErrors(what, run) {
|
|
|
20
20
|
try {
|
|
21
21
|
return run();
|
|
22
22
|
} catch (error) {
|
|
23
|
-
if (error instanceof
|
|
24
|
-
throw new (0,
|
|
23
|
+
if (error instanceof _chunkPJ6HB77Gcjs.PolytypoError) throw error;
|
|
24
|
+
throw new (0, _chunkPJ6HB77Gcjs.PolytypoError)(
|
|
25
25
|
"POLYTYPO_MALFORMED_INPUT",
|
|
26
26
|
`Input does not parse as ${what}: ${describe(error)}`
|
|
27
27
|
);
|
|
@@ -144,4 +144,4 @@ function htmlFragmentSpans(source, offset) {
|
|
|
144
144
|
|
|
145
145
|
|
|
146
146
|
exports.wrapParserErrors = wrapParserErrors; exports.isSkippedElement = isSkippedElement; exports.htmlSpans = htmlSpans; exports.htmlFragmentSpans = htmlFragmentSpans;
|
|
147
|
-
//# sourceMappingURL=chunk-
|
|
147
|
+
//# sourceMappingURL=chunk-GJJNGBRH.cjs.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-
|
|
1
|
+
{"version":3,"sources":["/home/runner/work/polytypo-js/polytypo-js/dist/chunk-GJJNGBRH.cjs","../src/modes/html.ts","../src/modes/parse-error.ts"],"names":[],"mappings":"AAAA;AACE;AACF,wDAA6B;AAC7B;AACA;ACJA,gCAAqC;ADMrC;AACA;AEAA,SAAS,QAAA,CAAS,KAAA,EAAwB;AACxC,EAAA,GAAA,CAAI,OAAO,MAAA,IAAU,SAAA,GAAY,MAAA,IAAU,IAAA,EAAM,OAAO,MAAA,CAAO,KAAK,CAAA;AACpE,EAAA,MAAM,WAAA,EAAa,KAAA;AACnB,EAAA,MAAM,OAAA,EACJ,OAAO,UAAA,CAAW,OAAA,IAAW,SAAA,EAAW,UAAA,CAAW,OAAA,EAAU,KAAA,CAAgB,OAAA;AAC/E,EAAA,MAAM,MAAA,EAAQ,UAAA,CAAW,KAAA;AACzB,EAAA,GAAA,CACE,MAAA,IAAU,KAAA,GACV,MAAA,IAAU,KAAA,EAAA,GACV,OAAO,KAAA,CAAM,KAAA,IAAS,SAAA,GACtB,OAAO,KAAA,CAAM,OAAA,IAAW,QAAA,EACxB;AACA,IAAA,OAAO,CAAA,EAAA;AACT,EAAA;AACO,EAAA;AACT;AAcgB;AACV,EAAA;AACK,IAAA;AACA,EAAA;AACH,IAAA;AACE,IAAA;AACJ,MAAA;AACA,MAAA;AACF,IAAA;AACF,EAAA;AACF;AFjBY;AACA;ACfN;AACJ,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACD;AAEe;AACP,EAAA;AACT;AAEM;AACA;AACO;AACP;AACA;AAGA;AAEG;AACA,EAAA;AACT;AAES;AACA,EAAA;AACT;AAES;AACA,EAAA;AACT;AAQS;AACH,EAAA;AACI,EAAA;AACC,EAAA;AAEL,EAAA;AACG,IAAA;AACC,IAAA;AACA,IAAA;AACG,IAAA;AACH,IAAA;AAEJ,IAAA;AAGK,MAAA;AACP,IAAA;AACI,IAAA;AACG,IAAA;AACT,EAAA;AAEM,EAAA;AAEJ,EAAA;AAIK,IAAA;AACP,EAAA;AACU,EAAA;AACH,EAAA;AACT;AAOgB;AACR,EAAA;AACF,EAAA;AACI,EAAA;AACD,EAAA;AACD,IAAA;AACI,MAAA;AACF,MAAA;AACE,QAAA;AACJ,QAAA;AACI,QAAA;AACJ,QAAA;AACF,MAAA;AACF,IAAA;AACK,IAAA;AACP,EAAA;AACU,EAAA;AACH,EAAA;AACT;AAES;AACA,EAAA;AACT;AAES;AACE,EAAA;AACD,IAAA;AACF,IAAA;AACJ,IAAA;AACQ,MAAA;AACR,IAAA;AACA,IAAA;AACF,EAAA;AAGS,EAAA;AAEL,EAAA;AAGK,EAAA;AACC,IAAA;AACR,IAAA;AACF,EAAA;AAEK,EAAA;AACL,EAAA;AACF;AAQgB;AACR,EAAA;AACA,EAAA;AACE,EAAA;AACD,EAAA;AACT;AAGgB;AACR,EAAA;AAA4B,IAAA;AAChC,IAAA;AACF,EAAA;AACM,EAAA;AACE,EAAA;AACD,EAAA;AAGT;AD9BY;AACA;AACA;AACA;AACA;AACA;AACA","file":"/home/runner/work/polytypo-js/polytypo-js/dist/chunk-GJJNGBRH.cjs","sourcesContent":[null,"import { parse, parseFragment } from \"parse5\";\nimport type { DefaultTreeAdapterMap } from \"parse5\";\nimport { wrapParserErrors } from \"./parse-error.js\";\nimport type { Span } from \"./spans.js\";\n\ntype Node = DefaultTreeAdapterMap[\"node\"];\ntype ParentNode = DefaultTreeAdapterMap[\"parentNode\"];\n\n/**\n * spec/rules/modes.md 3.6, exhaustive and **closed**: extending it is a spec change, not an\n * implementation decision. `svg` and `math` are here because in MathML a quotation mark, a\n * hyphen and a prime are operators and identifiers — substituting a curly glyph changes what\n * the expression means. Everything not listed is processable, **including unknown and custom\n * elements**: guessing from a tag name is exactly the heuristic that diverges across runtimes.\n */\nconst SKIPPED_ELEMENTS: ReadonlySet<string> = new Set([\n \"code\",\n \"pre\",\n \"kbd\",\n \"samp\",\n \"var\",\n \"script\",\n \"style\",\n \"textarea\",\n \"svg\",\n \"math\",\n]);\n\nexport function isSkippedElement(tagName: string): boolean {\n return SKIPPED_ELEMENTS.has(tagName);\n}\n\nconst AMPERSAND = 0x26;\nconst SEMICOLON = 0x3b;\nconst HASH = 0x23;\nconst LOWER_X = 0x78;\nconst UPPER_X = 0x58;\n\n/** Longest named reference in the WHATWG table is 32 characters; the guard is a bound, not a rule. */\nconst MAX_REFERENCE_BODY = 34;\n\nfunction isAsciiDigit(unit: number): boolean {\n return unit >= 0x30 && unit <= 0x39;\n}\n\nfunction isAsciiHexDigit(unit: number): boolean {\n return isAsciiDigit(unit) || (unit >= 0x61 && unit <= 0x66) || (unit >= 0x41 && unit <= 0x46);\n}\n\nfunction isAsciiAlphanumeric(unit: number): boolean {\n return isAsciiDigit(unit) || (unit >= 0x61 && unit <= 0x7a) || (unit >= 0x41 && unit <= 0x5a);\n}\n\n/**\n * The end offset of a well-formed character reference starting at `at`, or -1. A bare `&` that a\n * parser would repair is deliberately **not** treated as a reference: it stays inside the span,\n * where no rule can emit or delete it, rather than manufacturing a boundary that suppresses\n * conversions around it.\n */\nfunction characterReferenceEnd(source: string, at: number, limit: number): number {\n if (source.charCodeAt(at) !== AMPERSAND) return -1;\n let i = at + 1;\n if (i >= limit) return -1;\n\n if (source.charCodeAt(i) === HASH) {\n i += 1;\n const unit = i < limit ? source.charCodeAt(i) : -1;\n const hex = unit === LOWER_X || unit === UPPER_X;\n if (hex) i += 1;\n const digitsAt = i;\n while (\n i < limit &&\n (hex ? isAsciiHexDigit(source.charCodeAt(i)) : isAsciiDigit(source.charCodeAt(i)))\n ) {\n i += 1;\n }\n if (i === digitsAt) return -1;\n return i < limit && source.charCodeAt(i) === SEMICOLON ? i + 1 : -1;\n }\n\n const nameAt = i;\n while (\n i < limit &&\n i - nameAt < MAX_REFERENCE_BODY &&\n isAsciiAlphanumeric(source.charCodeAt(i))\n ) {\n i += 1;\n }\n if (i === nameAt) return -1;\n return i < limit && source.charCodeAt(i) === SEMICOLON ? i + 1 : -1;\n}\n\n/**\n * modes.md 3.6: every character reference is skipped as an **opaque unit**, and a text node\n * containing one is split into spans around it. That is the only way ` ` survives as\n * ` ` instead of becoming a literal U+00A0 on the way out.\n */\nexport function splitCharacterReferences(source: string, start: number, end: number): Span[] {\n const spans: Span[] = [];\n let cursor = start;\n let i = start;\n while (i < end) {\n if (source.charCodeAt(i) === AMPERSAND) {\n const stop = characterReferenceEnd(source, i, end);\n if (stop > 0) {\n if (i > cursor) spans.push({ start: cursor, end: i });\n cursor = stop;\n i = stop;\n continue;\n }\n }\n i += 1;\n }\n if (end > cursor) spans.push({ start: cursor, end });\n return spans;\n}\n\nfunction isParent(node: Node): node is ParentNode {\n return \"childNodes\" in node;\n}\n\nfunction collect(node: Node, source: string, spans: Span[]): void {\n if (node.nodeName === \"#text\") {\n const location = node.sourceCodeLocation;\n if (location === undefined || location === null) return;\n for (const span of splitCharacterReferences(source, location.startOffset, location.endOffset)) {\n spans.push(span);\n }\n return;\n }\n\n // Comments, doctype and processing instructions never contribute a span (modes.md 3.6).\n if (node.nodeName === \"#comment\" || node.nodeName === \"#documentType\") return;\n\n if (\"tagName\" in node && isSkippedElement(node.tagName)) return;\n\n // A `template`'s children live in a separate document fragment; `template` is not skipped.\n if (node.nodeName === \"template\" && \"content\" in node) {\n collect(node.content, source, spans);\n return;\n }\n\n if (!isParent(node)) return;\n for (const child of node.childNodes) collect(child, source, spans);\n}\n\n/**\n * Locate the processable spans of an HTML document. The tree is used only to find offsets and is\n * then discarded — modes.md 4 forbids reserialisation, which is what makes attribute quoting,\n * self-closing forms, entity spelling, tag case and malformed-input recovery preserved by\n * construction rather than by effort.\n */\nexport function htmlSpans(source: string): Span[] {\n const document = wrapParserErrors(\"HTML\", () => parse(source, { sourceCodeLocationInfo: true }));\n const spans: Span[] = [];\n collect(document, source, spans);\n return spans;\n}\n\n/** The same, for a fragment: used for HTML blocks embedded in markdown (modes.md 3.7). */\nexport function htmlFragmentSpans(source: string, offset: number): Span[] {\n const fragment = wrapParserErrors(\"HTML\", () =>\n parseFragment(source, { sourceCodeLocationInfo: true }),\n );\n const spans: Span[] = [];\n collect(fragment, source, spans);\n return offset === 0\n ? spans\n : spans.map((span) => ({ start: span.start + offset, end: span.end + offset }));\n}\n","import { PolytypoError } from \"../errors.js\";\n\ninterface PositionedError {\n readonly reason?: unknown;\n readonly place?: { readonly line?: unknown; readonly column?: unknown } | null;\n}\n\nfunction describe(error: unknown): string {\n if (typeof error !== \"object\" || error === null) return String(error);\n const positioned = error as PositionedError;\n const reason =\n typeof positioned.reason === \"string\" ? positioned.reason : (error as Error).message;\n const place = positioned.place;\n if (\n place !== null &&\n place !== undefined &&\n typeof place.line === \"number\" &&\n typeof place.column === \"number\"\n ) {\n return `${reason} (line ${place.line}, column ${place.column})`;\n }\n return reason;\n}\n\n/**\n * spec/rules/modes.md 3.7.2. **The parser's own error type must never escape.** A\n * `VFileMessage`, a `Nokogiri::SyntaxError` or a Python exception on the public surface puts a\n * dependency's type into the contract and is unreproducible in the other four runtimes; only the\n * code is contractual (ARCHITECTURE.md 4.6), while the message and the source position are\n * useful and are not.\n *\n * Reachable in exactly one place: `dialect: \"mdx\"`, because MDX embeds JavaScript. HTML parsing\n * is specified with total error recovery and every byte sequence is valid CommonMark, so neither\n * of the other two languages can fail. The wrapper is applied to all of them anyway — the\n * guarantee is that no parser type escapes, not that this particular parser is trusted.\n */\nexport function wrapParserErrors<T>(what: string, run: () => T): T {\n try {\n return run();\n } catch (error) {\n if (error instanceof PolytypoError) throw error;\n throw new PolytypoError(\n \"POLYTYPO_MALFORMED_INPUT\",\n `Input does not parse as ${what}: ${describe(error)}`,\n );\n }\n}\n"]}
|
|
@@ -1,15 +1,15 @@
|
|
|
1
1
|
import {
|
|
2
2
|
htmlSpans
|
|
3
|
-
} from "./chunk-
|
|
3
|
+
} from "./chunk-BGTMX6VS.js";
|
|
4
4
|
import {
|
|
5
5
|
analyzeOverSpans,
|
|
6
6
|
runOverSpans
|
|
7
|
-
} from "./chunk-
|
|
7
|
+
} from "./chunk-NLMP3JVF.js";
|
|
8
8
|
import {
|
|
9
9
|
getLocaleData,
|
|
10
10
|
planRules,
|
|
11
11
|
resolveNarrowTarget
|
|
12
|
-
} from "./chunk-
|
|
12
|
+
} from "./chunk-25RYOB5K.js";
|
|
13
13
|
|
|
14
14
|
// src/engine/html-pipeline.ts
|
|
15
15
|
function runHtmlPipeline(input, options) {
|
|
@@ -30,4 +30,4 @@ export {
|
|
|
30
30
|
runHtmlPipeline,
|
|
31
31
|
analyzeHtmlPipeline
|
|
32
32
|
};
|
|
33
|
-
//# sourceMappingURL=chunk-
|
|
33
|
+
//# sourceMappingURL=chunk-GPOCFMTF.js.map
|
|
@@ -2,17 +2,17 @@ import {
|
|
|
2
2
|
htmlFragmentSpans,
|
|
3
3
|
isSkippedElement,
|
|
4
4
|
wrapParserErrors
|
|
5
|
-
} from "./chunk-
|
|
5
|
+
} from "./chunk-BGTMX6VS.js";
|
|
6
6
|
import {
|
|
7
7
|
analyzeOverSpans,
|
|
8
8
|
runOverSpans
|
|
9
|
-
} from "./chunk-
|
|
9
|
+
} from "./chunk-NLMP3JVF.js";
|
|
10
10
|
import {
|
|
11
11
|
PolytypoError,
|
|
12
12
|
getLocaleData,
|
|
13
13
|
planRules,
|
|
14
14
|
resolveNarrowTarget
|
|
15
|
-
} from "./chunk-
|
|
15
|
+
} from "./chunk-25RYOB5K.js";
|
|
16
16
|
|
|
17
17
|
// src/modes/markdown.ts
|
|
18
18
|
import { parse, postprocess, preprocess } from "micromark";
|
|
@@ -184,4 +184,4 @@ export {
|
|
|
184
184
|
runMarkdownPipeline,
|
|
185
185
|
analyzeMarkdownPipeline
|
|
186
186
|
};
|
|
187
|
-
//# sourceMappingURL=chunk-
|
|
187
|
+
//# sourceMappingURL=chunk-HTWCCPWI.js.map
|