@albertomarturelo/sii-core 0.19.2 → 0.19.3

package/CHANGELOG.md

@@ -4,6 +4,17 @@ All notable changes to `@albertomarturelo/sii-core` are documented here. The for
 loosely based on [Keep a Changelog](https://keepachangelog.com/); the package is
 pre-1.0, so MINOR bumps may carry breaking changes (pin or use `~` downstream).
 
+## 0.19.3 — 2026-11-10
+
+### Fixed
+
+- **`repairMojibake` repairs SII's Windows-1252 double encoding (#271).** A UTF-8 second byte in
+  0x80–0x9F (the capitals « À »–« ß »: « Ó », « É », « Ñ », « Ú ») arrives as one of cp1252's
+  characters above U+00FF, which the guard took for clean text, leaving the whole string broken
+  (« LIQUIDACIÓN … deberá »). They map back to their byte before the test and the round trip.
+  Every text repaired since 0.19.1 gains it: the F22's estado, historial, observaciones, códigos
+  and terceros, and the mipyme DTE borradores. Clean text with « “ », « – » or « € » is untouched.
+
 ## 0.19.2 — 2026-11-09
 
 ### Fixed

package/dist/format/text.d.ts

@@ -1,7 +1,7 @@
-/** Repair SII's double-encoded text by re-encoding the decoded string as Latin-1 and decoding it
- *  as UTF-8 again. Applied ONLY when the string carries the tell-tale `Ã`/`Â` + continuation pair,
- *  holds nothing above U+00FF, and the round trip decodes cleanly, so legitimately-accented text
- *  is left untouched. */
+/** Repair SII's double-encoded text by re-encoding the decoded string as Windows-1252 and decoding
+ *  it as UTF-8 again. Applied ONLY when the string carries the tell-tale `Ã`/`Â` + continuation
+ *  pair (the continuation byte in its Latin-1 or cp1252 form), holds nothing else beyond one byte,
+ *  and the round trip decodes cleanly, so legitimately-accented text is left untouched. */
 export declare function repairMojibake(v: string): string;
 /** A wire value as text, repaired: null stays null, anything else is `String(v)` through
  *  `repairMojibake`. For the free-text fields a facade curates (glosas, descripciones). */

package/dist/format/text.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"text.d.ts","sourceRoot":"","sources":["../../src/format/text.ts"],"names":[],"mappings":"AAMA;;;yBAGyB;AACzB,wBAAgB,cAAc,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,CAUhD;AAED;2FAC2F;AAC3F,eAAO,MAAM,YAAY,GAAI,GAAG,OAAO,KAAG,MAAM,GAAG,IACe,CAAC"}
+{"version":3,"file":"text.d.ts","sourceRoot":"","sources":["../../src/format/text.ts"],"names":[],"mappings":"AAmBA;;;2FAG2F;AAC3F,wBAAgB,cAAc,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,CAehD;AAED;2FAC2F;AAC3F,eAAO,MAAM,YAAY,GAAI,GAAG,OAAO,KAAG,MAAM,GAAG,IACe,CAAC"}

package/dist/format/text.js

@@ -1,18 +1,31 @@
 // SII text as SII meant it. Some endpoints serve DOUBLE-ENCODED text: the UTF-8 bytes of an
 // already-mojibaked string, so even a correct decode yields "Fundación" / "Ñuñoa". Observed on
 // the DTE borradores (`listaBorrador`, 2026-09-08), one F22 terceros title (2026-10-21) and the
-// F22 estado glosas and historial glosas (2026-11-09, GH-264). Hoisted here after the third
-// facade needed it (conventions/sii-wire.md). Display-side only: nothing sent to SII passes here.
-/** Repair SII's double-encoded text by re-encoding the decoded string as Latin-1 and decoding it
- *  as UTF-8 again. Applied ONLY when the string carries the tell-tale `Ã`/`Â` + continuation pair,
- *  holds nothing above U+00FF, and the round trip decodes cleanly, so legitimately-accented text
- *  is left untouched. */
+// F22 estado and historial glosas (2026-11-09, GH-264) and the F22 código glosas (2026-11-10,
+// GH-267, GH-271). Hoisted here after the third facade needed it (conventions/sii-wire.md).
+// Display-side only: nothing sent to SII passes here.
+// SII's double encoding is Windows-1252, not Latin-1 (GH-271): a UTF-8 byte in 0x80–0x9F was read
+// as cp1252's printable character there, so « Ó » (C3 93) arrives as `Ó` (U+201C). These are the
+// characters cp1252 puts at 0x80–0x9F, in byte order; its five undefined bytes (81, 8D, 8F, 90, 9D)
+// decode to the C1 control of the same value, which already is that byte.
+const CP1252_HIGH = '\u20ac\u0081\u201a\u0192\u201e\u2026\u2020\u2021\u02c6\u2030\u0160\u2039\u0152\u008d\u017d\u008f' +
+    '\u0090\u2018\u2019\u201c\u201d\u2022\u2013\u2014\u02dc\u2122\u0161\u203a\u0153\u009d\u017e\u0178';
+const CP1252_BYTE = new Map(Array.from(CP1252_HIGH, (ch, i) => [ch, 0x80 + i]));
+/** The byte a character stood for when SII decoded UTF-8 as cp1252; its code point otherwise. */
+const cp1252Byte = (ch) => CP1252_BYTE.get(ch) ?? ch.codePointAt(0) ?? 0;
+/** Repair SII's double-encoded text by re-encoding the decoded string as Windows-1252 and decoding
+ *  it as UTF-8 again. Applied ONLY when the string carries the tell-tale `Ã`/`Â` + continuation
+ *  pair (the continuation byte in its Latin-1 or cp1252 form), holds nothing else beyond one byte,
+ *  and the round trip decodes cleanly, so legitimately-accented text is left untouched. */
 export function repairMojibake(v) {
-    if (!/[\u00c2-\u00c3][\u0080-\u00bf]/.test(v))
+    if (!/[ÂÃ]/.test(v))
+        return v; // every tell-tale pair starts with  or Ã
+    const codes = Array.from(v, cp1252Byte);
+    const tellTale = codes.some((c, i) => (c === 0xc2 || c === 0xc3) && (codes[i + 1] ?? 0) >= 0x80 && (codes[i + 1] ?? 0) <= 0xbf);
+    if (!tellTale)
         return v;
-    const codes = Array.from(v, (ch) => ch.charCodeAt(0));
     if (codes.some((c) => c > 0xff))
-        return v; // a non-Latin-1 char ⇒ not a mojibake string
+        return v; // a char that is no cp1252 byte ⇒ not a mojibake string
     try {
         const decoded = new TextDecoder('utf-8', { fatal: true }).decode(Uint8Array.from(codes));
         return decoded.includes('\ufffd') ? v : decoded;

package/dist/format/text.js.map

@@ -1 +1 @@
-{"version":3,"file":"text.js","sourceRoot":"","sources":["../../src/format/text.ts"],"names":[],"mappings":"AAAA,4FAA4F;AAC5F,kGAAkG;AAClG,gGAAgG;AAChG,4FAA4F;AAC5F,kGAAkG;AAElG;;;yBAGyB;AACzB,MAAM,UAAU,cAAc,CAAC,CAAS;IACtC,IAAI,CAAC,gCAAgC,CAAC,IAAI,CAAC,CAAC,CAAC;QAAE,OAAO,CAAC,CAAC;IACxD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;IACtD,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC,CAAC,6CAA6C;IACxF,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,IAAI,WAAW,CAAC,OAAO,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QACzF,OAAO,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;IAClD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,CAAC,CAAC,CAAC,6DAA6D;IACzE,CAAC;AACH,CAAC;AAED;2FAC2F;AAC3F,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAU,EAAiB,EAAE,CACxD,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC"}
+{"version":3,"file":"text.js","sourceRoot":"","sources":["../../src/format/text.ts"],"names":[],"mappings":"AAAA,4FAA4F;AAC5F,kGAAkG;AAClG,gGAAgG;AAChG,8FAA8F;AAC9F,4FAA4F;AAC5F,sDAAsD;AAEtD,kGAAkG;AAClG,kGAAkG;AAClG,oGAAoG;AACpG,0EAA0E;AAC1E,MAAM,WAAW,GACf,kGAAkG;IAClG,kGAAkG,CAAC;AACrG,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AAEhF,iGAAiG;AACjG,MAAM,UAAU,GAAG,CAAC,EAAU,EAAU,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;AAEzF;;;2FAG2F;AAC3F,MAAM,UAAU,cAAc,CAAC,CAAS;IACtC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC;QAAE,OAAO,CAAC,CAAC,CAAC,0CAA0C;IACzE,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;IACxC,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CACzB,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACP,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAC3F,CAAC;IACF,IAAI,CAAC,QAAQ;QAAE,OAAO,CAAC,CAAC;IACxB,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC,CAAC,wDAAwD;IACnG,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,IAAI,WAAW,CAAC,OAAO,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QACzF,OAAO,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;IAClD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,CAAC,CAAC,CAAC,6DAA6D;IACzE,CAAC;AACH,CAAC;AAED;2FAC2F;AAC3F,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAU,EAAiB,EAAE,CACxD,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@albertomarturelo/sii-core",
-  "version": "0.19.2",
+  "version": "0.19.3",
   "type": "module",
   "description": "Shared SII domain core: rut, config (prod hostnames), contracts, tasks, seams. A Node library; external side-effects (portal, secrets, audit, clock) are injected via seams (ADR-003).",
   "license": "MIT",