tsc-ts 0.0.4 → 0.0.6

package/src/compiler/link-tables.ts

@@ -0,0 +1,343 @@
+// Side tables of checker links, indexed by node or symbol id.
+//
+// The checker caches values per node and per symbol. A cached value that many nodes or symbols
+// have lives in a table of its own (a column) instead of a field of a links object, so a node or
+// symbol costs a slot in each column it uses rather than a whole object, and a lookup is two
+// array reads. Columns are stored in pages of `LinkPaging.Size` ids created on first write: a
+// checker that touches part of a program (as each of several parallel checkers does) stores only
+// the pages it touches.
+
+import type { Symbol } from './symbol.js';
+
+/** Links are stored in pages of 256 ids. */
+export const enum LinkPaging {
+  Bits = 8,
+  Size = 1 << Bits,
+  Mask = Size - 1,
+}
+
+/** A page of values, indexed by the low bits of an id; undefined until a value in it is stored. */
+export type LinkPage<T> = (T | undefined)[] | undefined;
+
+/** Values by id; undefined where none is stored. */
+export class LinkTable<T> {
+  readonly pages: LinkPage<T>[] = [];
+
+  get(id: number): T | undefined {
+    const page = this.pages[id >> LinkPaging.Bits];
+    return page !== undefined ? page[id & LinkPaging.Mask] : undefined;
+  }
+
+  set(id: number, value: T | undefined): void {
+    const index = id >> LinkPaging.Bits;
+    const page = this.pages[index];
+    if (page !== undefined) page[id & LinkPaging.Mask] = value;
+    else if (value !== undefined) newLinkPage(this.pages, index)[id & LinkPaging.Mask] = value;
+  }
+
+  clear(): void {
+    this.pages.length = 0;
+  }
+}
+
+/** Creates the page at `index`, filling the gap below it so the page list stays dense. */
+export function newLinkPage<T>(pages: LinkPage<T>[], index: number): (T | undefined)[] {
+  while (pages.length <= index) pages.push(undefined);
+  const page: (T | undefined)[] = [];
+  for (let i = 0; i < LinkPaging.Size; i++) page.push(undefined);
+  pages[index] = page;
+  return page;
+}
+
+/** Int32 values by id, zero where none is stored: flags and small counters. */
+export class IntLinkTable {
+  readonly pages: (Int32Array | undefined)[] = [];
+
+  get(id: number): number {
+    const page = this.pages[id >> LinkPaging.Bits];
+    return page !== undefined ? page[id & LinkPaging.Mask]! : 0;
+  }
+
+  set(id: number, value: number): void {
+    const index = id >> LinkPaging.Bits;
+    const page = this.pages[index];
+    if (page !== undefined) page[id & LinkPaging.Mask] = value;
+    else if (value !== 0) this.newPage(index)[id & LinkPaging.Mask] = value;
+  }
+
+  private newPage(index: number): Int32Array {
+    while (this.pages.length <= index) this.pages.push(undefined);
+    const page = new Int32Array(LinkPaging.Size);
+    this.pages[index] = page;
+    return page;
+  }
+
+  clear(): void {
+    this.pages.length = 0;
+  }
+}
+
+/**
+ * Values by id in a map, for ids that are sparse: a checker caches values for a small part of the
+ * nodes of a program, spread over the declaration files it reads, where a page would mostly hold
+ * empty slots.
+ */
+export class SparseLinkTable<T> {
+  readonly values = new Map<number, T>();
+
+  get(id: number): T | undefined {
+    return this.values.get(id);
+  }
+
+  set(id: number, value: T | undefined): void {
+    if (value !== undefined) this.values.set(id, value);
+    else this.values.delete(id);
+  }
+
+  clear(): void {
+    this.values.clear();
+  }
+}
+
+/**
+ * Values by id for ids in a few clusters, as one file's emit uses them: the file's nodes and the
+ * nodes the emit synthesizes, far above them. Pages of `LinkPaging.Size` ids are found through a
+ * map from the page index, so the ids between clusters cost nothing, and the page of the last
+ * lookup is kept, since lookups come in runs of nearby ids.
+ */
+export class ClusteredLinkTable<T> {
+  private readonly pages = new Map<number, (T | undefined)[]>();
+  private lastIndex = -1;
+  private lastPage: (T | undefined)[] | undefined = undefined;
+
+  get(id: number): T | undefined {
+    const page = this.pageOf(id >> LinkPaging.Bits);
+    return page !== undefined ? page[id & LinkPaging.Mask] : undefined;
+  }
+
+  set(id: number, value: T | undefined): void {
+    const index = id >> LinkPaging.Bits;
+    const page = this.pageOf(index);
+    if (page !== undefined) page[id & LinkPaging.Mask] = value;
+    else if (value !== undefined) {
+      const created: (T | undefined)[] = [];
+      for (let i = 0; i < LinkPaging.Size; i++) created.push(undefined);
+      created[id & LinkPaging.Mask] = value;
+      this.pages.set(index, created);
+      this.lastPage = created;
+    }
+  }
+
+  private pageOf(index: number): (T | undefined)[] | undefined {
+    if (index !== this.lastIndex) {
+      this.lastIndex = index;
+      this.lastPage = this.pages.get(index);
+    }
+    return this.lastPage;
+  }
+}
+
+/** Int32 values by id, zero where none is stored, for ids in a few clusters (`ClusteredLinkTable`). */
+export class ClusteredIntLinkTable {
+  private readonly pages = new Map<number, Int32Array>();
+  private lastIndex = -1;
+  private lastPage: Int32Array | undefined = undefined;
+
+  get(id: number): number {
+    const page = this.pageOf(id >> LinkPaging.Bits);
+    return page !== undefined ? page[id & LinkPaging.Mask]! : 0;
+  }
+
+  set(id: number, value: number): void {
+    const index = id >> LinkPaging.Bits;
+    const page = this.pageOf(index);
+    if (page !== undefined) page[id & LinkPaging.Mask] = value;
+    else if (value !== 0) {
+      const created = new Int32Array(LinkPaging.Size);
+      created[id & LinkPaging.Mask] = value;
+      this.pages.set(index, created);
+      this.lastPage = created;
+    }
+  }
+
+  private pageOf(index: number): Int32Array | undefined {
+    if (index !== this.lastIndex) {
+      this.lastIndex = index;
+      this.lastPage = this.pages.get(index);
+    }
+    return this.lastPage;
+  }
+}
+
+/** The number of set bits of a 32-bit value. */
+function bitCount(x: number): number {
+  x = x - ((x >>> 1) & 0x55555555);
+  x = (x & 0x33333333) + ((x >>> 2) & 0x33333333);
+  x = (x + (x >>> 4)) & 0x0f0f0f0f;
+  return Math.imul(x, 0x01010101) >>> 24;
+}
+
+/**
+ * A page of a `CompactSymbolLinkTable`: one bit per id of the page in eight 32-id words, the number
+ * of symbols stored before each word, and the symbols in id order. The words are fields, so a
+ * lookup reads one object before the symbols. The class isn't generic, so natively the words are
+ * 32-bit fields: every write stores an int32.
+ */
+class CompactLinkPage {
+  readonly values: Symbol[];
+  private w0: number;
+  private w1: number;
+  private w2: number;
+  private w3: number;
+  private w4: number;
+  private w5: number;
+  private w6: number;
+  private w7: number;
+  /** Byte `w & 3` of `before0` (words 1 to 3) or of `before1` (words 4 to 7): the values before word `w`. */
+  private before0: number;
+  private before1: number;
+
+  constructor() {
+    this.values = [];
+    this.w0 = 0;
+    this.w1 = 0;
+    this.w2 = 0;
+    this.w3 = 0;
+    this.w4 = 0;
+    this.w5 = 0;
+    this.w6 = 0;
+    this.w7 = 0;
+    this.before0 = 0;
+    this.before1 = 0;
+  }
+
+  /** The presence bits of ids `32 * word` to `32 * word + 31`. */
+  word(word: number): number {
+    switch (word) {
+      case 0:
+        return this.w0;
+      case 1:
+        return this.w1;
+      case 2:
+        return this.w2;
+      case 3:
+        return this.w3;
+      case 4:
+        return this.w4;
+      case 5:
+        return this.w5;
+      case 6:
+        return this.w6;
+      default:
+        return this.w7;
+    }
+  }
+
+  setWord(word: number, bits: number): void {
+    switch (word) {
+      case 0:
+        this.w0 = bits | 0;
+        break;
+      case 1:
+        this.w1 = bits | 0;
+        break;
+      case 2:
+        this.w2 = bits | 0;
+        break;
+      case 3:
+        this.w3 = bits | 0;
+        break;
+      case 4:
+        this.w4 = bits | 0;
+        break;
+      case 5:
+        this.w5 = bits | 0;
+        break;
+      case 6:
+        this.w6 = bits | 0;
+        break;
+      default:
+        this.w7 = bits | 0;
+    }
+  }
+
+  /** The index in `values` of the value of `slot` (an id's low bits), or of where it goes. */
+  rank(slot: number, bits: number): number {
+    const word = slot >> 5;
+    const before = ((word < 4 ? this.before0 : this.before1) >>> ((word & 3) << 3)) & 0xff;
+    return before + bitCount(bits & ((1 << (slot & 31)) - 1));
+  }
+
+  /** Adds `delta` to the counts of the words after `word`. */
+  shiftCounts(word: number, delta: number): void {
+    for (let w = word + 1; w < 8; w++) {
+      const step = delta << ((w & 3) << 3);
+      if (w < 4) this.before0 = (this.before0 + step) | 0;
+      else this.before1 = (this.before1 + step) | 0;
+    }
+  }
+}
+
+/**
+ * Symbols by node id, for ids a checker resolves in one in four to one in eight of the ids of the
+ * pages it touches, as it resolves the identifiers and names of the files it checks. Pages of
+ * `LinkPaging.Size` ids keep a presence bit per id and only the symbols stored, so a symbol costs
+ * about one array slot. A `LinkTable` page costs a slot for every id, and a `Map` entry also
+ * stores its key and a share of the hash buckets. At fewer values per page than that, the page's
+ * own allocations outweigh what it saves, and a `SparseLinkTable` costs less.
+ */
+export class CompactSymbolLinkTable {
+  readonly pages: (CompactLinkPage | undefined)[] = [];
+  /** The number of symbols stored. */
+  size = 0;
+
+  get(id: number): Symbol | undefined {
+    const page = this.pages[id >> LinkPaging.Bits];
+    if (page === undefined) return undefined;
+    const slot = id & LinkPaging.Mask;
+    const bits = page.word(slot >> 5);
+    if ((bits & (1 << (slot & 31))) === 0) return undefined;
+    return page.values[page.rank(slot, bits)];
+  }
+
+  set(id: number, value: Symbol | undefined): void {
+    const index = id >> LinkPaging.Bits;
+    let page = this.pages[index];
+    if (page === undefined) {
+      if (value === undefined) return;
+      while (this.pages.length <= index) this.pages.push(undefined);
+      page = new CompactLinkPage();
+      this.pages[index] = page;
+    }
+    const slot = id & LinkPaging.Mask;
+    const word = slot >> 5;
+    const bit = 1 << (slot & 31);
+    const bits = page.word(word);
+    const at = page.rank(slot, bits);
+    const values = page.values;
+    if ((bits & bit) !== 0) {
+      if (value !== undefined) {
+        values[at] = value;
+        return;
+      }
+      for (let i = at; i < values.length - 1; i++) values[i] = values[i + 1]!;
+      values.pop();
+      page.setWord(word, bits & ~bit);
+      page.shiftCounts(word, -1);
+      this.size--;
+      return;
+    }
+    if (value === undefined) return;
+    values.push(value);
+    for (let i = values.length - 1; i > at; i--) values[i] = values[i - 1]!;
+    values[at] = value;
+    page.setWord(word, bits | bit);
+    page.shiftCounts(word, 1);
+    this.size++;
+  }
+
+  clear(): void {
+    this.pages.length = 0;
+    this.size = 0;
+  }
+}

package/src/compiler/locale-data.ts

@@ -0,0 +1,203 @@
+// Locale data of golang.org/x/text/language v0.38.0: the subtag registry tsgo validates `--locale`
+// against, and the CLDR matching results it picks a translation with.
+//
+// Registry lists are space-separated and lowercase.
+
+/** Known three-letter language subtags as a bitset over `aaa`..`zzz`, six bits per base64 character. */
+export const languages3 =
+  '/3//+////u9+/7v+3/BPX92l7h////Pc/Ow///cBFKX+/3//////3//////////4p////3EvK43////////9//////9b0HP8dd+n' +
+  'UcAIf3/n/lcI4////////////9///////////f/f////////v///////v//////v//////9//v////3///v////////9//f/////' +
+  '//////v////3///37///+/////////3v9Xdr/B0/cG8QswQJBBAUEBAw//v/fSrQFZ9m0g497d+FWVVfOwxN/N/7//fX4UGCAAB/' +
+  '//7////f34zxH/9/BEIAwWAgAACAAMAUAAkkhA9//9v3+7FBAIAZNmRwfv3IAAAAf3m3mUe2z7////8nUEEDGCQ0AA/xn9l1Z61t' +
+  't/+AAIAAAAAw3p9VS2WAuY9e0////BAAAEADWNAAAALFnElFaAQAAAAERkEAAAGEAAAAQAAAERAAQAYBYAAAFAAgoQAAAAV1tABZ' +
+  '3QyU1X9vj38AAAIAABBAAAAAEcROB091ZGCmniAAFAkgAAQAEAAAAIQAABARAAAs+/eOAIAAAAABAAAAAAAIAAEBAAAACAAAAAxo' +
+  'qWAAAAAAEACBmiABAgQgQBAAIEjhABAAAAAQAdQ9DBBCEAAAAA+uzOBAACQEAAAAAAAAAAAAAAAA///////3PPooA//3n8+XnPAE' +
+  'gIbxmWUJbOw35/9AUjEkB4Qw3PJXbjz8/9fb5////RACAEQISw339/QhPREQAQw///v1ovB93MaDAAAI7lnAHQIAw////DAGFEIA' +
+  'AAAgQQLHhAAAAAAAQAEAEgAt+fKDABAARSCBsBAYw////ZPGF853uNAARAAAAAUB16rgoQAAEEl4///PFkACFAEAAAAAQEAAMAAA' +
+  'AAQoCw25JRRigA8RQeAAAAAzQBEJFeEhABCEAEiAQhYEAkN7vT1NdERaO036zgAAgUAQRAAAAAFRWmk1dfawHd5+AABAIAAgAAUR' +
+  'AEgAAEAQACgRIg/63vDxdaBAAwABBACIdLaAAAAASQEAAQAEw3ZlTQAgAEA0WAEAQAPkixk0CEgSAYEBAigAAACwAAoBAgAAAAAA' +
+  'wj97VIACAAQAAAAAAARAAABAAAw////////////f//v/////////////f////////////7//////////////9//3////////////' +
+  '/7/+///////////////////////7////////////////d/9/8iZBs8/Bw////BACAMcffYg5yB////HRiAzn75R//fl8//z/y7Rl' +
+  '3///HBYACAAIAV1nK+d+uFxhRh/9/fHQDUQ0YxFAABAEEIAAgoAg/////7/////////7/////9///////////f////9f////////' +
+  '///v///+////f////f///f////////////////33//773////////////////9///////////9//f6/////f//nACSw/////X/+/' +
+  '/f//i755/f//21fz/////////////9wpRZAxXCdXva6/9//Ydt1////PgQBAXFlwlVfTy///fLUhFXAAiAAdpBBGFEEABYAAAAAA' +
+  'AEFYFQQAAAgBRACAYUgkx3/ZrZQlGcV773Un7NIBiBEAVIEAAAAVHn/XQ4cyG999TEDAgAAAAAJAAAAAAoAEAEAQAA/3//bf++8/' +
+  '/KEFEGGs/3neEIAAJ1CFn8f7x/77/AAACcM4ew/v/3/+33/9//P/9/+R0/HEBEAx/9/9dn/XFY471fX/9AAAAMUcCBAQAAQADlBJ' +
+  'IAw///////////////////////////////////////////////////////////////////////////////////////fvnfl7T0VC' +
+  'BDUI6fRAACAAAAA8+//vAMCAgAAAIAAAwU74QAAAAEBJWAQACAzgjGAUAEwgRgEAAAA8f///SoKE/jvUAAIIAAAAAAEECIQPABhQ' +
+  'QE2XfHDAAAQAcJAIAAQAAIUAgAAA/////////8t3P37/v+///tEQQEf/v3/9///+///bxvX8/9/f///////////3/3///f1/vF4d' +
+  'fy/9//////v///1V/9+XQhhc///nV8ZFP0nx9Fqk1/////////f/d///9b/3//DQYKw/v//927/////fQ7TF7x7/++//3///3/3/' +
+  '/7//3fd/v/Cg/W8///////////f////fG9xd/n/2d9cfWkr67BqHgAAMCQAJIRAAAAE1GQAAAQAAEAAMJYAUAgAAABAJAQAAQwNe' +
+  'X30DU9U8WQW1KpAQCAAQAgAAAAA//v+HYhUQkAKBAAjEAJCAQAAAAAAAAEAAAgAgQAxvvNJABAAAAAAAAAAwA+CAAAAAAAgCAAAA' +
+  'AfowCAQAAEw/YIAASA//59DAlAAAIoAAAAAAAAAQAAxAAkAIAAQAAAwgAAAAAEAAAAAAAAAAAAAAAAw////z+5bEQAAASGgRNnvX' +
+  'AEAAwQBBVBRAEY/P6VABAALggUVdf233xxMeVPU9XfGFBAAAAAAAsc/2fQFYDiWBQsIO6GAAAADAkQEAAAxAxIQAAAA81//1XzLc' +
+  'WjneVAVBkSY6BBAAAsWOSRHQpDDkqJJAAIw/v//fFOF9v///y39xMVvQ8///fAAgAbFzW85+3037/3LpvGARYEQXOpECQhCMgDIE' +
+  'gQCA////v5fAGgoKAZRABUxK+EAAAABkpVkACEQ6/+7AAAAEcfa0U5JRf3/jmNbVgQ94YDTPACAAAwE9RUMhvBFAiA1f/u9BAACE' +
+  'EKbRQYARAAgECEBAw/f//VgAWkIAAAACAABDDAAAAAwgwIALEGwMAPCJAAAALTuOGhMVx////ZRABQIUHw////QAAAEE4EQAcIBQ' +
+  'hbnFIOAEAAAABAAAAAAAiQiCABIAAA';
+
+export const languages2 =
+  'aa ab ae af ak am an ar as av ay az ba be bg bh bi bm bn bo br bs ca ce ch co cr cs cu cv cy da de ' +
+  'dv dz ee el en eo es et eu fa ff fi fj fo fr fy ga gd gl gn gu gv ha he hi ho hr ht hu hy hz ia id ' +
+  'ie ig ii ik in io is it iu iw ja ji jv jw ka kg ki kj kk kl km kn ko kr ks ku kv kw ky la lb lg li ' +
+  'ln lo lt lu lv mg mh mi mk ml mn mo mr ms mt my na nb nd ne ng nl nn no nr nv ny oc oj om or os pa ' +
+  'pi pl ps pt qu rm rn ro ru rw sa sc sd se sg sh si sk sl sm sn so sq sr ss st su sv sw ta te tg th ' +
+  'ti tk tl tn to tr ts tt tw ty ug uk ur uz ve vi vo wa wo xh yi yo za zh zu';
+
+export const scripts =
+  'adlm afak aghb ahom arab aran armi armn avst bali bamu bass batk beng bhks blis bopo brah brai bugi' +
+  ' buhd cakm cans cari cham cher chrs cirt copt cpmn cprt cyrl cyrs deva diak dogr dsrt dupl egyd egy' +
+  'h egyp elba elym ethi geok geor glag gong gonm goth gran grek gujr guru hanb hang hani hano hans ha' +
+  'nt hatr hebr hira hluw hmng hmnp hrkt hung inds ital jamo java jpan jurc kali kana kawi khar khmr k' +
+  'hoj kitl kits knda kore kpel kthi lana laoo latf latg latn leke lepc limb lina linb lisu loma lyci ' +
+  'lydi mahj maka mand mani marc maya medf mend merc mero mlym modi mong moon mroo mtei mult mymr nagm' +
+  ' nand narb nbat newa nkdb nkgb nkoo nshu ogam olck orkh orya osge osma ougr palm pauc pcun pelm per' +
+  'm phag phli phlp phlv phnx piqd plrd prti psin qaaa qaab qaac qaad qaae qaaf qaag qaah qaai qaaj qa' +
+  'ak qaal qaam qaan qaao qaap qaaq qaar qaas qaat qaau qaav qaaw qaax qaay qaaz qaba qabb qabc qabd q' +
+  'abe qabf qabg qabh qabi qabj qabk qabl qabm qabn qabo qabp qabq qabr qabs qabt qabu qabv qabw qabx ' +
+  'ranj rjng rohg roro runr samr sara sarb saur sgnw shaw shrd shui sidd sind sinh sogd sogo sora soyo' +
+  ' sund sunu sylo syrc syre syrj syrn tagb takr tale talu taml tang tavt telu teng tfng tglg thaa tha' +
+  'i tibt tirh tnsa toto ugar vaii visp vith wara wcho wole xpeo xsux yezi yiii zanb zinh zmth zsye zs' +
+  'ym zxxx zyyy zzzz';
+
+export const regions =
+  'aa ac ad ae af ag ai al am an ao aq ar as at au aw ax az ba bb bd be bf bg bh bi bj bl bm bn bo bq ' +
+  'br bs bt bu bv bw by bz ca cc cd cf cg ch ci ck cl cm cn co cp cq cr cs ct cu cv cw cx cy cz dd de ' +
+  'dg dj dk dm do dy dz ea ec ee eg eh er es et eu ez fi fj fk fm fo fq fr fx ga gb gd ge gf gg gh gi ' +
+  'gl gm gn gp gq gr gs gt gu gw gy hk hm hn hr ht hu hv ic id ie il im in io iq ir is it je jm jo jp ' +
+  'jt ke kg kh ki km kn kp kr kw ky kz la lb lc li lk lr ls lt lu lv ly ma mc md me mf mg mh mi mk ml ' +
+  'mm mn mo mp mq mr ms mt mu mv mw mx my mz na nc ne nf ng nh ni nl no np nq nr nt nu nz om pa pc pe ' +
+  'pf pg ph pk pl pm pn pr ps pt pu pw py pz qa qm qn qo qp qq qr qs qt qu qv qw qx qy qz re rh ro rs ' +
+  'ru rw sa sb sc sd se sg sh si sj sk sl sm sn so sr ss st su sv sx sy sz ta tc td tf tg th tj tk tl ' +
+  'tm tn to tp tr tt tv tw tz ua ug uk um un us uy uz va vc vd ve vg vi vn vu wf wk ws xa xb xc xd xe ' +
+  'xf xg xh xi xj xk xl xm xn xo xp xq xr xs xt xu xv xw xx xy xz yd ye yt yu za zm zr zw zz aaa abw a' +
+  'fg ago aia ala alb and ant are arg arm asc asm ata atf atg atn aus aut aze bdi bel ben bes bfa bgd ' +
+  'bgr bhr bhs bih blm blr blz bmu bol bra brb brn btn bur bvt bwa caf can cck che chl chn civ cmr cod' +
+  ' cog cok col com cpt cpv cri cte cub cuw cxr cym cyp cze ddr deu dga dhy dji dma dnk dom dza ecu eg' +
+  'y eri esh esp est eth fin fji flk fra fro fsm fxx gab gbr geo ggy gha gib gin glp gmb gnb gnq grc g' +
+  'rd grl gtm guf gum guy hkg hmd hnd hrv hti hun hvo idn imn ind iot irl irn irq isl isr ita jam jey ' +
+  'jor jpn jtn kaz ken kgz khm kir kna kor kwt lao lbn lbr lby lca lie lka lso ltu lux lva mac maf mar' +
+  ' mco mda mdg mdv mex mhl mid mkd mli mlt mmr mne mng mnp moz mrt msr mtq mus mwi mys myt nam ncl ne' +
+  'r nfk nga nhb nic niu nld nor npl nru ntz nzl omn pak pan pci pcn pcz per phl plw png pol pri prk p' +
+  'rt pry pse pus pyf qat qmm qnn qoo qpp qqq qrr qss qtt quu qvv qww qxx qyy qzz reu rho rou rus rwa ' +
+  'sau scg sdn sen sgp sgs shn sjm slb sle slv smr som spm srb ssd stp sun sur svk svn swe swz sxm syc' +
+  ' syr taa tca tcd tgo tha tjk tkl tkm tls tmp ton tto tun tur tuv twn tza uga ukr umi ury usa uzb va' +
+  't vct vdr ven vgb vir vnm vut wak wlf wsm xaa xbb xcc xdd xee xff xgg xhh xii xjj xkk xll xmm xnn x' +
+  'oo xpp xqq xrr xss xtt xuu xvv xww xxx xyy xzz yem ymd yug zaf zar zmb zwe zzz 001 002 003 004 005 ' +
+  '008 009 010 011 012 013 014 015 016 017 018 019 020 021 024 028 029 030 031 032 034 035 036 039 040' +
+  ' 044 048 050 051 052 053 054 056 057 060 061 064 068 070 072 074 076 084 086 090 092 096 100 104 10' +
+  '8 112 116 120 124 132 136 140 142 143 144 145 148 150 151 152 154 155 156 158 162 166 170 174 175 1' +
+  '78 180 184 188 191 192 196 202 203 204 208 212 214 218 222 226 230 231 232 233 234 238 239 242 246 ' +
+  '248 249 250 254 258 260 262 266 268 270 275 276 278 280 288 292 296 300 304 308 312 316 320 324 328' +
+  ' 332 334 336 340 344 348 352 356 360 364 368 372 376 380 384 388 392 398 400 404 408 410 414 417 41' +
+  '8 419 422 426 428 430 434 438 440 442 446 450 454 458 462 466 470 474 478 480 484 492 496 498 499 5' +
+  '00 504 508 512 516 520 524 528 530 531 533 534 535 536 540 548 554 558 562 566 570 574 578 580 581 ' +
+  '583 584 585 586 591 598 600 604 608 612 616 620 624 626 630 634 638 642 643 646 652 654 659 660 662' +
+  ' 663 666 670 674 678 682 686 688 690 694 702 703 704 705 706 710 716 720 724 728 729 732 736 740 74' +
+  '4 748 752 756 760 762 764 768 772 776 780 784 788 792 795 796 798 800 804 807 810 818 826 831 832 8' +
+  '33 834 840 850 854 858 860 862 876 882 886 887 891 894 958 959 960 961 962 963 964 965 966 967 968 ' +
+  '969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993' +
+  ' 994 995 996 997 998 999';
+
+export const variants =
+  '1606nict 1694acad 1901 1959acad 1994 1996 abl1943 akuapem alalc97 aluku ao1990 aranes arevela arevm' +
+  'da arkaika asante auvern baku1926 balanka barla basiceng bauddha bciav bcizbl biscayan biske bohori' +
+  'c boont bornholm cisaup colb1945 cornu creiss dajnko ekavsk emodeng fonipa fonkirsh fonnapa fonupa ' +
+  'fonxsamp gallo gascon grclass grital grmistr hepburn heploc hognorsk hsistemo ijekavsk itihasa ivan' +
+  'chov jauer jyutping kkcor kociewie kscor laukika lemosin lengadoc lipaw ltg1929 ltg2007 luna1918 me' +
+  'telko monoton ndyuka nedis newfound nicard njiva nulik osojs oxendict pahawh2 pahawh3 pahawh4 pamak' +
+  'a peano petr1708 pinyin polyton provenc puter rigik rozaj rumgr scotland scouse simple solba sotav ' +
+  'spanglis surmiran sursilv sutsilv synnejyl tarask tongyong tunumiit uccor ucrcor ulster unifon vaid' +
+  'ika valencia vallader vecdruka vivaraup wadegile xsistemo';
+
+//// Matching: space-separated `key=value` pairs, where numbers name a `Locale`.
+
+/** ISO 639-2 codes of languages that can select a translation, with their canonical code. */
+export const languageAliases = `
+  abk=ab aym=ay aze=az bel=be bre=br ces=cs chi=zh cos=co cze=cs deu=de eus=eu fra=fr fre=fr ger=de glg=gl
+  grn=gn hat=ht hye=hy ita=it jpn=ja kaz=kk kin=rw kir=ky kon=kg kor=ko kur=ku lat=la lin=ln ltz=lb mlg=mg
+  mon=mn oci=oc pol=pl por=pt que=qu roh=rm rus=ru spa=es tat=tt tgk=tg tuk=tk tur=tr uig=ug uzb=uz wol=wo
+  zho=zh
+`;
+
+/** Bibliographic codes x/text only canonicalizes when the whole tag parses without error. */
+export const bibliographicAliases = 'arm=hy baq=eu';
+
+/** The translation chosen for a language in a script, where there is one. */
+export const scriptMatches = `
+  ab-Cyrl=12 ay-Latn=5 az-Cyrl=12 az-Latn=12 be-Cyrl=12 br-Latn=6 cmn-Hans=1 cmn-Hant=2 co-Latn=6 crs-Latn=6
+  cs-Latn=3 de-Latn=4 es-Latn=5 eu-Latn=5 fr-Latn=6 gl-Latn=5 gn-Latn=5 gsw-Latn=4 ht-Latn=6 hy-Armn=12
+  hy-Cyrl=12 it-Latn=7 ja-Jpan=8 kg-Latn=6 kk-Cyrl=12 ko-Kore=9 ku-Latn=13 ky-Cyrl=12 la-Latn=7 lb-Latn=4
+  ln-Latn=6 lua-Latn=6 mg-Latn=6 mn-Cyrl=12 oc-Latn=6 pl-Latn=10 pt-Latn=11 qu-Latn=5 rm-Latn=4 ru-Cyrl=12
+  ru-Latn=12 rw-Latn=6 tg-Cyrl=12 tk-Cyrl=12 tk-Latn=12 tr-Latn=13 tt-Cyrl=12 ug-Hans=1 ug-Hant=2 uz-Cyrl=12
+  uz-Latn=12 wo-Latn=6 zh-Hans=1 zh-Hant=2
+`;
+
+/** The usual script of each language that can select a translation. */
+export const likelyScripts = `
+  ab=Cyrl ay=Latn az=Latn be=Cyrl br=Latn cmn=Hans co=Latn crs=Latn cs=Latn de=Latn es=Latn eu=Latn fr=Latn
+  gl=Latn gn=Latn gsw=Latn ht=Latn hy=Armn it=Latn ja=Jpan kg=Latn kk=Cyrl ko=Kore ku=Latn ky=Cyrl la=Latn
+  lb=Latn ln=Latn lua=Latn mg=Latn mn=Cyrl oc=Latn pl=Latn pt=Latn qu=Latn rm=Latn ru=Cyrl rw=Latn tg=Cyrl
+  tk=Latn tr=Latn tt=Cyrl ug=Arab uz=Latn wo=Latn zh=Hans
+`;
+
+/** Regions where a language is usually written in another script, when that changes the translation. */
+export const regionScripts = `
+  az-034=Arab az-364=Arab az-368=Arab az-IQ=Arab az-IR=Arab cmn-003=Hant cmn-005=Hant cmn-009=Hant cmn-013=Hant
+  cmn-019=Hant cmn-021=Hant cmn-035=Hant cmn-036=Hant cmn-053=Hant cmn-061=Hant cmn-096=Hant cmn-150=Hant
+  cmn-154=Hant cmn-158=Hant cmn-254=Hant cmn-258=Hant cmn-344=Hant cmn-360=Hant cmn-419=Hant cmn-446=Hant
+  cmn-458=Hant cmn-591=Hant cmn-608=Hant cmn-704=Hant cmn-740=Hant cmn-764=Hant cmn-826=Hant cmn-840=Hant
+  cmn-967=Hant cmn-AU=Hant cmn-BN=Hant cmn-EU=Hant cmn-GB=Hant cmn-GF=Hant cmn-HK=Hant cmn-ID=Hant cmn-MO=Hant
+  cmn-MY=Hant cmn-PA=Hant cmn-PF=Hant cmn-PH=Hant cmn-PZ=Hant cmn-QU=Hant cmn-SR=Hant cmn-TH=Hant cmn-TW=Hant
+  cmn-UK=Hant cmn-US=Hant cmn-VD=Hant cmn-VN=Hant kk-004=Arab kk-030=Arab kk-034=Arab kk-156=Arab kk-364=Arab
+  kk-496=Arab kk-AF=Arab kk-CN=Arab kk-IR=Arab kk-MN=Arab ku-422=Arab ku-LB=Arab ky-030=Arab ky-145=Latn
+  ky-156=Arab ky-792=Latn ky-CN=Arab ky-TR=Latn mn-156=Mong mn-CN=Mong tg-034=Arab tg-586=Arab tg-PK=Arab
+  uz-004=Arab uz-034=Arab uz-AF=Arab zh-003=Hant zh-005=Hant zh-009=Hant zh-013=Hant zh-019=Hant zh-021=Hant
+  zh-035=Hant zh-036=Hant zh-053=Hant zh-061=Hant zh-096=Hant zh-150=Hant zh-154=Hant zh-158=Hant zh-254=Hant
+  zh-258=Hant zh-344=Hant zh-360=Hant zh-419=Hant zh-446=Hant zh-458=Hant zh-591=Hant zh-608=Hant zh-704=Hant
+  zh-740=Hant zh-764=Hant zh-826=Hant zh-840=Hant zh-967=Hant zh-AU=Hant zh-BN=Hant zh-EU=Hant zh-GB=Hant
+  zh-GF=Hant zh-HK=Hant zh-ID=Hant zh-MO=Hant zh-MY=Hant zh-PA=Hant zh-PF=Hant zh-PH=Hant zh-PZ=Hant zh-QU=Hant
+  zh-SR=Hant zh-TH=Hant zh-TW=Hant zh-UK=Hant zh-US=Hant zh-VD=Hant zh-VN=Hant
+`;
+
+/** For `und-REGION`: the region's usual script and translation, where there is one. */
+export const undRegions = `
+  005=Latn:11 013=Latn:5 024=Latn:11 029=Latn:5 030=Hans:1 031=Latn:12 032=Latn:5 039=Latn:7 040=Latn:4
+  051=Armn:12 068=Latn:5 076=Latn:11 112=Cyrl:12 120=Latn:6 132=Latn:11 140=Latn:6 142=Hans:1 143=Latn:12
+  148=Latn:6 150=Cyrl:12 151=Cyrl:12 152=Latn:5 155=Latn:4 156=Hans:1 158=Hant:2 170=Latn:5 175=Latn:6
+  178=Latn:6 188=Latn:5 192=Latn:5 203=Latn:3 204=Latn:6 214=Latn:5 218=Latn:5 222=Latn:5 226=Latn:5 249=Latn:6
+  250=Latn:6 254=Latn:6 258=Latn:6 260=Latn:6 266=Latn:6 276=Latn:4 278=Latn:4 280=Latn:4 312=Latn:6 320=Latn:5
+  324=Latn:6 332=Latn:6 336=Latn:7 340=Latn:5 344=Hant:2 380=Latn:7 384=Latn:6 392=Jpan:8 398=Cyrl:12 408=Kore:9
+  410=Kore:9 417=Cyrl:12 419=Latn:5 438=Latn:4 442=Latn:6 446=Hant:2 450=Latn:6 474=Latn:6 484=Latn:5 492=Latn:6
+  496=Cyrl:12 508=Latn:11 540=Latn:6 558=Latn:5 591=Latn:5 600=Latn:5 604=Latn:5 616=Latn:10 620=Latn:11
+  624=Latn:11 626=Latn:11 630=Latn:5 638=Latn:6 643=Cyrl:12 646=Latn:6 652=Latn:6 663=Latn:6 666=Latn:6
+  674=Latn:7 678=Latn:11 686=Latn:6 690=Latn:6 724=Latn:5 756=Latn:4 762=Cyrl:12 768=Latn:6 792=Latn:13
+  795=Latn:12 854=Latn:6 858=Latn:5 860=Latn:12 862=Latn:5 876=Latn:6 AM=Armn:12 AO=Latn:11 AR=Latn:5 AT=Latn:4
+  AZ=Latn:12 BF=Latn:6 BJ=Latn:6 BL=Latn:6 BO=Latn:5 BR=Latn:11 BY=Cyrl:12 CF=Latn:6 CG=Latn:6 CH=Latn:4
+  CI=Latn:6 CL=Latn:5 CM=Latn:6 CN=Hans:1 CO=Latn:5 CR=Latn:5 CU=Latn:5 CV=Latn:11 CZ=Latn:3 DD=Latn:4 DE=Latn:4
+  DO=Latn:5 DY=Latn:6 EA=Latn:5 EC=Latn:5 ES=Latn:5 EZ=Latn:4 FR=Latn:6 FX=Latn:6 GA=Latn:6 GF=Latn:6 GN=Latn:6
+  GP=Latn:6 GQ=Latn:5 GT=Latn:5 GW=Latn:11 HK=Hant:2 HN=Latn:5 HT=Latn:6 HV=Latn:6 IC=Latn:5 IT=Latn:7 JP=Jpan:8
+  KG=Cyrl:12 KP=Kore:9 KR=Kore:9 KZ=Cyrl:12 LI=Latn:4 LU=Latn:6 MC=Latn:6 MF=Latn:6 MG=Latn:6 MN=Cyrl:12
+  MO=Hant:2 MQ=Latn:6 MX=Latn:5 MZ=Latn:11 NC=Latn:6 NI=Latn:5 PA=Latn:5 PE=Latn:5 PF=Latn:6 PL=Latn:10
+  PM=Latn:6 PR=Latn:5 PT=Latn:11 PY=Latn:5 PZ=Latn:5 RE=Latn:6 RU=Cyrl:12 RW=Latn:6 SC=Latn:6 SM=Latn:7
+  SN=Latn:6 ST=Latn:11 SV=Latn:5 TD=Latn:6 TF=Latn:6 TG=Latn:6 TJ=Cyrl:12 TL=Latn:11 TM=Latn:12 TP=Latn:11
+  TR=Latn:13 TW=Hant:2 UY=Latn:5 UZ=Latn:12 VA=Latn:7 VE=Latn:5 WF=Latn:6 YT=Latn:6
+`;
+
+/** For `und-Script`: the translation for the script's usual language, where there is one. */
+export const undScripts = 'Armn=12 Cyrl=12 Hans=1 Hant=2 Jpan=8 Kore=9';
+
+/** Tags the rules above get wrong, with x/text's answer. */
+export const exceptions = `
+  ug-156=1 ug-158=2 ug-CN=1 ug-TW=2 und-Cyrl-008=0 und-Cyrl-070=0 und-Cyrl-100=0 und-Cyrl-300=0 und-Cyrl-498=0
+  und-Cyrl-642=0 und-Cyrl-688=0 und-Cyrl-703=0 und-Cyrl-792=0 und-Cyrl-804=0 und-Cyrl-807=0 und-Cyrl-983=0
+  und-Cyrl-AL=0 und-Cyrl-BA=0 und-Cyrl-BG=0 und-Cyrl-GR=0 und-Cyrl-MD=0 und-Cyrl-MK=0 und-Cyrl-RO=0
+  und-Cyrl-RS=0 und-Cyrl-SK=0 und-Cyrl-TR=0 und-Cyrl-UA=0 und-Cyrl-XK=0 und-Latn-004=12 und-Latn-012=6
+  und-Latn-051=13 und-Latn-150=12 und-Latn-151=12 und-Latn-174=6 und-Latn-196=13 und-Latn-268=13 und-Latn-364=12
+  und-Latn-398=12 und-Latn-446=11 und-Latn-478=6 und-Latn-504=6 und-Latn-760=6 und-Latn-788=6 und-Latn-804=10
+  und-Latn-AF=12 und-Latn-AM=13 und-Latn-CY=13 und-Latn-DZ=6 und-Latn-GE=13 und-Latn-IR=12 und-Latn-KM=6
+  und-Latn-KZ=12 und-Latn-MA=6 und-Latn-MO=11 und-Latn-MR=6 und-Latn-SY=6 und-Latn-TN=6 und-Latn-UA=10
+`;
+
+/** Deprecated region codes x/text only replaces when the whole tag parses without error. */
+export const deprecatedRegions = 'BU CT DD DY FX HV JT MI NH NQ PU PZ QU RH TP UK VD WK YD ZR 249 278 720';