1
0
mirror of https://github.com/gorhill/uBlock.git synced 2024-11-02 00:42:45 +01:00

Implement bidirectional plain-string trie

The bidirectional trie allows storing the right
and left parts of a string into a trie given a
pivot position.

Releated issue:
- https://github.com/uBlockOrigin/uBlock-issues/issues/528

Additionally, the mandatory token-at-index-0 rule
for FilterPlainHnAnchored has been lifted, thus
allowing the engine to pick a potentially better token
at any position in the filter string.

***

TODO: Eventually rename `strie.js` to `biditrie.js`.

TODO: Fix dump() method, it currently only show the
      right-hand side of a filter string.
This commit is contained in:
Raymond Hill 2019-06-18 19:16:39 -04:00
parent 4a062728ff
commit cfc2ce333d
No known key found for this signature in database
GPG Key ID: 25E1490B761470C2
3 changed files with 516 additions and 260 deletions

View File

@ -137,8 +137,8 @@ const µBlock = (function() { // jshint ignore:line
// Read-only
systemSettings: {
compiledMagic: 16, // Increase when compiled format changes
selfieMagic: 16 // Increase when selfie format changes
compiledMagic: 17, // Increase when compiled format changes
selfieMagic: 17 // Increase when selfie format changes
},
restoreBackupSettings: {

View File

@ -20,7 +20,7 @@
*/
/* jshint bitwise: false */
/* global punycode, HNTrieContainer, STrieContainer */
/* global punycode, HNTrieContainer */
'use strict';
@ -130,6 +130,7 @@ const reIsWildcarded = /[\^\*]/;
// See the following as short-lived registers, used during evaluation. They are
// valid until the next evaluation.
let urlRegister = '';
let pageHostnameRegister = '';
let requestHostnameRegister = '';
@ -311,13 +312,12 @@ registerFilterClass(FilterTrue);
/******************************************************************************/
const FilterPlain = class {
constructor(s, tokenBeg) {
constructor(s) {
this.s = s;
this.tokenBeg = tokenBeg;
}
match(url, tokenBeg) {
return url.startsWith(this.s, tokenBeg - this.tokenBeg);
return url.startsWith(this.s, tokenBeg);
}
logData() {
@ -332,56 +332,55 @@ const FilterPlain = class {
return [ this.fid, this.s, this.tokenBeg ];
}
addToTrie(trie) {
trie.add(this.s, this.tokenBeg);
}
static compile(details) {
return [ FilterPlain.fid, details.f, details.tokenBeg ];
}
static load(args) {
return new FilterPlain(args[1], args[2]);
if ( args[2] === 0 ) {
return new FilterPlain(args[1]);
}
if ( args[2] === 1 ) {
return new FilterPlain1(args[1]);
}
return new FilterPlainX(args[1], args[2]);
}
static addToTrie(args, trie) {
trie.add(args[1], args[2]);
}
};
FilterPlain.trieableId = 0;
FilterPlain.prototype.trieableId = FilterPlain.trieableId;
FilterPlain.prototype.tokenBeg = 0;
registerFilterClass(FilterPlain);
/******************************************************************************/
const FilterPlainPrefix1 = class {
constructor(s) {
this.s = s;
}
const FilterPlain1 = class extends FilterPlain {
match(url, tokenBeg) {
return url.startsWith(this.s, tokenBeg - 1);
}
logData() {
return {
raw: rawToPlainStr(this.s, 0),
regex: rawToRegexStr(this.s, 0),
compiled: this.compile()
};
}
compile() {
return [ this.fid, this.s ];
}
static compile(details) {
return [ FilterPlainPrefix1.fid, details.f ];
}
static load(args) {
return new FilterPlainPrefix1(args[1]);
}
static trieableStringFromArgs(args) {
return args[1];
}
};
FilterPlainPrefix1.prototype.trieableId = 0;
FilterPlain1.prototype.tokenBeg = 1;
registerFilterClass(FilterPlainPrefix1);
const FilterPlainX = class extends FilterPlain {
constructor(s, tokenBeg) {
super(s);
this.tokenBeg = tokenBeg;
}
match(url, tokenBeg) {
return url.startsWith(this.s, tokenBeg - this.tokenBeg);
}
};
/******************************************************************************/
@ -538,32 +537,54 @@ const FilterPlainHnAnchored = class {
logData() {
return {
raw: `||${this.s}`,
regex: rawToRegexStr(this.s, 0),
regex: rawToRegexStr(this.s, this.tokenBeg),
compiled: this.compile()
};
}
compile() {
return [ this.fid, this.s ];
return [ this.fid, this.s, this.tokenBeg ];
}
addToTrie(trie) {
trie.add(this.s, this.tokenBeg);
}
static compile(details) {
return [ FilterPlainHnAnchored.fid, details.f ];
return [ FilterPlainHnAnchored.fid, details.f, details.tokenBeg ];
}
static load(args) {
if ( args[2] === 0 ) {
return new FilterPlainHnAnchored(args[1]);
}
return new FilterPlainHnAnchoredX(args[1], args[2]);
}
static trieableStringFromArgs(args) {
return args[1];
static addToTrie(args, trie) {
trie.add(args[1], args[2]);
}
};
FilterPlainHnAnchored.prototype.trieableId = 1;
FilterPlainHnAnchored.trieableId = 1;
FilterPlainHnAnchored.prototype.trieableId = FilterPlainHnAnchored.trieableId;
FilterPlainHnAnchored.prototype.tokenBeg = 0;
registerFilterClass(FilterPlainHnAnchored);
const FilterPlainHnAnchoredX = class extends FilterPlainHnAnchored {
constructor(s, tokenBeg) {
super(s);
this.tokenBeg = tokenBeg;
}
match(url, tokenBeg) {
const beg = tokenBeg - this.tokenBeg;
return url.startsWith(this.s, beg) && isHnAnchored(url, beg);
}
};
/*******************************************************************************
Filters with only one single occurrence of wildcard `*`
@ -1605,12 +1626,13 @@ const FilterBucket = class {
this.filters.push(a, b, c);
this._countTrieable();
}
this.trieResult = 0;
}
get size() {
let size = this.filters.length;
if ( this.plainPrefix1Trie !== null ) {
size += this.plainPrefix1Trie.size;
if ( this.plainTrie !== null ) {
size += this.plainTrie.size;
}
if ( this.plainHnAnchoredTrie !== null ) {
size += this.plainHnAnchoredTrie.size;
@ -1619,39 +1641,26 @@ const FilterBucket = class {
}
add(fdata) {
if ( fdata[0] === this.plainPrefix1Id ) {
if ( this.plainPrefix1Trie !== null ) {
return this.plainPrefix1Trie.add(
FilterPlainPrefix1.trieableStringFromArgs(fdata)
);
const fclass = filterClasses[fdata[0]];
if ( fclass.trieableId === 0 ) {
if ( this.plainTrie !== null ) {
return fclass.addToTrie(fdata, this.plainTrie);
}
if ( this.plainPrefix1Count === 3 ) {
this.plainPrefix1Trie = FilterBucket.trieContainer.createOne();
this._transferTrieable(
this.plainPrefix1Id,
this.plainPrefix1Trie
);
return this.plainPrefix1Trie.add(
FilterPlainPrefix1.trieableStringFromArgs(fdata)
);
if ( this.plainCount === 3 ) {
this.plainTrie = FilterBucket.trieContainer.createOne();
this._transferTrieable(0, this.plainTrie);
return fclass.addToTrie(fdata, this.plainTrie);
}
this.plainPrefix1Count += 1;
this.plainCount += 1;
}
if ( fdata[0] === this.plainHnAnchoredId ) {
if ( fclass.trieableId === 1 ) {
if ( this.plainHnAnchoredTrie !== null ) {
return this.plainHnAnchoredTrie.add(
FilterPlainHnAnchored.trieableStringFromArgs(fdata)
);
return fclass.addToTrie(fdata, this.plainHnAnchoredTrie);
}
if ( this.plainHnAnchoredCount === 3 ) {
this.plainHnAnchoredTrie = FilterBucket.trieContainer.createOne();
this._transferTrieable(
this.plainHnAnchoredId,
this.plainHnAnchoredTrie
);
return this.plainHnAnchoredTrie.add(
FilterPlainHnAnchored.trieableStringFromArgs(fdata)
);
this._transferTrieable(1, this.plainHnAnchoredTrie);
return fclass.addToTrie(fdata, this.plainHnAnchoredTrie);
}
this.plainHnAnchoredCount += 1;
}
@ -1659,22 +1668,21 @@ const FilterBucket = class {
}
match(url, tokenBeg) {
if ( this.plainPrefix1Trie !== null ) {
const pos = this.plainPrefix1Trie.matches(url, tokenBeg - 1);
if ( this.plainTrie !== null ) {
const pos = this.plainTrie.matches(url, tokenBeg);
if ( pos !== -1 ) {
this.plainPrefix1Filter.s = url.slice(tokenBeg - 1, pos);
this.f = this.plainPrefix1Filter;
this.trieResult = pos;
this.f = this.plainFilter;
this.f.tokenBeg = tokenBeg - (pos >>> 16);
return true;
}
}
if (
this.plainHnAnchoredTrie !== null &&
isHnAnchored(url, tokenBeg)
) {
if ( this.plainHnAnchoredTrie !== null ) {
const pos = this.plainHnAnchoredTrie.matches(url, tokenBeg);
if ( pos !== -1 ) {
this.plainHnAnchoredFilter.s = url.slice(tokenBeg, pos);
if ( pos !== -1 && isHnAnchored(url, pos >>> 16) ) {
this.trieResult = pos;
this.f = this.plainHnAnchoredFilter;
this.f.tokenBeg = tokenBeg - (pos >>> 16);
return true;
}
}
@ -1690,6 +1698,15 @@ const FilterBucket = class {
}
logData() {
if (
this.f === this.plainFilter ||
this.f === this.plainHnAnchoredFilter
) {
this.f.s = urlRegister.slice(
this.trieResult >>> 16,
this.trieResult & 0xFFFF
);
}
return this.f.logData();
}
@ -1697,8 +1714,8 @@ const FilterBucket = class {
return [
this.fid,
this.filters.map(filter => filter.compile(toSelfie)),
this.plainPrefix1Trie !== null &&
FilterBucket.trieContainer.compileOne(this.plainPrefix1Trie),
this.plainTrie !== null &&
FilterBucket.trieContainer.compileOne(this.plainTrie),
this.plainHnAnchoredTrie !== null &&
FilterBucket.trieContainer.compileOne(this.plainHnAnchoredTrie),
];
@ -1706,21 +1723,22 @@ const FilterBucket = class {
_countTrieable() {
for ( const f of this.filters ) {
if ( f.fid === this.plainPrefix1Id ) {
this.plainPrefix1Count += 1;
} else if ( f.fid === this.plainHnAnchoredId ) {
if ( f.trieableId === 0 ) {
this.plainCount += 1;
} else if ( f.trieableId === 1 ) {
this.plainHnAnchoredCount += 1;
}
}
}
_transferTrieable(fid, trie) {
let i = this.filters.length;
_transferTrieable(trieableId, trie) {
const filters = this.filters;
let i = filters.length;
while ( i-- ) {
const f = this.filters[i];
if ( f.fid !== fid || f.s.length > 255 ) { continue; }
trie.add(f.s);
this.filters.splice(i, 1);
const f = filters[i];
if ( f.trieableId !== trieableId || f.s.length > 255 ) { continue; }
f.addToTrie(trie);
filters.splice(i, 1);
}
}
@ -1757,7 +1775,7 @@ const FilterBucket = class {
const bucket = new FilterBucket();
bucket.filters = args[1].map(data => filterFromCompiledData(data));
if ( Array.isArray(args[2]) ) {
bucket.plainPrefix1Trie =
bucket.plainTrie =
FilterBucket.trieContainer.createOne(args[2]);
}
if ( Array.isArray(args[3]) ) {
@ -1771,17 +1789,15 @@ const FilterBucket = class {
FilterBucket.prototype.f = null;
FilterBucket.prototype.promoted = 0;
FilterBucket.prototype.plainPrefix1Id = FilterPlainPrefix1.fid;
FilterBucket.prototype.plainPrefix1Count = 0;
FilterBucket.prototype.plainPrefix1Trie = null;
FilterBucket.prototype.plainPrefix1Filter = new FilterPlainPrefix1('');
FilterBucket.prototype.plainCount = 0;
FilterBucket.prototype.plainTrie = null;
FilterBucket.prototype.plainFilter = new FilterPlainX('', 0);
FilterBucket.prototype.plainHnAnchoredId = FilterPlainHnAnchored.fid;
FilterBucket.prototype.plainHnAnchoredCount = 0;
FilterBucket.prototype.plainHnAnchoredTrie = null;
FilterBucket.prototype.plainHnAnchoredFilter = new FilterPlainHnAnchored('');
FilterBucket.prototype.plainHnAnchoredFilter = new FilterPlainHnAnchoredX('', 0);
FilterBucket.trieContainer = (function() {
FilterBucket.trieContainer = (( ) => {
let trieDetails;
try {
trieDetails = JSON.parse(
@ -1789,7 +1805,7 @@ FilterBucket.trieContainer = (function() {
);
} catch(ex) {
}
return new STrieContainer(trieDetails);
return new µBlock.BidiTrieContainer(trieDetails);
})();
registerFilterClass(FilterBucket);
@ -2215,7 +2231,6 @@ FilterParser.prototype.parse = function(raw) {
// These "bad tokens" are collated manually.
// Hostname-anchored with no wildcard always have a token index of 0.
const reHostnameToken = /^[0-9a-z]+/;
const reGoodToken = /[%0-9a-z]{2,}/g;
const reRegexToken = /[%0-9A-Za-z]{2,}/g;
const reRegexTokenAbort = /[([]/;
@ -2298,13 +2313,7 @@ FilterParser.prototype.makeToken = function() {
if ( this.f === '*' ) { return; }
let matches = null;
if ( (this.anchor & 0x4) !== 0 && this.wildcarded === false ) {
matches = reHostnameToken.exec(this.f);
}
if ( matches === null ) {
matches = this.findFirstGoodToken();
}
let matches = this.findFirstGoodToken();
if ( matches !== null ) {
this.token = matches[0];
this.tokenHash = µb.urlTokenizer.tokenHashFromString(this.token);
@ -2361,7 +2370,6 @@ FilterContainer.prototype.reset = function() {
FilterBucket.reset();
// Runtime registers
this.urlRegister = '';
this.catbitsRegister = 0;
this.tokenRegister = 0;
this.filterRegister = null;
@ -2673,8 +2681,7 @@ FilterContainer.prototype.compile = function(raw, writer) {
} else if ( parsed.anchor === 0x4 ) {
if (
parsed.wildcarded === false &&
parsed.tokenHash !== parsed.noTokenHash &&
parsed.tokenBeg === 0
parsed.tokenHash !== parsed.noTokenHash
) {
fdata = FilterPlainHnAnchored.compile(parsed);
} else {
@ -2704,8 +2711,6 @@ FilterContainer.prototype.compile = function(raw, writer) {
fdata = FilterPlainRightAnchored.compile(parsed);
} else if ( parsed.anchor === 0x3 ) {
fdata = FilterExactMatch.compile(parsed);
} else if ( parsed.tokenBeg === 1 ) {
fdata = FilterPlainPrefix1.compile(parsed);
} else {
fdata = FilterPlain.compile(parsed);
}
@ -2962,7 +2967,7 @@ FilterContainer.prototype.realmMatchString = function(
}
// Pattern-based filters
else {
const url = this.urlRegister;
const url = urlRegister;
const tokenHashes = this.urlTokenizer.getTokens();
let i = 0, tokenBeg = 0;
for (;;) {
@ -3026,7 +3031,7 @@ FilterContainer.prototype.matchStringGenericHide = function(requestURL) {
const typeBits = typeNameToTypeValue['generichide'] | 0x80000000;
// Prime tokenizer: we get a normalized URL in return.
this.urlRegister = this.urlTokenizer.setURL(requestURL);
urlRegister = this.urlTokenizer.setURL(requestURL);
this.filterRegister = null;
// These registers will be used by various filters
@ -3071,7 +3076,7 @@ FilterContainer.prototype.matchString = function(fctxt, modifiers = 0) {
const partyBits = fctxt.is3rdPartyToDoc() ? ThirdParty : FirstParty;
// Prime tokenizer: we get a normalized URL in return.
this.urlRegister = this.urlTokenizer.setURL(fctxt.url);
urlRegister = this.urlTokenizer.setURL(fctxt.url);
this.filterRegister = null;
// These registers will be used by various filters
@ -3176,6 +3181,20 @@ FilterContainer.prototype.benchmark = async function(action) {
}
};
/******************************************************************************/
FilterContainer.prototype.test = function(docURL, type, url) {
const fctxt = µb.filteringContext.duplicate();
fctxt.setDocOriginFromURL(docURL);
fctxt.setType(type);
fctxt.setURL(url);
const r = this.matchString(fctxt);
console.log(`${r}`);
if ( r !== 0 ) {
console.log(this.toLogData());
}
};
/******************************************************************************-
With default filter lists:
@ -3316,8 +3335,8 @@ FilterContainer.prototype.filterClassHistogram = function() {
filterClassDetails.set(i, { name: filterClasses[i].name, count: 0, });
}
// Artificial classes to report content of tries
filterClassDetails.set(1000, { name: 'FilterPrefix1Trie', count: 0, });
filterClassDetails.set(1001, { name: 'FilterHnAnchoredTrie', count: 0, });
filterClassDetails.set(1000, { name: 'FilterPlainTrie', count: 0, });
filterClassDetails.set(1001, { name: 'FilterPlainHnAnchoredTrie', count: 0, });
const countFilter = function(f) {
if ( f instanceof Object === false ) { return; }
@ -3332,8 +3351,8 @@ FilterContainer.prototype.filterClassHistogram = function() {
countFilter(f);
if ( f instanceof FilterBucket ) {
for ( const g of f.filters ) { countFilter(g); }
if ( f.plainPrefix1Trie !== null ) {
filterClassDetails.get(1000).count += f.plainPrefix1Trie.size;
if ( f.plainTrie !== null ) {
filterClassDetails.get(1000).count += f.plainTrie.size;
}
if ( f.plainHnAnchoredTrie !== null ) {
filterClassDetails.get(1001).count += f.plainHnAnchoredTrie.size;

View File

@ -19,13 +19,16 @@
Home: https://github.com/gorhill/uBlock
*/
/* exported STrieContainer */
'use strict';
// *****************************************************************************
// start of local namespace
{
/*******************************************************************************
A STrieContainer is mostly a large buffer in which distinct but related
A BidiTrieContainer is mostly a large buffer in which distinct but related
tries are stored. The memory layout of the buffer is as follow:
0-255: reserved
@ -35,28 +38,96 @@
268-271: offset to end of character data section (=> char1)
272: start of trie data section
+--------------+
Normal cell: | And | If "Segment info" matches:
(aka CELL) +--------------+ Goto "And"
| Or | Else
+--------------+ Goto "Or"
| Segment info |
+--------------+
+--------------+
Boundary cell: | Right And | "Right And" and/or "Left And"
(aka BCELL) +--------------+ can be 0 in last-segment condition.
| Left And |
+--------------+
| 0 |
+--------------+
Given following filters and assuming token is "ad" for all of them:
-images/ad-
/google_ad.
/images_ad.
_images/ad.
We get the following internal representation:
+-----------+ +-----------+ +---+
| |---->| |---->| 0 |
+-----------+ +-----------+ +---+ +-----------+
| 0 | +--| | | |---->| 0 |
+-----------+ | +-----------+ +---+ +-----------+
| ad | | | - | | 0 | | 0 |
+-----------+ | +-----------+ +---+ +-----------+
| | -images/ |
| +-----------+ +---+ +-----------+
+->| |---->| 0 |
+-----------+ +---+ +-----------+ +-----------+
| 0 | | |---->| |---->| 0 |
+-----------+ +---+ +-----------+ +-----------+
| . | | 0 | +--| | +--| |
+-----------+ +---+ | +-----------+ | +-----------+
| | _ | | | /google |
| +-----------+ | +-----------+
| |
| | +-----------+
| +->| 0 |
| +-----------+
| | 0 |
| +-----------+
| | /images |
| +-----------+
|
| +-----------+
+->| 0 |
+-----------+
| 0 |
+-----------+
| _images/ |
+-----------+
*/
const STRIE_PAGE_SIZE = 65536;
const PAGE_SIZE = 65536;
// i32 / i8
const STRIE_TRIE0_SLOT = 256 >>> 2; // 64 / 256
const STRIE_TRIE1_SLOT = STRIE_TRIE0_SLOT + 1; // 65 / 260
const STRIE_CHAR0_SLOT = STRIE_TRIE0_SLOT + 2; // 66 / 264
const STRIE_CHAR1_SLOT = STRIE_TRIE0_SLOT + 3; // 67 / 268
const STRIE_TRIE0_START = STRIE_TRIE0_SLOT + 4 << 2; // 272
const TRIE0_SLOT = 256 >>> 2; // 64 / 256
const TRIE1_SLOT = TRIE0_SLOT + 1; // 65 / 260
const CHAR0_SLOT = TRIE0_SLOT + 2; // 66 / 264
const CHAR1_SLOT = TRIE0_SLOT + 3; // 67 / 268
const TRIE0_START = TRIE0_SLOT + 4 << 2; // 272
const CELL_BYTE_LENGTH = 12;
const MIN_FREE_CELL_BYTE_LENGTH = CELL_BYTE_LENGTH * 4;
const CELL_AND = 0;
const CELL_OR = 1;
const BCELL_RIGHT_AND = 0;
const BCELL_LEFT_AND = 1;
const SEGMENT_INFO = 2;
const STrieContainer = class {
µBlock.BidiTrieContainer = class {
constructor(details) {
if ( details instanceof Object === false ) { details = {}; }
const len = (details.byteLength || 0) + STRIE_PAGE_SIZE-1 & ~(STRIE_PAGE_SIZE-1);
const len = (details.byteLength || 0) + PAGE_SIZE-1 & ~(PAGE_SIZE-1);
this.buf = new Uint8Array(Math.max(len, 131072));
this.buf32 = new Uint32Array(this.buf.buffer);
this.buf32[STRIE_TRIE0_SLOT] = STRIE_TRIE0_START;
this.buf32[STRIE_TRIE1_SLOT] = this.buf32[STRIE_TRIE0_SLOT];
this.buf32[STRIE_CHAR0_SLOT] = details.char0 || 65536;
this.buf32[STRIE_CHAR1_SLOT] = this.buf32[STRIE_CHAR0_SLOT];
this.buf32[TRIE0_SLOT] = TRIE0_START;
this.buf32[TRIE1_SLOT] = this.buf32[TRIE0_SLOT];
this.buf32[CHAR0_SLOT] = details.char0 || 65536;
this.buf32[CHAR1_SLOT] = this.buf32[CHAR0_SLOT];
}
//--------------------------------------------------------------------------
@ -64,43 +135,86 @@ const STrieContainer = class {
//--------------------------------------------------------------------------
reset() {
this.buf32[STRIE_TRIE1_SLOT] = this.buf32[STRIE_TRIE0_SLOT];
this.buf32[STRIE_CHAR1_SLOT] = this.buf32[STRIE_CHAR0_SLOT];
this.buf32[TRIE1_SLOT] = this.buf32[TRIE0_SLOT];
this.buf32[CHAR1_SLOT] = this.buf32[CHAR0_SLOT];
}
matches(iroot, a, al) {
const ar = a.length;
const char0 = this.buf32[STRIE_CHAR0_SLOT];
matches(iroot, a, i) {
const buf32 = this.buf32;
const buf8 = this.buf;
const char0 = buf32[CHAR0_SLOT];
const aR = a.length;
let icell = iroot;
let al = i;
let c, v, bl, n;
for (;;) {
let c = a.charCodeAt(al);
c = a.charCodeAt(al);
al += 1;
let v, bl;
// find first segment with a first-character match
for (;;) {
v = this.buf32[icell+2];
v = buf32[icell+SEGMENT_INFO];
bl = char0 + (v & 0x00FFFFFF);
if ( this.buf[bl] === c ) { break; }
icell = this.buf32[icell+0];
if ( buf8[bl] === c ) { break; }
icell = buf32[icell+CELL_OR];
if ( icell === 0 ) { return -1; }
}
// all characters in segment must match
let n = v >>> 24;
n = v >>> 24;
if ( n > 1 ) {
n -= 1;
if ( (al + n) > ar ) { return -1; }
if ( (al + n) > aR ) { return -1; }
bl += 1;
const br = bl + n;
do {
if ( a.charCodeAt(al) !== this.buf[bl] ) { return -1; }
al += 1;
bl += 1;
} while ( bl < br );
for ( let i = 0; i < n; i++ ) {
if ( a.charCodeAt(al+i) !== buf8[bl+i] ) { return -1; }
}
al += n;
}
// next segment
icell = this.buf32[icell+1];
if ( icell === 0 || this.buf32[icell+2] === 0 ) { return al; }
if ( al === ar ) { return -1; }
icell = buf32[icell+CELL_AND];
if ( /* icell === 0 || */ buf32[icell+SEGMENT_INFO] === 0 ) {
const inext = buf32[icell+BCELL_LEFT_AND];
if ( inext === 0 ) { return (i << 16) | al; }
const r = this.matchesLeft(inext, a, i);
if ( r !== -1 ) { return (r << 16) | al; }
icell = buf32[icell+CELL_AND];
if ( icell === 0 ) { return -1; }
}
if ( al === aR ) { return -1; }
}
}
matchesLeft(iroot, a, i) {
const buf32 = this.buf32;
const buf8 = this.buf;
const char0 = buf32[CHAR0_SLOT];
let icell = iroot;
let ar = i;
let c, v, br, n;
for (;;) {
ar -= 1;
c = a.charCodeAt(ar);
// find first segment with a first-character match
for (;;) {
v = buf32[icell+SEGMENT_INFO];
n = v >>> 24;
br = char0 + (v & 0x00FFFFFF) + n - 1;
if ( buf8[br] === c ) { break; }
icell = buf32[icell+CELL_OR];
if ( icell === 0 ) { return -1; }
}
// all characters in segment must match
if ( n > 1 ) {
n -= 1;
if ( n > ar ) { return -1; }
for ( let i = 1; i <= n; i++ ) {
if ( a.charCodeAt(ar-i) !== buf8[br-i] ) { return -1; }
}
ar -= n;
}
// next segment
icell = buf32[icell+CELL_AND];
if ( icell === 0 || buf32[icell+SEGMENT_INFO] === 0 ) { return ar; }
if ( ar === 0 ) { return -1; }
}
}
@ -109,14 +223,14 @@ const STrieContainer = class {
return new this.STrieRef(this, args[0], args[1]);
}
// grow buffer if needed
if ( (this.buf32[STRIE_CHAR0_SLOT] - this.buf32[STRIE_TRIE1_SLOT]) < 12 ) {
this.growBuf(12, 0);
if ( (this.buf32[CHAR0_SLOT] - this.buf32[TRIE1_SLOT]) < CELL_BYTE_LENGTH ) {
this.growBuf(CELL_BYTE_LENGTH, 0);
}
const iroot = this.buf32[STRIE_TRIE1_SLOT] >>> 2;
this.buf32[STRIE_TRIE1_SLOT] += 12;
this.buf32[iroot+0] = 0;
this.buf32[iroot+1] = 0;
this.buf32[iroot+2] = 0;
const iroot = this.buf32[TRIE1_SLOT] >>> 2;
this.buf32[TRIE1_SLOT] += CELL_BYTE_LENGTH;
this.buf32[iroot+CELL_OR] = 0;
this.buf32[iroot+CELL_AND] = 0;
this.buf32[iroot+SEGMENT_INFO] = 0;
return new this.STrieRef(this, iroot, 0);
}
@ -124,68 +238,177 @@ const STrieContainer = class {
return [ trieRef.iroot, trieRef.size ];
}
add(iroot, s) {
const lschar = s.length;
if ( lschar === 0 ) { return 0; }
let ischar = 0;
add(iroot, a, i = 0) {
const aR = a.length;
if ( aR === 0 ) { return 0; }
let icell = iroot;
// special case: first node in trie
if ( this.buf32[icell+2] === 0 ) {
this.buf32[icell+2] = this.addSegment(s.slice(ischar));
return 1;
if ( this.buf32[icell+SEGMENT_INFO] === 0 ) {
this.buf32[icell+SEGMENT_INFO] = this.addSegment(a, i, aR);
return this.addLeft(icell, a, i);
}
// grow buffer if needed
if (
(this.buf32[STRIE_CHAR0_SLOT] - this.buf32[STRIE_TRIE1_SLOT]) < 24 ||
(this.buf.length - this.buf32[STRIE_CHAR1_SLOT]) < 256
(this.buf32[CHAR0_SLOT] - this.buf32[TRIE1_SLOT]) < MIN_FREE_CELL_BYTE_LENGTH ||
(this.buf.length - this.buf32[CHAR1_SLOT]) < 256
) {
this.growBuf(24, 256);
this.growBuf(MIN_FREE_CELL_BYTE_LENGTH, 256);
}
//
const char0 = this.buf32[STRIE_CHAR0_SLOT];
const buf32 = this.buf32;
const buf8 = this.buf;
const char0 = buf32[CHAR0_SLOT];
let al = i;
let inext;
// find a matching cell: move down
for (;;) {
const vseg = this.buf32[icell+2];
const binfo = buf32[icell+SEGMENT_INFO];
// skip boundary cells
if ( vseg === 0 ) {
icell = this.buf32[icell+1];
if ( binfo === 0 ) {
icell = buf32[icell+BCELL_RIGHT_AND];
continue;
}
let isegchar0 = char0 + (vseg & 0x00FFFFFF);
let bl = char0 + (binfo & 0x00FFFFFF);
// if first character is no match, move to next descendant
if ( this.buf[isegchar0] !== s.charCodeAt(ischar) ) {
inext = this.buf32[icell+0];
if ( buf8[bl] !== a.charCodeAt(al) ) {
inext = buf32[icell+CELL_OR];
if ( inext === 0 ) {
this.buf32[icell+0] = this.addCell(0, 0, this.addSegment(s.slice(ischar)));
inext = this.addCell(0, 0, this.addSegment(a, al, aR));
buf32[icell+CELL_OR] = inext;
return this.addLeft(inext, a, i);
}
icell = inext;
continue;
}
// 1st character was tested
let bi = 1;
al += 1;
// find 1st mismatch in rest of segment
const bR = binfo >>> 24;
if ( bR !== 1 ) {
for (;;) {
if ( bi === bR ) { break; }
if ( al === aR ) { break; }
if ( buf8[bl+bi] !== a.charCodeAt(al) ) { break; }
bi += 1;
al += 1;
}
}
// all segment characters matched
if ( bi === bR ) {
// needle remainder: no
if ( al === aR ) {
return this.addLeft(icell, a, i);
}
// needle remainder: yes
inext = buf32[icell+CELL_AND];
if ( buf32[inext+CELL_AND] !== 0 ) {
icell = inext;
continue;
}
// add needle remainder
icell = this.addCell(0, 0, this.addSegment(a, al, aR));
buf32[inext+CELL_AND] = icell;
return this.addLeft(icell, a, i);
}
// some characters matched
// split current segment
bl -= char0;
buf32[icell+SEGMENT_INFO] = bi << 24 | bl;
inext = this.addCell(
buf32[icell+CELL_AND],
0,
bR - bi << 24 | bl + bi
);
buf32[icell+CELL_AND] = inext;
// needle remainder: no = need boundary cell
if ( al === aR ) {
return this.addLeft(icell, a, i);
}
// needle remainder: yes = need new cell for remaining characters
icell = this.addCell(0, 0, this.addSegment(a, al, aR));
buf32[inext+CELL_OR] = icell;
return this.addLeft(icell, a, i);
}
}
addLeft(icell, a, i) {
const buf32 = this.buf32;
// fetch boundary cell
let inext = buf32[icell+CELL_AND];
// add boundary cell if none exist
if ( inext === 0 || buf32[inext+SEGMENT_INFO] !== 0 ) {
const iboundary = this.allocateCell();
buf32[icell+CELL_AND] = iboundary;
buf32[iboundary+BCELL_RIGHT_AND] = inext;
if ( i === 0 ) { return 1; }
buf32[iboundary+BCELL_LEFT_AND] = this.allocateCell();
inext = iboundary;
}
// shortest match is always first so no point storing whatever is left
if ( buf32[inext+BCELL_LEFT_AND] === 0 ) {
return i === 0 ? 0 : 1;
}
// bail out if no left segment
if ( i === 0 ) {
buf32[inext+BCELL_LEFT_AND] = 0;
return 1;
}
// fetch root cell of left segment
icell = buf32[inext+BCELL_LEFT_AND];
// special case: first node in trie
if ( buf32[icell+SEGMENT_INFO] === 0 ) {
buf32[icell+SEGMENT_INFO] = this.addSegment(a, 0, i);
return 1;
}
const buf8 = this.buf;
const char0 = buf32[CHAR0_SLOT];
let ar = i;
// find a matching cell: move down
for (;;) {
const binfo = buf32[icell+SEGMENT_INFO];
// skip boundary cells
if ( binfo === 0 ) {
icell = buf32[icell+CELL_AND];
continue;
}
const bL = char0 + (binfo & 0x00FFFFFF);
const bR = bL + (binfo >>> 24);
let br = bR;
// if first character is no match, move to next descendant
if ( buf8[br-1] !== a.charCodeAt(ar-1) ) {
inext = buf32[icell+CELL_OR];
if ( inext === 0 ) {
inext = this.addCell(0, 0, this.addSegment(a, 0, ar));
buf32[icell+CELL_OR] = inext;
return 1;
}
icell = inext;
continue;
}
// 1st character was tested
let isegchar = 1;
ischar += 1;
br -= 1;
ar -= 1;
// find 1st mismatch in rest of segment
const lsegchar = vseg >>> 24;
if ( lsegchar !== 1 ) {
if ( br !== bL ) {
for (;;) {
if ( isegchar === lsegchar ) { break; }
if ( ischar === lschar ) { break; }
if ( this.buf[isegchar0+isegchar] !== s.charCodeAt(ischar) ) { break; }
isegchar += 1;
ischar += 1;
if ( br === bL ) { break; }
if ( ar === 0 ) { break; }
if ( buf8[br-1] !== a.charCodeAt(ar-1) ) { break; }
br -= 1;
ar -= 1;
}
}
// all segment characters matched
if ( isegchar === lsegchar ) {
inext = this.buf32[icell+1];
if ( br === bL ) {
inext = buf32[icell+CELL_AND];
// needle remainder: no
if ( ischar === lschar ) {
if ( ar === 0 ) {
// boundary cell already present
if ( inext === 0 || this.buf32[inext+2] === 0 ) { return 0; }
if ( inext === 0 || buf32[inext+SEGMENT_INFO] === 0 ) {
return 0;
}
// need boundary cell
this.buf32[icell+1] = this.addCell(0, inext, 0);
buf32[icell+CELL_AND] = this.addCell(inext, 0, 0);
}
// needle remainder: yes
else {
@ -195,28 +418,29 @@ const STrieContainer = class {
}
// boundary cell + needle remainder
inext = this.addCell(0, 0, 0);
this.buf32[icell+1] = inext;
this.buf32[inext+1] = this.addCell(0, 0, this.addSegment(s.slice(ischar)));
buf32[icell+CELL_AND] = inext;
buf32[inext+CELL_AND] =
this.addCell(0, 0, this.addSegment(a, 0, ar));
}
}
// some segment characters matched
else {
// split current cell
isegchar0 -= char0;
this.buf32[icell+2] = isegchar << 24 | isegchar0;
buf32[icell+SEGMENT_INFO] = (bR - br) << 24 | (br - char0);
inext = this.addCell(
buf32[icell+CELL_AND],
0,
this.buf32[icell+1],
lsegchar - isegchar << 24 | isegchar0 + isegchar
(br - bL) << 24 | (bL - char0)
);
this.buf32[icell+1] = inext;
buf32[icell+CELL_AND] = inext;
// needle remainder: no = need boundary cell
if ( ischar === lschar ) {
this.buf32[icell+1] = this.addCell(0, inext, 0);
if ( ar === 0 ) {
buf32[icell+CELL_AND] = this.addCell(inext, 0, 0);
}
// needle remainder: yes = need new cell for remaining characters
else {
this.buf32[inext+0] = this.addCell(0, 0, this.addSegment(s.slice(ischar)));
buf32[inext+CELL_OR] =
this.addCell(0, 0, this.addSegment(a, 0, ar));
}
}
return 1;
@ -227,7 +451,7 @@ const STrieContainer = class {
this.shrinkBuf();
return {
byteLength: this.buf.byteLength,
char0: this.buf32[STRIE_CHAR0_SLOT],
char0: this.buf32[CHAR0_SLOT],
};
}
@ -235,14 +459,14 @@ const STrieContainer = class {
if ( encoder instanceof Object ) {
return encoder.encode(
this.buf32.buffer,
this.buf32[STRIE_CHAR1_SLOT]
this.buf32[CHAR1_SLOT]
);
}
return Array.from(
new Uint32Array(
this.buf32.buffer,
0,
this.buf32[STRIE_CHAR1_SLOT] + 3 >>> 2
this.buf32[CHAR1_SLOT] + 3 >>> 2
)
);
}
@ -253,7 +477,7 @@ const STrieContainer = class {
? decoder.decodeSize(selfie)
: selfie.length << 2;
if ( byteLength === 0 ) { return false; }
byteLength = byteLength + STRIE_PAGE_SIZE-1 & ~(STRIE_PAGE_SIZE-1);
byteLength = byteLength + PAGE_SIZE-1 & ~(PAGE_SIZE-1);
if ( byteLength > this.buf.length ) {
this.buf = new Uint8Array(byteLength);
this.buf32 = new Uint32Array(this.buf.buffer);
@ -270,92 +494,100 @@ const STrieContainer = class {
// Private methods
//--------------------------------------------------------------------------
addCell(idown, iright, v) {
let icell = this.buf32[STRIE_TRIE1_SLOT];
this.buf32[STRIE_TRIE1_SLOT] = icell + 12;
allocateCell() {
let icell = this.buf32[TRIE1_SLOT];
this.buf32[TRIE1_SLOT] = icell + CELL_BYTE_LENGTH;
icell >>>= 2;
this.buf32[icell+0] = idown;
this.buf32[icell+1] = iright;
this.buf32[icell+2] = v;
this.buf32[icell+0] = 0;
this.buf32[icell+1] = 0;
this.buf32[icell+2] = 0;
return icell;
}
addSegment(segment) {
const lsegchar = segment.length;
if ( lsegchar === 0 ) { return 0; }
let char1 = this.buf32[STRIE_CHAR1_SLOT];
const isegchar = char1 - this.buf32[STRIE_CHAR0_SLOT];
let i = 0;
do {
this.buf[char1++] = segment.charCodeAt(i++);
} while ( i !== lsegchar );
this.buf32[STRIE_CHAR1_SLOT] = char1;
return (lsegchar << 24) | isegchar;
addCell(iand, ior, v) {
const icell = this.allocateCell();
this.buf32[icell+CELL_AND] = iand;
this.buf32[icell+CELL_OR] = ior;
this.buf32[icell+SEGMENT_INFO] = v;
return icell;
}
addSegment(s, l, r) {
const n = r - l;
if ( n === 0 ) { return 0; }
const buf32 = this.buf32;
const des = buf32[CHAR1_SLOT];
buf32[CHAR1_SLOT] = des + n;
const buf8 = this.buf;
for ( let i = 0; i < n; i++ ) {
buf8[des+i] = s.charCodeAt(l+i);
}
return (n << 24) | (des - buf32[CHAR0_SLOT]);
}
growBuf(trieGrow, charGrow) {
const char0 = Math.max(
(this.buf32[STRIE_TRIE1_SLOT] + trieGrow + STRIE_PAGE_SIZE-1) & ~(STRIE_PAGE_SIZE-1),
this.buf32[STRIE_CHAR0_SLOT]
(this.buf32[TRIE1_SLOT] + trieGrow + PAGE_SIZE-1) & ~(PAGE_SIZE-1),
this.buf32[CHAR0_SLOT]
);
const char1 = char0 + this.buf32[STRIE_CHAR1_SLOT] - this.buf32[STRIE_CHAR0_SLOT];
const char1 = char0 + this.buf32[CHAR1_SLOT] - this.buf32[CHAR0_SLOT];
const bufLen = Math.max(
(char1 + charGrow + STRIE_PAGE_SIZE-1) & ~(STRIE_PAGE_SIZE-1),
(char1 + charGrow + PAGE_SIZE-1) & ~(PAGE_SIZE-1),
this.buf.length
);
this.resizeBuf(bufLen, char0);
}
shrinkBuf() {
const char0 = this.buf32[STRIE_TRIE1_SLOT] + 24;
const char1 = char0 + this.buf32[STRIE_CHAR1_SLOT] - this.buf32[STRIE_CHAR0_SLOT];
const char0 = this.buf32[TRIE1_SLOT] + MIN_FREE_CELL_BYTE_LENGTH;
const char1 = char0 + this.buf32[CHAR1_SLOT] - this.buf32[CHAR0_SLOT];
const bufLen = char1 + 256;
this.resizeBuf(bufLen, char0);
}
resizeBuf(bufLen, char0) {
bufLen = bufLen + STRIE_PAGE_SIZE-1 & ~(STRIE_PAGE_SIZE-1);
bufLen = bufLen + PAGE_SIZE-1 & ~(PAGE_SIZE-1);
if (
bufLen === this.buf.length &&
char0 === this.buf32[STRIE_CHAR0_SLOT]
char0 === this.buf32[CHAR0_SLOT]
) {
return;
}
const charDataLen = this.buf32[STRIE_CHAR1_SLOT] - this.buf32[STRIE_CHAR0_SLOT];
const charDataLen = this.buf32[CHAR1_SLOT] - this.buf32[CHAR0_SLOT];
if ( bufLen !== this.buf.length ) {
const newBuf = new Uint8Array(bufLen);
newBuf.set(
new Uint8Array(
this.buf.buffer,
0,
this.buf32[STRIE_TRIE1_SLOT]
this.buf32[TRIE1_SLOT]
),
0
);
newBuf.set(
new Uint8Array(
this.buf.buffer,
this.buf32[STRIE_CHAR0_SLOT],
this.buf32[CHAR0_SLOT],
charDataLen
),
char0
);
this.buf = newBuf;
this.buf32 = new Uint32Array(this.buf.buffer);
this.buf32[STRIE_CHAR0_SLOT] = char0;
this.buf32[STRIE_CHAR1_SLOT] = char0 + charDataLen;
this.buf32[CHAR0_SLOT] = char0;
this.buf32[CHAR1_SLOT] = char0 + charDataLen;
}
if ( char0 !== this.buf32[STRIE_CHAR0_SLOT] ) {
if ( char0 !== this.buf32[CHAR0_SLOT] ) {
this.buf.set(
new Uint8Array(
this.buf.buffer,
this.buf32[STRIE_CHAR0_SLOT],
this.buf32[CHAR0_SLOT],
charDataLen
),
char0
);
this.buf32[STRIE_CHAR0_SLOT] = char0;
this.buf32[STRIE_CHAR1_SLOT] = char0 + charDataLen;
this.buf32[CHAR0_SLOT] = char0;
this.buf32[CHAR1_SLOT] = char0 + charDataLen;
}
}
};
@ -366,23 +598,23 @@ const STrieContainer = class {
*/
STrieContainer.prototype.STrieRef = class {
µBlock.BidiTrieContainer.prototype.STrieRef = class {
constructor(container, iroot, size) {
this.container = container;
this.iroot = iroot;
this.size = size;
}
add(pattern) {
if ( this.container.add(this.iroot, pattern) === 1 ) {
add(s, i = 0) {
if ( this.container.add(this.iroot, s, i) === 1 ) {
this.size += 1;
return true;
}
return false;
}
matches(a, al) {
return this.container.matches(this.iroot, a, al);
matches(a, i) {
return this.container.matches(this.iroot, a, i);
}
dump() {
@ -406,24 +638,24 @@ STrieContainer.prototype.STrieRef = class {
this.icell = this.forks.pop();
}
for (;;) {
const idown = this.container.buf32[this.icell+0];
const idown = this.container.buf32[this.icell+CELL_OR];
if ( idown !== 0 ) {
this.forks.push(idown, this.charPtr);
}
const v = this.container.buf32[this.icell+2];
let i0 = this.container.buf32[STRIE_CHAR0_SLOT] + (v & 0x00FFFFFF);
const v = this.container.buf32[this.icell+SEGMENT_INFO];
let i0 = this.container.buf32[CHAR0_SLOT] + (v & 0x00FFFFFF);
const i1 = i0 + (v >>> 24);
while ( i0 < i1 ) {
this.charBuf[this.charPtr] = this.container.buf[i0];
this.charPtr += 1;
i0 += 1;
}
this.icell = this.container.buf32[this.icell+1];
this.icell = this.container.buf32[this.icell+CELL_AND];
if ( this.icell === 0 ) {
return this.toPattern();
}
if ( this.container.buf32[this.icell+2] === 0 ) {
this.icell = this.container.buf32[this.icell+1];
if ( this.container.buf32[this.icell+SEGMENT_INFO] === 0 ) {
this.icell = this.container.buf32[this.icell+CELL_AND];
return this.toPattern();
}
}
@ -443,3 +675,8 @@ STrieContainer.prototype.STrieRef = class {
};
}
};
// end of local namespace
// *****************************************************************************
}