parse-siwe
Version:
Standalone, high-performance, validating 'ERC-4361: Sign in with Ethereum' Parser
732 lines (725 loc) • 23.6 kB
JavaScript
export { Parser, utilities, identifiers };
/*
* This is the [apg-lite](https://github.com/ldthomas/apg-lite) parser.
* Included here is only the `Parser` class and it's accompanying `identities` and `utilities`.
* See [apg-lite](https://github.com/ldthomas/apg-lite) for details and detailed examples of its use.
* The license for this code is reproduced here:
* ````
* copyright: Copyright (c) 2023 Lowell D. Thomas, all rights reserved
* license: BSD-2-Clause (https://opensource.org/licenses/BSD-2-Clause)
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice, this
* list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
* SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
* OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
* ````
*/
const Parser = function fnparser() {
const id = identifiers;
const utils = utilities;
const p = this;
const thisFileName = 'parser.js: Parser(): ';
const systemData = function systemData() {
this.state = id.ACTIVE;
this.phraseLength = 0;
this.refresh = () => {
this.state = id.ACTIVE;
this.phraseLength = 0;
};
};
p.ast = undefined;
p.stats = undefined;
p.trace = undefined;
p.callbacks = [];
let lookAhead = 0;
let treeDepth = 0;
let maxTreeDepth = 0;
let nodeHits = 0;
let maxMatched = 0;
let rules = undefined;
let udts = undefined;
let opcodes = undefined;
let chars = undefined;
let sysData = new systemData();
let ruleCallbacks = undefined;
let udtCallbacks = undefined;
let userData = undefined;
const clear = () => {
lookAhead = 0;
treeDepth = 0;
maxTreeDepth = 0;
nodeHits = 0;
maxMatched = 0;
rules = undefined;
udts = undefined;
opcodes = undefined;
chars = undefined;
sysData.refresh();
ruleCallbacks = undefined;
udtCallbacks = undefined;
userData = undefined;
};
const initializeCallbacks = () => {
const functionName = `${thisFileName}initializeCallbacks(): `;
let i;
ruleCallbacks = [];
udtCallbacks = [];
for (i = 0; i < rules.length; i += 1) {
ruleCallbacks[i] = undefined;
}
for (i = 0; i < udts.length; i += 1) {
udtCallbacks[i] = undefined;
}
let func;
const list = [];
for (i = 0; i < rules.length; i += 1) {
list.push(rules[i].lower);
}
for (i = 0; i < udts.length; i += 1) {
list.push(udts[i].lower);
}
for (const index in p.callbacks) {
if (p.callbacks.hasOwnProperty(index)) {
i = list.indexOf(index.toLowerCase());
if (i < 0) {
throw new Error(`${functionName}syntax callback '${index}' not a rule or udt name`);
}
func = p.callbacks[index] ? p.callbacks[index] : undefined;
if (typeof func === 'function' || func === undefined) {
if (i < rules.length) {
ruleCallbacks[i] = func;
} else {
udtCallbacks[i - rules.length] = func;
}
} else {
throw new Error(`${functionName}syntax callback[${index}] must be function reference or falsy)`);
}
}
}
};
p.parse = (grammar, startName, inputString, callbackData) => {
const functionName = `${thisFileName}parse(): `;
clear();
chars = utils.stringToChars(inputString);
rules = grammar.rules;
udts = grammar.udts;
const lower = startName.toLowerCase();
let startIndex = undefined;
for (const i in rules) {
if (rules.hasOwnProperty(i)) {
if (lower === rules[i].lower) {
startIndex = rules[i].index;
break;
}
}
}
if (startIndex === undefined) {
throw new Error(`${functionName}start rule name '${startRule}' not recognized`);
}
initializeCallbacks();
if (p.trace) {
p.trace.init(rules, udts, chars);
}
if (p.stats) {
p.stats.init(rules, udts);
}
if (p.ast) {
p.ast.init(rules, udts, chars);
}
userData = callbackData;
/* create a dummy opcode for the start rule */
opcodes = [
{
type: id.RNM,
index: startIndex,
},
];
/* execute the start rule */
opExecute(0, 0);
opcodes = undefined;
/* test and return the sysData */
let success = false;
switch (sysData.state) {
case id.ACTIVE:
throw new Error(`${functionName}final state should never be 'ACTIVE'`);
case id.NOMATCH:
success = false;
break;
case id.EMPTY:
case id.MATCH:
if (sysData.phraseLength === chars.length) {
success = true;
} else {
success = false;
}
break;
default:
throw new Error('unrecognized state');
}
return {
success,
state: sysData.state,
stateName: id.idName(sysData.state),
length: chars.length,
matched: sysData.phraseLength,
maxMatched,
maxTreeDepth,
nodeHits,
};
};
// The `ALT` operator.<br>
// Executes its child nodes, from left to right, until it finds a match.
// Fails if *all* of its child nodes fail.
const opALT = (opIndex, phraseIndex) => {
const op = opcodes[opIndex];
for (let i = 0; i < op.children.length; i += 1) {
opExecute(op.children[i], phraseIndex);
if (sysData.state !== id.NOMATCH) {
break;
}
}
};
// The `CAT` operator.<br>
// Executes all of its child nodes, from left to right,
// concatenating the matched phrases.
// Fails if *any* child nodes fail.
const opCAT = (opIndex, phraseIndex) => {
let success;
let astLength;
let catCharIndex;
let catPhrase;
const op = opcodes[opIndex];
if (p.ast) {
astLength = p.ast.getLength();
}
success = true;
catCharIndex = phraseIndex;
catPhrase = 0;
for (let i = 0; i < op.children.length; i += 1) {
opExecute(op.children[i], catCharIndex);
if (sysData.state === id.NOMATCH) {
success = false;
break;
} else {
catCharIndex += sysData.phraseLength;
catPhrase += sysData.phraseLength;
}
}
if (success) {
sysData.state = catPhrase === 0 ? id.EMPTY : id.MATCH;
sysData.phraseLength = catPhrase;
} else {
sysData.state = id.NOMATCH;
sysData.phraseLength = 0;
if (p.ast) {
p.ast.setLength(astLength);
}
}
};
// The `REP` operator.<br>
// Repeatedly executes its single child node,
// concatenating each of the matched phrases found.
// The number of repetitions executed and its final sysData depends
// on its `min` & `max` repetition values.
const opREP = (opIndex, phraseIndex) => {
let astLength;
let repCharIndex;
let repPhrase;
let repCount;
const op = opcodes[opIndex];
if (op.max === 0) {
// this is an empty-string acceptor
// deprecated: use the TLS empty string operator, "", instead
sysData.state = id.EMPTY;
sysData.phraseLength = 0;
return;
}
repCharIndex = phraseIndex;
repPhrase = 0;
repCount = 0;
if (p.ast) {
astLength = p.ast.getLength();
}
while (1) {
if (repCharIndex >= chars.length) {
/* exit on end of input string */
break;
}
opExecute(opIndex + 1, repCharIndex);
if (sysData.state === id.NOMATCH) {
/* always end if the child node fails */
break;
}
if (sysData.state === id.EMPTY) {
/* REP always succeeds when the child node returns an empty phrase */
/* this may not seem obvious, but that's the way it works out */
break;
}
repCount += 1;
repPhrase += sysData.phraseLength;
repCharIndex += sysData.phraseLength;
if (repCount === op.max) {
/* end on maxed out reps */
break;
}
}
/* evaluate the match count according to the min, max values */
if (sysData.state === id.EMPTY) {
sysData.state = repPhrase === 0 ? id.EMPTY : id.MATCH;
sysData.phraseLength = repPhrase;
} else if (repCount >= op.min) {
sysData.state = repPhrase === 0 ? id.EMPTY : id.MATCH;
sysData.phraseLength = repPhrase;
} else {
sysData.state = id.NOMATCH;
sysData.phraseLength = 0;
if (p.ast) {
p.ast.setLength(astLength);
}
}
};
// Validate the callback function's returned sysData values.
// It's the user's responsibility to get them right
// but `RNM` fails if not.
const validateRnmCallbackResult = (rule, sysData, charsLeft, down) => {
if (sysData.phraseLength > charsLeft) {
let str = `${thisFileName}opRNM(${rule.name}): callback function error: `;
str += `sysData.phraseLength: ${sysData.phraseLength}`;
str += ` must be <= remaining chars: ${charsLeft}`;
throw new Error(str);
}
switch (sysData.state) {
case id.ACTIVE:
if (!down) {
throw new Error(
`${thisFileName}opRNM(${rule.name}): callback function return error. ACTIVE state not allowed.`
);
}
break;
case id.EMPTY:
sysData.phraseLength = 0;
break;
case id.MATCH:
if (sysData.phraseLength === 0) {
sysData.state = id.EMPTY;
}
break;
case id.NOMATCH:
sysData.phraseLength = 0;
break;
default:
throw new Error(
`${thisFileName}opRNM(${rule.name}): callback function return error. Unrecognized return state: ${sysData.state}`
);
}
};
// The `RNM` operator.<br>
// This operator will acts as a root node for a parse tree branch below and
// returns the matched phrase to its parent.
// However, its larger responsibility is handling user-defined callback functions and `AST` nodes.
// Note that the `AST` is a separate object, but `RNM` calls its functions to create its nodes.
const opRNM = (opIndex, phraseIndex) => {
let astLength;
let astDefined;
let savedOpcodes;
const op = opcodes[opIndex];
const rule = rules[op.index];
const callback = ruleCallbacks[rule.index];
/* ignore AST in look ahead (AND or NOT operator above) */
if (!lookAhead) {
astDefined = p.ast && p.ast.ruleDefined(op.index);
if (astDefined) {
astLength = p.ast.getLength();
p.ast.down(op.index, rules[op.index].name);
}
}
if (callback) {
/* call user's callback going down the parse tree*/
const charsLeft = chars.length - phraseIndex;
callback(sysData, chars, phraseIndex, userData);
validateRnmCallbackResult(rule, sysData, charsLeft, true);
if (sysData.state === id.ACTIVE) {
savedOpcodes = opcodes;
opcodes = rule.opcodes;
opExecute(0, phraseIndex);
opcodes = savedOpcodes;
/* call user's callback going up the parse tree*/
callback(sysData, chars, phraseIndex, userData);
validateRnmCallbackResult(rule, sysData, charsLeft, false);
} /* implied else clause: just accept the callback sysData - RNM acting as UDT */
} else {
/* no callback - just execute the rule */
savedOpcodes = opcodes;
opcodes = rule.opcodes;
opExecute(0, phraseIndex, sysData);
opcodes = savedOpcodes;
}
if (!lookAhead) {
/* end AST */
if (astDefined) {
if (sysData.state === id.NOMATCH) {
p.ast.setLength(astLength);
} else {
p.ast.up(op.index, rule.name, phraseIndex, sysData.phraseLength);
}
}
}
};
// The `TRG` operator.<br>
// Succeeds if the single first character of the phrase is
// within the `min - max` range.
const opTRG = (opIndex, phraseIndex) => {
const op = opcodes[opIndex];
sysData.state = id.NOMATCH;
if (phraseIndex < chars.length) {
if (op.min <= chars[phraseIndex] && chars[phraseIndex] <= op.max) {
sysData.state = id.MATCH;
sysData.phraseLength = 1;
}
}
};
// The `TBS` operator.<br>
// Matches its pre-defined phrase against the input string.
// All characters must match exactly.
// Case-sensitive literal strings (`'string'` & `%s"string"`) are translated to `TBS`
// operators by `apg`.
// Phrase length of zero is not allowed.
// Empty phrases can only be defined with `TLS` operators.
const opTBS = (opIndex, phraseIndex) => {
const op = opcodes[opIndex];
const len = op.string.length;
sysData.state = id.NOMATCH;
if (phraseIndex + len <= chars.length) {
for (let i = 0; i < len; i += 1) {
if (chars[phraseIndex + i] !== op.string[i]) {
return;
}
}
sysData.state = id.MATCH;
sysData.phraseLength = len;
} /* implied else NOMATCH */
};
// The `TLS` operator.<br>
// Matches its pre-defined phrase against the input string.
// A case-insensitive match is attempted for ASCII alphbetical characters.
// `TLS` is the only operator that explicitly allows empty phrases.
// `apg` will fail for empty `TBS`, case-sensitive strings (`''`) or
// zero repetitions (`0*0RuleName` or `0RuleName`).
const opTLS = (opIndex, phraseIndex) => {
let code;
const op = opcodes[opIndex];
sysData.state = id.NOMATCH;
const len = op.string.length;
if (len === 0) {
/* EMPTY match allowed for TLS */
sysData.state = id.EMPTY;
return;
}
if (phraseIndex + len <= chars.length) {
for (let i = 0; i < len; i += 1) {
code = chars[phraseIndex + i];
if (code >= 65 && code <= 90) {
code += 32;
}
if (code !== op.string[i]) {
return;
}
}
sysData.state = id.MATCH;
sysData.phraseLength = len;
} /* implied else NOMATCH */
};
// Validate the callback function's returned sysData values.
// It's the user's responsibility to get it right but `UDT` fails if not.
const validateUdtCallbackResult = (udt, sysData, charsLeft) => {
if (sysData.phraseLength > charsLeft) {
let str = `${thisFileName}opUDT(${udt.name}): callback function error: `;
str += `sysData.phraseLength: ${sysData.phraseLength}`;
str += ` must be <= remaining chars: ${charsLeft}`;
throw new Error(str);
}
switch (sysData.state) {
case id.ACTIVE:
throw new Error(`${thisFileName}opUDT(${udt.name}) ACTIVE state return not allowed.`);
case id.EMPTY:
if (udt.empty) {
sysData.phraseLength = 0;
} else {
throw new Error(`${thisFileName}opUDT(${udt.name}) may not return EMPTY.`);
}
break;
case id.MATCH:
if (sysData.phraseLength === 0) {
if (udt.empty) {
sysData.state = id.EMPTY;
} else {
throw new Error(`${thisFileName}opUDT(${udt.name}) may not return EMPTY.`);
}
}
break;
case id.NOMATCH:
sysData.phraseLength = 0;
break;
default:
throw new Error(
`${thisFileName}opUDT(${udt.name}): callback function return error. Unrecognized return state: ${sysData.state}`
);
}
};
// The `UDT` operator.<br>
// Simply calls the user's callback function, but operates like `RNM` with regard to the `AST`
// and back referencing.
// There is some ambiguity here. `UDT`s act as terminals for phrase recognition but as named rules
// for `AST` nodes and back referencing.
// See [`ast.js`](./ast.html) for usage.
const opUDT = (opIndex, phraseIndex) => {
let astLength;
let astIndex;
let astDefined;
const op = opcodes[opIndex];
const udt = udts[op.index];
sysData.UdtIndex = udt.index;
/* ignore AST in look ahead */
if (!lookAhead) {
astDefined = p.ast && p.ast.udtDefined(op.index);
if (astDefined) {
astIndex = rules.length + op.index;
astLength = p.ast.getLength();
p.ast.down(astIndex, udt.name);
}
}
/* call the UDT */
const charsLeft = chars.length - phraseIndex;
udtCallbacks[op.index](sysData, chars, phraseIndex, userData);
validateUdtCallbackResult(udt, sysData, charsLeft);
if (!lookAhead) {
/* end AST */
if (astDefined) {
if (sysData.state === id.NOMATCH) {
p.ast.setLength(astLength);
} else {
p.ast.up(astIndex, udt.name, phraseIndex, sysData.phraseLength);
}
}
}
};
// The `AND` operator.<br>
// This is the positive `look ahead` operator.
// Executes its single child node, returning the EMPTY state
// if it succeedsand NOMATCH if it fails.
// *Always* backtracks on any matched phrase and returns EMPTY on success.
const opAND = (opIndex, phraseIndex) => {
lookAhead += 1;
opExecute(opIndex + 1, phraseIndex);
lookAhead -= 1;
sysData.phraseLength = 0;
switch (sysData.state) {
case id.EMPTY:
sysData.state = id.EMPTY;
break;
case id.MATCH:
sysData.state = id.EMPTY;
break;
case id.NOMATCH:
sysData.state = id.NOMATCH;
break;
default:
throw new Error(`opAND: invalid state ${sysData.state}`);
}
};
// The `NOT` operator.<br>
// This is the negative `look ahead` operator.
// Executes its single child node, returning the EMPTY state
// if it *fails* and NOMATCH if it succeeds.
// *Always* backtracks on any matched phrase and returns EMPTY
// on success (failure of its child node).
const opNOT = (opIndex, phraseIndex) => {
lookAhead += 1;
opExecute(opIndex + 1, phraseIndex);
lookAhead -= 1;
sysData.phraseLength = 0;
switch (sysData.state) {
case id.EMPTY:
case id.MATCH:
sysData.state = id.NOMATCH;
break;
case id.NOMATCH:
sysData.state = id.EMPTY;
break;
default:
throw new Error(`opNOT: invalid state ${sysData.state}`);
}
};
const opExecute = (opIndex, phraseIndex) => {
const functionName = `${thisFileName}opExecute(): `;
const op = opcodes[opIndex];
nodeHits += 1;
if (treeDepth > maxTreeDepth) {
maxTreeDepth = treeDepth;
}
treeDepth += 1;
sysData.refresh();
if (p.trace) {
p.trace.down(op, phraseIndex);
}
switch (op.type) {
case id.ALT:
opALT(opIndex, phraseIndex);
break;
case id.CAT:
opCAT(opIndex, phraseIndex);
break;
case id.REP:
opREP(opIndex, phraseIndex);
break;
case id.RNM:
opRNM(opIndex, phraseIndex);
break;
case id.TRG:
opTRG(opIndex, phraseIndex);
break;
case id.TBS:
opTBS(opIndex, phraseIndex);
break;
case id.TLS:
opTLS(opIndex, phraseIndex);
break;
case id.UDT:
opUDT(opIndex, phraseIndex);
break;
case id.AND:
opAND(opIndex, phraseIndex);
break;
case id.NOT:
opNOT(opIndex, phraseIndex);
break;
default:
throw new Error(`${functionName}unrecognized operator`);
}
if (!lookAhead) {
if (phraseIndex + sysData.phraseLength > maxMatched) {
maxMatched = phraseIndex + sysData.phraseLength;
}
}
if (p.stats) {
p.stats.collect(op, sysData);
}
if (p.trace) {
p.trace.up(op, sysData.state, phraseIndex, sysData.phraseLength);
}
treeDepth -= 1;
};
};
const utilities = {
// utility functions
stringToChars: (string) => [...string].map((cp) => cp.codePointAt(0)),
charsToString: (chars, beg, len) => {
let subChars = chars;
while (1) {
if (beg === undefined || beg < 0) {
break;
}
if (len === undefined) {
subChars = chars.slice(beg);
break;
}
if (len <= 0) {
// always an empty string
return '';
}
subChars = chars.slice(beg, beg + len);
break;
}
return String.fromCodePoint(...subChars);
},
};
const identifiers = {
// Identifies the operator type.
// NB: These must match the values in apg-js 4.3.0, apg-lib/identifiers.
/* the original ABNF operators */
ALT: 1 /* alternation */,
CAT: 2 /* concatenation */,
REP: 3 /* repetition */,
RNM: 4 /* rule name */,
TRG: 5 /* terminal range */,
TBS: 6 /* terminal binary string, case sensitive */,
TLS: 7 /* terminal literal string, case insensitive */,
/* the super set, SABNF operators */
UDT: 11 /* user-defined terminal */,
AND: 12 /* positive look ahead */,
NOT: 13 /* negative look ahead */,
// Used by the parser and the user's `RNM` and `UDT` callback functions.
// Identifies the parser state as it traverses the parse tree nodes.
// - *ACTIVE* - indicates the downward direction through the parse tree node.
// - *MATCH* - indicates the upward direction and a phrase, of length \> 0, has been successfully matched
// - *EMPTY* - indicates the upward direction and a phrase, of length = 0, has been successfully matched
// - *NOMATCH* - indicates the upward direction and the parser failed to match any phrase at all
ACTIVE: 100,
MATCH: 101,
EMPTY: 102,
NOMATCH: 103,
// Used by [`AST` translator](./ast.html) (semantic analysis) and the user's callback functions
// to indicate the direction of flow through the `AST` nodes.
// - *SEM_PRE* - indicates the downward (pre-branch) direction through the `AST` node.
// - *SEM_POST* - indicates the upward (post-branch) direction through the `AST` node.
SEM_PRE: 200,
SEM_POST: 201,
// Ignored. Retained for backwords compatibility.
SEM_OK: 300,
idName: (s) => {
switch (s) {
case identifiers.ALT:
return 'ALT';
case identifiers.CAT:
return 'CAT';
case identifiers.REP:
return 'REP';
case identifiers.RNM:
return 'RNM';
case identifiers.TRG:
return 'TRG';
case identifiers.TBS:
return 'TBS';
case identifiers.TLS:
return 'TLS';
case identifiers.UDT:
return 'UDT';
case identifiers.AND:
return 'AND';
case identifiers.NOT:
return 'NOT';
case identifiers.ACTIVE:
return 'ACTIVE';
case identifiers.EMPTY:
return 'EMPTY';
case identifiers.MATCH:
return 'MATCH';
case identifiers.NOMATCH:
return 'NOMATCH';
case identifiers.SEM_PRE:
return 'SEM_PRE';
case identifiers.SEM_POST:
return 'SEM_POST';
case identifiers.SEM_OK:
return 'SEM_OK';
default:
return 'UNRECOGNIZED STATE';
}
},
};