yaml 2.0.0-3 → 2.0.0-7

package/browser/dist/parse/lexer.js

@@ -0,0 +1,674 @@
+import { BOM, DOCUMENT, FLOW_END, SCALAR } from './cst.js';
+
+/*
+START -> stream
+
+stream
+  directive -> line-end -> stream
+  indent + line-end -> stream
+  [else] -> line-start
+
+line-end
+  comment -> line-end
+  newline -> .
+  input-end -> END
+
+line-start
+  doc-start -> doc
+  doc-end -> stream
+  [else] -> indent -> block-start
+
+block-start
+  seq-item-start -> block-start
+  explicit-key-start -> block-start
+  map-value-start -> block-start
+  [else] -> doc
+
+doc
+  line-end -> line-start
+  spaces -> doc
+  anchor -> doc
+  tag -> doc
+  flow-start -> flow -> doc
+  flow-end -> error -> doc
+  seq-item-start -> error -> doc
+  explicit-key-start -> error -> doc
+  map-value-start -> doc
+  alias -> doc
+  quote-start -> quoted-scalar -> doc
+  block-scalar-header -> line-end -> block-scalar(min) -> line-start
+  [else] -> plain-scalar(false, min) -> doc
+
+flow
+  line-end -> flow
+  spaces -> flow
+  anchor -> flow
+  tag -> flow
+  flow-start -> flow -> flow
+  flow-end -> .
+  seq-item-start -> error -> flow
+  explicit-key-start -> flow
+  map-value-start -> flow
+  alias -> flow
+  quote-start -> quoted-scalar -> flow
+  comma -> flow
+  [else] -> plain-scalar(true, 0) -> flow
+
+quoted-scalar
+  quote-end -> .
+  [else] -> quoted-scalar
+
+block-scalar(min)
+  newline + peek(indent < min) -> .
+  [else] -> block-scalar(min)
+
+plain-scalar(is-flow, min)
+  scalar-end(is-flow) -> .
+  peek(newline + (indent < min)) -> .
+  [else] -> plain-scalar(min)
+*/
+function isEmpty(ch) {
+    switch (ch) {
+        case undefined:
+        case ' ':
+        case '\n':
+        case '\r':
+        case '\t':
+            return true;
+        default:
+            return false;
+    }
+}
+const hexDigits = '0123456789ABCDEFabcdef'.split('');
+const tagChars = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz-#;/?:@&=+$_.!~*'()".split('');
+const invalidFlowScalarChars = ',[]{}'.split('');
+const invalidAnchorChars = ' ,[]{}\n\r\t'.split('');
+const isNotAnchorChar = (ch) => !ch || invalidAnchorChars.includes(ch);
+/**
+ * Splits an input string into lexical tokens, i.e. smaller strings that are
+ * easily identifiable by `tokens.tokenType()`.
+ *
+ * Lexing starts always in a "stream" context. Incomplete input may be buffered
+ * until a complete token can be emitted.
+ *
+ * In addition to slices of the original input, the following control characters
+ * may also be emitted:
+ *
+ * - `\x02` (Start of Text): A document starts with the next token
+ * - `\x18` (Cancel): Unexpected end of flow-mode (indicates an error)
+ * - `\x1f` (Unit Separator): Next token is a scalar value
+ * - `\u{FEFF}` (Byte order mark): Emitted separately outside documents
+ */
+class Lexer {
+    constructor() {
+        /**
+         * Flag indicating whether the end of the current buffer marks the end of
+         * all input
+         */
+        this.atEnd = false;
+        /**
+         * Explicit indent set in block scalar header, as an offset from the current
+         * minimum indent, so e.g. set to 1 from a header `|2+`. Set to -1 if not
+         * explicitly set.
+         */
+        this.blockScalarIndent = -1;
+        /**
+         * Block scalars that include a + (keep) chomping indicator in their header
+         * include trailing empty lines, which are otherwise excluded from the
+         * scalar's contents.
+         */
+        this.blockScalarKeep = false;
+        /** Current input */
+        this.buffer = '';
+        /**
+         * Flag noting whether the map value indicator : can immediately follow this
+         * node within a flow context.
+         */
+        this.flowKey = false;
+        /** Count of surrounding flow collection levels. */
+        this.flowLevel = 0;
+        /**
+         * Minimum level of indentation required for next lines to be parsed as a
+         * part of the current scalar value.
+         */
+        this.indentNext = 0;
+        /** Indentation level of the current line. */
+        this.indentValue = 0;
+        /** Stores the state of the lexer if reaching the end of incpomplete input */
+        this.next = null;
+        /** A pointer to `buffer`; the current position of the lexer. */
+        this.pos = 0;
+    }
+    /**
+     * Generate YAML tokens from the `source` string. If `incomplete`,
+     * a part of the last line may be left as a buffer for the next call.
+     *
+     * @returns A generator of lexical tokens
+     */
+    *lex(source, incomplete = false) {
+        if (source)
+            this.buffer = this.buffer ? this.buffer + source : source;
+        this.atEnd = !incomplete;
+        let next = this.next || 'stream';
+        while (next && (incomplete || this.hasChars(1)))
+            next = yield* this.parseNext(next);
+    }
+    atLineEnd() {
+        let i = this.pos;
+        let ch = this.buffer[i];
+        while (ch === ' ' || ch === '\t')
+            ch = this.buffer[++i];
+        if (!ch || ch === '#' || ch === '\n')
+            return true;
+        if (ch === '\r')
+            return this.buffer[i + 1] === '\n';
+        return false;
+    }
+    charAt(n) {
+        return this.buffer[this.pos + n];
+    }
+    continueScalar(offset) {
+        let ch = this.buffer[offset];
+        if (this.indentNext > 0) {
+            let indent = 0;
+            while (ch === ' ')
+                ch = this.buffer[++indent + offset];
+            if (ch === '\r') {
+                const next = this.buffer[indent + offset + 1];
+                if (next === '\n' || (!next && !this.atEnd))
+                    return offset + indent + 1;
+            }
+            return ch === '\n' || indent >= this.indentNext || (!ch && !this.atEnd)
+                ? offset + indent
+                : -1;
+        }
+        if (ch === '-' || ch === '.') {
+            const dt = this.buffer.substr(offset, 3);
+            if ((dt === '---' || dt === '...') && isEmpty(this.buffer[offset + 3]))
+                return -1;
+        }
+        return offset;
+    }
+    getLine() {
+        let end = this.buffer.indexOf('\n', this.pos);
+        if (end === -1)
+            return this.atEnd ? this.buffer.substring(this.pos) : null;
+        if (this.buffer[end - 1] === '\r')
+            end -= 1;
+        return this.buffer.substring(this.pos, end);
+    }
+    hasChars(n) {
+        return this.pos + n <= this.buffer.length;
+    }
+    setNext(state) {
+        this.buffer = this.buffer.substring(this.pos);
+        this.pos = 0;
+        this.next = state;
+        return null;
+    }
+    peek(n) {
+        return this.buffer.substr(this.pos, n);
+    }
+    *parseNext(next) {
+        switch (next) {
+            case 'stream':
+                return yield* this.parseStream();
+            case 'line-start':
+                return yield* this.parseLineStart();
+            case 'block-start':
+                return yield* this.parseBlockStart();
+            case 'doc':
+                return yield* this.parseDocument();
+            case 'flow':
+                return yield* this.parseFlowCollection();
+            case 'quoted-scalar':
+                return yield* this.parseQuotedScalar();
+            case 'block-scalar':
+                return yield* this.parseBlockScalar();
+            case 'plain-scalar':
+                return yield* this.parsePlainScalar();
+        }
+    }
+    *parseStream() {
+        let line = this.getLine();
+        if (line === null)
+            return this.setNext('stream');
+        if (line[0] === BOM) {
+            yield* this.pushCount(1);
+            line = line.substring(1);
+        }
+        if (line[0] === '%') {
+            let dirEnd = line.length;
+            const cs = line.indexOf('#');
+            if (cs !== -1) {
+                const ch = line[cs - 1];
+                if (ch === ' ' || ch === '\t')
+                    dirEnd = cs - 1;
+            }
+            while (true) {
+                const ch = line[dirEnd - 1];
+                if (ch === ' ' || ch === '\t')
+                    dirEnd -= 1;
+                else
+                    break;
+            }
+            const n = (yield* this.pushCount(dirEnd)) + (yield* this.pushSpaces(true));
+            yield* this.pushCount(line.length - n); // possible comment
+            this.pushNewline();
+            return 'stream';
+        }
+        if (this.atLineEnd()) {
+            const sp = yield* this.pushSpaces(true);
+            yield* this.pushCount(line.length - sp);
+            yield* this.pushNewline();
+            return 'stream';
+        }
+        yield DOCUMENT;
+        return yield* this.parseLineStart();
+    }
+    *parseLineStart() {
+        const ch = this.charAt(0);
+        if (!ch && !this.atEnd)
+            return this.setNext('line-start');
+        if (ch === '-' || ch === '.') {
+            if (!this.atEnd && !this.hasChars(4))
+                return this.setNext('line-start');
+            const s = this.peek(3);
+            if (s === '---' && isEmpty(this.charAt(3))) {
+                yield* this.pushCount(3);
+                this.indentValue = 0;
+                this.indentNext = 0;
+                return 'doc';
+            }
+            else if (s === '...' && isEmpty(this.charAt(3))) {
+                yield* this.pushCount(3);
+                return 'stream';
+            }
+        }
+        this.indentValue = yield* this.pushSpaces(false);
+        if (this.indentNext > this.indentValue && !isEmpty(this.charAt(1)))
+            this.indentNext = this.indentValue;
+        return yield* this.parseBlockStart();
+    }
+    *parseBlockStart() {
+        const [ch0, ch1] = this.peek(2);
+        if (!ch1 && !this.atEnd)
+            return this.setNext('block-start');
+        if ((ch0 === '-' || ch0 === '?' || ch0 === ':') && isEmpty(ch1)) {
+            const n = (yield* this.pushCount(1)) + (yield* this.pushSpaces(true));
+            this.indentNext = this.indentValue + 1;
+            this.indentValue += n;
+            return yield* this.parseBlockStart();
+        }
+        return 'doc';
+    }
+    *parseDocument() {
+        yield* this.pushSpaces(true);
+        const line = this.getLine();
+        if (line === null)
+            return this.setNext('doc');
+        let n = yield* this.pushIndicators();
+        switch (line[n]) {
+            case '#':
+                yield* this.pushCount(line.length - n);
+            // fallthrough
+            case undefined:
+                yield* this.pushNewline();
+                return yield* this.parseLineStart();
+            case '{':
+            case '[':
+                yield* this.pushCount(1);
+                this.flowKey = false;
+                this.flowLevel = 1;
+                return 'flow';
+            case '}':
+            case ']':
+                // this is an error
+                yield* this.pushCount(1);
+                return 'doc';
+            case '*':
+                yield* this.pushUntil(isNotAnchorChar);
+                return 'doc';
+            case '"':
+            case "'":
+                return yield* this.parseQuotedScalar();
+            case '|':
+            case '>':
+                n += yield* this.parseBlockScalarHeader();
+                n += yield* this.pushSpaces(true);
+                yield* this.pushCount(line.length - n);
+                yield* this.pushNewline();
+                return yield* this.parseBlockScalar();
+            default:
+                return yield* this.parsePlainScalar();
+        }
+    }
+    *parseFlowCollection() {
+        let nl, sp;
+        let indent = -1;
+        do {
+            nl = yield* this.pushNewline();
+            sp = yield* this.pushSpaces(true);
+            if (nl > 0)
+                this.indentValue = indent = sp;
+        } while (nl + sp > 0);
+        const line = this.getLine();
+        if (line === null)
+            return this.setNext('flow');
+        if ((indent !== -1 && indent < this.indentNext) ||
+            (indent === 0 &&
+                (line.startsWith('---') || line.startsWith('...')) &&
+                isEmpty(line[3]))) {
+            // Allowing for the terminal ] or } at the same (rather than greater)
+            // indent level as the initial [ or { is technically invalid, but
+            // failing here would be surprising to users.
+            const atFlowEndMarker = indent === this.indentNext - 1 &&
+                this.flowLevel === 1 &&
+                (line[0] === ']' || line[0] === '}');
+            if (!atFlowEndMarker) {
+                // this is an error
+                this.flowLevel = 0;
+                yield FLOW_END;
+                return yield* this.parseLineStart();
+            }
+        }
+        let n = 0;
+        while (line[n] === ',')
+            n += (yield* this.pushCount(1)) + (yield* this.pushSpaces(true));
+        n += yield* this.pushIndicators();
+        switch (line[n]) {
+            case undefined:
+                return 'flow';
+            case '#':
+                yield* this.pushCount(line.length - n);
+                return 'flow';
+            case '{':
+            case '[':
+                yield* this.pushCount(1);
+                this.flowKey = false;
+                this.flowLevel += 1;
+                return 'flow';
+            case '}':
+            case ']':
+                yield* this.pushCount(1);
+                this.flowKey = true;
+                this.flowLevel -= 1;
+                return this.flowLevel ? 'flow' : 'doc';
+            case '*':
+                yield* this.pushUntil(isNotAnchorChar);
+                return 'flow';
+            case '"':
+            case "'":
+                this.flowKey = true;
+                return yield* this.parseQuotedScalar();
+            case ':': {
+                const next = this.charAt(1);
+                if (this.flowKey || isEmpty(next) || next === ',') {
+                    yield* this.pushCount(1);
+                    yield* this.pushSpaces(true);
+                    return 'flow';
+                }
+            }
+            // fallthrough
+            default:
+                this.flowKey = false;
+                return yield* this.parsePlainScalar();
+        }
+    }
+    *parseQuotedScalar() {
+        const quote = this.charAt(0);
+        let end = this.buffer.indexOf(quote, this.pos + 1);
+        if (quote === "'") {
+            while (end !== -1 && this.buffer[end + 1] === "'")
+                end = this.buffer.indexOf("'", end + 2);
+        }
+        else {
+            // double-quote
+            while (end !== -1) {
+                let n = 0;
+                while (this.buffer[end - 1 - n] === '\\')
+                    n += 1;
+                if (n % 2 === 0)
+                    break;
+                end = this.buffer.indexOf('"', end + 1);
+            }
+        }
+        let nl = this.buffer.indexOf('\n', this.pos);
+        if (nl !== -1 && nl < end) {
+            while (nl !== -1 && nl < end) {
+                const cs = this.continueScalar(nl + 1);
+                if (cs === -1)
+                    break;
+                nl = this.buffer.indexOf('\n', cs);
+            }
+            if (nl !== -1 && nl < end) {
+                // this is an error caused by an unexpected unindent
+                end = nl - 1;
+            }
+        }
+        if (end === -1) {
+            if (!this.atEnd)
+                return this.setNext('quoted-scalar');
+            end = this.buffer.length;
+        }
+        yield* this.pushToIndex(end + 1, false);
+        return this.flowLevel ? 'flow' : 'doc';
+    }
+    *parseBlockScalarHeader() {
+        this.blockScalarIndent = -1;
+        this.blockScalarKeep = false;
+        let i = this.pos;
+        while (true) {
+            const ch = this.buffer[++i];
+            if (ch === '+')
+                this.blockScalarKeep = true;
+            else if (ch > '0' && ch <= '9')
+                this.blockScalarIndent = Number(ch) - 1;
+            else if (ch !== '-')
+                break;
+        }
+        return yield* this.pushUntil(ch => isEmpty(ch) || ch === '#');
+    }
+    *parseBlockScalar() {
+        let nl = this.pos - 1; // may be -1 if this.pos === 0
+        let indent = 0;
+        let ch;
+        loop: for (let i = this.pos; (ch = this.buffer[i]); ++i) {
+            switch (ch) {
+                case ' ':
+                    indent += 1;
+                    break;
+                case '\n':
+                    nl = i;
+                    indent = 0;
+                    break;
+                case '\r': {
+                    const next = this.buffer[i + 1];
+                    if (!next && !this.atEnd)
+                        return this.setNext('block-scalar');
+                    if (next === '\n')
+                        break;
+                } // fallthrough
+                default:
+                    break loop;
+            }
+        }
+        if (!ch && !this.atEnd)
+            return this.setNext('block-scalar');
+        if (indent >= this.indentNext) {
+            if (this.blockScalarIndent === -1)
+                this.indentNext = indent;
+            else
+                this.indentNext += this.blockScalarIndent;
+            do {
+                const cs = this.continueScalar(nl + 1);
+                if (cs === -1)
+                    break;
+                nl = this.buffer.indexOf('\n', cs);
+            } while (nl !== -1);
+            if (nl === -1) {
+                if (!this.atEnd)
+                    return this.setNext('block-scalar');
+                nl = this.buffer.length;
+            }
+        }
+        if (!this.blockScalarKeep) {
+            do {
+                let i = nl - 1;
+                let ch = this.buffer[i];
+                if (ch === '\r')
+                    ch = this.buffer[--i];
+                while (ch === ' ' || ch === '\t')
+                    ch = this.buffer[--i];
+                if (ch === '\n' && i >= this.pos)
+                    nl = i;
+                else
+                    break;
+            } while (true);
+        }
+        yield SCALAR;
+        yield* this.pushToIndex(nl + 1, true);
+        return yield* this.parseLineStart();
+    }
+    *parsePlainScalar() {
+        const inFlow = this.flowLevel > 0;
+        let end = this.pos - 1;
+        let i = this.pos - 1;
+        let ch;
+        while ((ch = this.buffer[++i])) {
+            if (ch === ':') {
+                const next = this.buffer[i + 1];
+                if (isEmpty(next) || (inFlow && next === ','))
+                    break;
+                end = i;
+            }
+            else if (isEmpty(ch)) {
+                const next = this.buffer[i + 1];
+                if (next === '#' || (inFlow && invalidFlowScalarChars.includes(next)))
+                    break;
+                if (ch === '\r') {
+                    if (next === '\n') {
+                        i += 1;
+                        ch = '\n';
+                    }
+                    else
+                        end = i;
+                }
+                if (ch === '\n') {
+                    const cs = this.continueScalar(i + 1);
+                    if (cs === -1)
+                        break;
+                    i = Math.max(i, cs - 2); // to advance, but still account for ' #'
+                }
+            }
+            else {
+                if (inFlow && invalidFlowScalarChars.includes(ch))
+                    break;
+                end = i;
+            }
+        }
+        if (!ch && !this.atEnd)
+            return this.setNext('plain-scalar');
+        yield SCALAR;
+        yield* this.pushToIndex(end + 1, true);
+        return inFlow ? 'flow' : 'doc';
+    }
+    *pushCount(n) {
+        if (n > 0) {
+            yield this.buffer.substr(this.pos, n);
+            this.pos += n;
+            return n;
+        }
+        return 0;
+    }
+    *pushToIndex(i, allowEmpty) {
+        const s = this.buffer.slice(this.pos, i);
+        if (s) {
+            yield s;
+            this.pos += s.length;
+            return s.length;
+        }
+        else if (allowEmpty)
+            yield '';
+        return 0;
+    }
+    *pushIndicators() {
+        switch (this.charAt(0)) {
+            case '!':
+                return ((yield* this.pushTag()) +
+                    (yield* this.pushSpaces(true)) +
+                    (yield* this.pushIndicators()));
+            case '&':
+                return ((yield* this.pushUntil(isNotAnchorChar)) +
+                    (yield* this.pushSpaces(true)) +
+                    (yield* this.pushIndicators()));
+            case ':':
+            case '?': // this is an error outside flow collections
+            case '-': // this is an error
+                if (isEmpty(this.charAt(1))) {
+                    if (this.flowLevel === 0)
+                        this.indentNext = this.indentValue + 1;
+                    return ((yield* this.pushCount(1)) +
+                        (yield* this.pushSpaces(true)) +
+                        (yield* this.pushIndicators()));
+                }
+        }
+        return 0;
+    }
+    *pushTag() {
+        if (this.charAt(1) === '<') {
+            let i = this.pos + 2;
+            let ch = this.buffer[i];
+            while (!isEmpty(ch) && ch !== '>')
+                ch = this.buffer[++i];
+            return yield* this.pushToIndex(ch === '>' ? i + 1 : i, false);
+        }
+        else {
+            let i = this.pos + 1;
+            let ch = this.buffer[i];
+            while (ch) {
+                if (tagChars.includes(ch))
+                    ch = this.buffer[++i];
+                else if (ch === '%' &&
+                    hexDigits.includes(this.buffer[i + 1]) &&
+                    hexDigits.includes(this.buffer[i + 2])) {
+                    ch = this.buffer[(i += 3)];
+                }
+                else
+                    break;
+            }
+            return yield* this.pushToIndex(i, false);
+        }
+    }
+    *pushNewline() {
+        const ch = this.buffer[this.pos];
+        if (ch === '\n')
+            return yield* this.pushCount(1);
+        else if (ch === '\r' && this.charAt(1) === '\n')
+            return yield* this.pushCount(2);
+        else
+            return 0;
+    }
+    *pushSpaces(allowTabs) {
+        let i = this.pos - 1;
+        let ch;
+        do {
+            ch = this.buffer[++i];
+        } while (ch === ' ' || (allowTabs && ch === '\t'));
+        const n = i - this.pos;
+        if (n > 0) {
+            yield this.buffer.substr(this.pos, n);
+            this.pos = i;
+        }
+        return n;
+    }
+    *pushUntil(test) {
+        let i = this.pos;
+        let ch = this.buffer[i];
+        while (!test(ch))
+            ch = this.buffer[++i];
+        return yield* this.pushToIndex(i, false);
+    }
+}
+
+export { Lexer };

package/browser/dist/parse/line-counter.js

@@ -0,0 +1,39 @@
+/**
+ * Tracks newlines during parsing in order to provide an efficient API for
+ * determining the one-indexed `{ line, col }` position for any offset
+ * within the input.
+ */
+class LineCounter {
+    constructor() {
+        this.lineStarts = [];
+        /**
+         * Should be called in ascending order. Otherwise, call
+         * `lineCounter.lineStarts.sort()` before calling `linePos()`.
+         */
+        this.addNewLine = (offset) => this.lineStarts.push(offset);
+        /**
+         * Performs a binary search and returns the 1-indexed { line, col }
+         * position of `offset`. If `line === 0`, `addNewLine` has never been
+         * called or `offset` is before the first known newline.
+         */
+        this.linePos = (offset) => {
+            let low = 0;
+            let high = this.lineStarts.length;
+            while (low < high) {
+                const mid = (low + high) >> 1; // Math.floor((low + high) / 2)
+                if (this.lineStarts[mid] < offset)
+                    low = mid + 1;
+                else
+                    high = mid;
+            }
+            if (this.lineStarts[low] === offset)
+                return { line: low + 1, col: 1 };
+            if (low === 0)
+                return { line: 0, col: offset };
+            const start = this.lineStarts[low - 1];
+            return { line: low, col: offset - start + 1 };
+        };
+    }
+}
+
+export { LineCounter };