@bufbuild/protobuf 2.12.1 → 2.13.0

package/dist/esm/txtpb/from-text.js

@@ -0,0 +1,680 @@
+// Copyright 2021-2026 Buf Technologies, Inc.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+import { ScalarType, } from "../descriptors.js";
+import { protoInt64 } from "../proto-int64.js";
+import { reflect } from "../reflect/reflect.js";
+import { isFieldError } from "../reflect/error.js";
+import { scalarZeroValue } from "../reflect/scalar.js";
+import { toBinary } from "../to-binary.js";
+import { createExtensionContainer, getExtension, hasExtension, setExtension, } from "../extensions.js";
+import { getTextEncoding } from "../wire/text-encoding.js";
+import { Reader } from "./reader.js";
+import { isGroupLike } from "./is-group-like.js";
+function makeReadContext(reader, options) {
+    return Object.assign(Object.assign({ recursionLimit: 100 }, options), { reader, depth: 0 });
+}
+/**
+ * Parse a message from the protobuf text format.
+ *
+ * Requires BigInt: throws immediately if the environment does not support it.
+ */
+export function fromText(schema, text, options) {
+    const msg = reflect(schema);
+    parseText(msg, text, options);
+    return msg.message;
+}
+/**
+ * Parse a message from the protobuf text format, merging into the target.
+ *
+ * Repeated fields are appended, singular fields are overwritten (last wins),
+ * message fields are merged, and map entries are added (overwriting an existing
+ * key).
+ *
+ * Requires BigInt: throws immediately if the environment does not support it.
+ */
+export function mergeFromText(schema, target, text, options) {
+    parseText(reflect(schema, target), text, options);
+    return target;
+}
+function parseText(msg, text, options) {
+    if (!protoInt64.supported) {
+        throw new Error("the protobuf text format requires BigInt, which is unavailable in this environment");
+    }
+    const ctx = makeReadContext(new Reader(text), options);
+    try {
+        readMessageBody(msg, ctx, "eof");
+    }
+    catch (e) {
+        if (isFieldError(e)) {
+            throw new Error(`cannot decode ${e.field()} from text format: ${e.message}`, 
+            // @ts-expect-error we use the ES2022 error CTOR option "cause" for better stack traces
+            { cause: e });
+        }
+        throw e;
+    }
+}
+/**
+ * Read a message body (its fields until `close`), guarding nesting depth. The
+ * top-level message and every nested "{...}"/"<...>" block go through here, so
+ * the recursion limit counts the root too, matching fromJson and fromBinary.
+ */
+function readMessageBody(msg, ctx, close) {
+    if (++ctx.depth > ctx.recursionLimit) {
+        throw new Error(`recursion limit of ${ctx.recursionLimit} reached decoding ${msg.desc}`);
+    }
+    readFields(msg, ctx, close);
+    ctx.depth--;
+}
+/**
+ * Read the fields of a message until `close` (the matching close token, or
+ * "eof" for the top-level message).
+ */
+function readFields(msg, ctx, close) {
+    const seen = { fields: new Set(), oneofs: new Set() };
+    for (;;) {
+        const tok = ctx.reader.peek();
+        if (tok.type === "eof") {
+            if (close !== "eof") {
+                throw new Error(`unexpected end of input, expected "${close}"`);
+            }
+            return;
+        }
+        if (close !== "eof" && tok.type === close) {
+            ctx.reader.next();
+            return;
+        }
+        readField(msg, ctx, seen);
+        // A single optional "," or ";" may follow a field. Because the next
+        // iteration treats a separator as a field name and rejects it, a doubled
+        // separator is an error, while a single trailing one is allowed.
+        consumeSeparator(ctx);
+    }
+}
+function readField(msg, ctx, seen) {
+    var _a;
+    const nameTok = ctx.reader.next();
+    if (nameTok.type === "identifier") {
+        const field = fieldByTextName(msg.desc, nameTok.value);
+        if (field !== undefined) {
+            checkSeen(field, seen);
+            readFieldValue(msg, field, ctx);
+            return;
+        }
+        // Reserved field names are silently skipped; any other unknown name is an
+        // error. This matches protobuf-go.
+        if (msg.desc.proto.reservedName.includes(nameTok.value)) {
+            skipFieldValue(ctx);
+            return;
+        }
+        throw new Error(`unknown field "${nameTok.value}" for ${msg.desc}`);
+    }
+    if (nameTok.type === "[") {
+        const name = ctx.reader.readTypeName();
+        // Inside google.protobuf.Any, a bracketed name is always a type URL; in any
+        // other message it is an extension name.
+        if (msg.desc.typeName === "google.protobuf.Any") {
+            readExpandedAny(msg, ctx, name, seen);
+            return;
+        }
+        const ext = (_a = ctx.registry) === null || _a === void 0 ? void 0 : _a.getExtension(name);
+        if (ext !== undefined && ext.extendee.typeName === msg.desc.typeName) {
+            checkSeen(ext, seen);
+            readExtensionField(msg, ext, ctx);
+            return;
+        }
+        throw new Error(`unknown extension "[${name}]" for ${msg.desc}`);
+    }
+    if (nameTok.type === "int") {
+        // Like protobuf-go, a field cannot be addressed by number, so the numbered
+        // output of printUnknownFields cannot be read back.
+        throw new Error(`cannot specify field by number: ${nameTok.text}`);
+    }
+    throw new Error(`expected a field name, got ${describe(nameTok)}`);
+}
+// Rejects a repeated occurrence of a singular field, or a second member of the
+// same oneof. Repeated and map fields may appear any number of times.
+function checkSeen(field, seen) {
+    if (field.fieldKind === "list" || field.fieldKind === "map") {
+        return;
+    }
+    if (field.oneof !== undefined) {
+        if (seen.oneofs.has(field.oneof)) {
+            throw new Error(`oneof "${field.oneof.name}" is already set`);
+        }
+        seen.oneofs.add(field.oneof);
+    }
+    if (seen.fields.has(field.number)) {
+        const what = field.kind === "extension"
+            ? `extension "[${field.typeName}]"`
+            : `field "${field.name}"`;
+        throw new Error(`non-repeated ${what} is repeated`);
+    }
+    seen.fields.add(field.number);
+}
+function readFieldValue(target, field, ctx) {
+    // The ":" separator is optional before a message, group, or map value, but
+    // required for scalars, enums, and lists of them.
+    const hasColon = consumeColon(ctx);
+    if (!colonOptional(field) && !hasColon) {
+        throw new Error(`expected ":" before value of field "${field.name}"`);
+    }
+    switch (field.fieldKind) {
+        case "scalar":
+            target.set(field, readScalarValue(field, field.scalar, ctx));
+            break;
+        case "enum":
+            target.set(field, readEnumValue(field.enum, ctx));
+            break;
+        case "message": {
+            const sub = target.isSet(field)
+                ? target.get(field)
+                : reflect(field.message);
+            readMessageValue(sub, ctx);
+            target.set(field, sub);
+            break;
+        }
+        case "list":
+            readListField(field, target.get(field), ctx);
+            break;
+        case "map":
+            readMapField(field, target.get(field), ctx);
+            break;
+    }
+}
+function readExtensionField(msg, ext, ctx) {
+    // Extensions live in the unknown-field set. We read the new value into a
+    // container seeded with any existing value, so a repeated extension appends.
+    const existing = hasExtension(msg.message, ext)
+        ? getExtension(msg.message, ext)
+        : undefined;
+    const [container, field, get] = createExtensionContainer(ext, existing);
+    readFieldValue(container, field, ctx);
+    setExtension(msg.message, ext, get());
+}
+function colonOptional(field) {
+    return (field.fieldKind === "message" ||
+        field.fieldKind === "map" ||
+        (field.fieldKind === "list" && field.listKind === "message"));
+}
+/**
+ * Read a "{ ... }" or "< ... >" block into the given message.
+ */
+function readMessageValue(msg, ctx) {
+    readMessageBody(msg, ctx, readMessageOpen(ctx));
+}
+function readMessageOpen(ctx) {
+    const open = ctx.reader.next();
+    if (open.type === "{") {
+        return "}";
+    }
+    if (open.type === "<") {
+        return ">";
+    }
+    throw new Error(`expected "{" or "<", got ${describe(open)}`);
+}
+/**
+ * Read a repeated value: either a single element, or a bracketed list
+ * "[ e, e, ... ]". This is the one place the list grammar lives, so list
+ * fields, map fields, and the reserved-skip path cannot drift in how they
+ * accept (and reject) separators.
+ */
+function readBracketedList(ctx, readElement) {
+    if (ctx.reader.peek().type !== "[") {
+        readElement();
+        return;
+    }
+    ctx.reader.next(); // "["
+    if (ctx.reader.peek().type === "]") {
+        ctx.reader.next();
+        return;
+    }
+    for (;;) {
+        readElement();
+        const sep = ctx.reader.next();
+        if (sep.type === "]") {
+            return;
+        }
+        if (sep.type !== ",") {
+            throw new Error(`expected "," or "]" in list, got ${describe(sep)}`);
+        }
+    }
+}
+function readListField(field, list, ctx) {
+    readBracketedList(ctx, () => list.add(readListItem(field, ctx)));
+}
+function readListItem(field, ctx) {
+    switch (field.listKind) {
+        case "scalar":
+            return readScalarValue(field, field.scalar, ctx);
+        case "enum":
+            return readEnumValue(field.enum, ctx);
+        case "message": {
+            const sub = reflect(field.message);
+            readMessageValue(sub, ctx);
+            return sub;
+        }
+    }
+}
+function readMapField(field, map, ctx) {
+    readBracketedList(ctx, () => readMapEntry(field, map, ctx));
+}
+/**
+ * Read a map entry: a "{ key: ... value: ... }" block. A missing key or value
+ * defaults to the zero value, like protobuf-go. A duplicate "key" or "value"
+ * within one entry is an error; duplicate keys across separate entries are
+ * legal, with the last entry winning (handled by the caller's map.set).
+ */
+function readMapEntry(field, map, ctx) {
+    if (++ctx.depth > ctx.recursionLimit) {
+        throw new Error(`recursion limit of ${ctx.recursionLimit} reached decoding a map entry`);
+    }
+    const close = readMessageOpen(ctx);
+    let key = scalarZeroValue(field.mapKey, false);
+    let value = mapValueZero(field);
+    let keySeen = false;
+    let valueSeen = false;
+    for (;;) {
+        const tok = ctx.reader.peek();
+        if (tok.type === close) {
+            ctx.reader.next();
+            break;
+        }
+        if (tok.type === "eof") {
+            throw new Error(`unexpected end of input, expected "${close}"`);
+        }
+        const nameTok = ctx.reader.next();
+        if (nameTok.type !== "identifier") {
+            throw new Error(`expected "key" or "value", got ${describe(nameTok)}`);
+        }
+        if (nameTok.value === "key") {
+            if (keySeen) {
+                throw new Error('map entry "key" is already set');
+            }
+            keySeen = true;
+            requireColon(ctx);
+            key = readScalarValue(field, field.mapKey, ctx);
+        }
+        else if (nameTok.value === "value") {
+            if (valueSeen) {
+                throw new Error('map entry "value" is already set');
+            }
+            valueSeen = true;
+            value = readMapValue(field, ctx);
+        }
+        else {
+            throw new Error(`unknown field "${nameTok.value}" in map entry`);
+        }
+        consumeSeparator(ctx);
+    }
+    ctx.depth--;
+    map.set(key, value);
+}
+function readMapValue(field, ctx) {
+    switch (field.mapKind) {
+        case "scalar":
+            requireColon(ctx);
+            return readScalarValue(field, field.scalar, ctx);
+        case "enum":
+            requireColon(ctx);
+            return readEnumValue(field.enum, ctx);
+        case "message": {
+            consumeColon(ctx);
+            const sub = reflect(field.message);
+            readMessageValue(sub, ctx);
+            return sub;
+        }
+    }
+}
+function mapValueZero(field) {
+    switch (field.mapKind) {
+        case "scalar":
+            return scalarZeroValue(field.scalar, false);
+        case "enum":
+            return field.enum.values[0].number;
+        case "message":
+            return reflect(field.message);
+    }
+}
+/**
+ * Read `google.protobuf.Any` in its expanded form `[type.url]: { ... }`.
+ *
+ * The expanded form is mutually exclusive with the raw `type_url` (field 1) and
+ * `value` (field 2) fields, and may appear only once. We enforce that through
+ * the same seen-set the duplicate-field check uses: the expansion is rejected
+ * if either field is already set, and it marks both as set so a following
+ * `type_url` or `value` is rejected too.
+ */
+function readExpandedAny(msg, ctx, typeUrl, seen) {
+    var _a;
+    if (seen.fields.has(1) || seen.fields.has(2)) {
+        throw new Error("google.protobuf.Any cannot mix the expanded form with type_url/value");
+    }
+    const slash = typeUrl.lastIndexOf("/");
+    const typeName = slash >= 0 ? typeUrl.substring(slash + 1) : typeUrl;
+    const desc = (_a = ctx.registry) === null || _a === void 0 ? void 0 : _a.getMessage(typeName);
+    if (desc === undefined) {
+        throw new Error(`unable to resolve "${typeUrl}" for google.protobuf.Any`);
+    }
+    consumeColon(ctx);
+    const unpacked = reflect(desc);
+    readMessageValue(unpacked, ctx);
+    const any = msg.message;
+    // Preserve the exact type URL, including any custom domain prefix.
+    any.typeUrl = typeUrl;
+    any.value = toBinary(desc, unpacked.message);
+    seen.fields.add(1);
+    seen.fields.add(2);
+}
+// Consume an optional leading "-" sign and report whether one was present.
+// This sees a sign token only before a number (the scanner glues a sign onto an
+// identifier as in "-inf"), and whitespace and comments between the sign and the
+// number are insignificant, so "- 42" means -42, matching protobuf-go.
+function consumeSign(ctx) {
+    if (ctx.reader.peek().type === "-") {
+        ctx.reader.next();
+        return true;
+    }
+    return false;
+}
+function readScalarValue(field, type, ctx) {
+    const negative = consumeSign(ctx);
+    const tok = ctx.reader.next();
+    switch (type) {
+        case ScalarType.STRING:
+        case ScalarType.BYTES: {
+            if (negative) {
+                throw new Error("a string value cannot have a sign");
+            }
+            if (tok.type !== "string") {
+                throw new Error(`expected a string, got ${describe(tok)}`);
+            }
+            const bytes = concatStrings(tok, ctx);
+            if (type === ScalarType.BYTES) {
+                return bytes;
+            }
+            try {
+                return getTextEncoding().decodeUtf8(bytes, field.utf8Validation);
+            }
+            catch (_a) {
+                throw new Error("invalid UTF-8 in string");
+            }
+        }
+        case ScalarType.BOOL:
+            return readBoolValue(tok, negative);
+        case ScalarType.FLOAT:
+            // Round to 32-bit precision: an out-of-range value becomes ±inf, which is
+            // what the text format requires for float overflow.
+            return Math.fround(readFloatValue(tok, negative));
+        case ScalarType.DOUBLE:
+            return readFloatValue(tok, negative);
+        case ScalarType.UINT32:
+        case ScalarType.FIXED32:
+            return Number(readUnsignedInt(tok, negative));
+        case ScalarType.UINT64:
+        case ScalarType.FIXED64:
+            return readUnsignedInt(tok, negative);
+        case ScalarType.INT64:
+        case ScalarType.SINT64:
+        case ScalarType.SFIXED64:
+            return readSignedInt(tok, negative);
+        default:
+            // INT32, SINT32, SFIXED32: the reflect layer range-checks the number.
+            return Number(readSignedInt(tok, negative));
+    }
+}
+function readEnumValue(descEnum, ctx) {
+    const negative = consumeSign(ctx);
+    const tok = ctx.reader.next();
+    if (tok.type === "identifier") {
+        if (negative) {
+            throw new Error(`invalid enum value "-${tok.value}" for ${descEnum}`);
+        }
+        const value = descEnum.values.find((v) => v.name === tok.value);
+        if (value === undefined) {
+            throw new Error(`unknown enum value "${tok.value}" for ${descEnum}`);
+        }
+        return value.number;
+    }
+    if (tok.type === "int") {
+        // The reflect layer validates the number: any int32 for open enums, a known
+        // value for closed (proto2) enums.
+        return Number(readSignedInt(tok, negative));
+    }
+    throw new Error(`expected an enum value for ${descEnum}, got ${describe(tok)}`);
+}
+function readBoolValue(tok, negative) {
+    if (!negative) {
+        if (tok.type === "identifier") {
+            switch (tok.value) {
+                case "true":
+                case "True":
+                case "t":
+                    return true;
+                case "false":
+                case "False":
+                case "f":
+                    return false;
+            }
+        }
+        if (tok.type === "int") {
+            const value = intTokenToBigInt(tok);
+            if (value === BigInt(0)) {
+                return false;
+            }
+            if (value === BigInt(1)) {
+                return true;
+            }
+        }
+    }
+    throw new Error(`expected a bool, got ${negative ? "-" : ""}${describe(tok)}`);
+}
+function readFloatValue(tok, negative) {
+    if (tok.type === "identifier") {
+        // A separate "-" token (negative) before a float literal is invalid; the
+        // only signed literals are "-inf"/"-infinity", which the scanner glues into
+        // one identifier token. "-nan" is not a literal, so it falls through to the
+        // error below. This matches protobuf-go's identifier-path sign handling.
+        if (negative) {
+            throw new Error(`invalid float value "-${tok.value}"`);
+        }
+        switch (tok.value.toLowerCase()) {
+            case "inf":
+            case "infinity":
+                return Number.POSITIVE_INFINITY;
+            case "-inf":
+            case "-infinity":
+                return Number.NEGATIVE_INFINITY;
+            case "nan":
+                return Number.NaN;
+        }
+        throw new Error(`invalid float value "${tok.value}"`);
+    }
+    if (tok.type === "float") {
+        const n = Number(tok.text);
+        return negative ? -n : n;
+    }
+    if (tok.type === "int") {
+        // Octal and hexadecimal literals are not valid for float and double fields.
+        if (tok.base !== 10) {
+            throw new Error("octal and hexadecimal are not valid for a float field");
+        }
+        const n = Number(tok.text);
+        return negative ? -n : n;
+    }
+    throw new Error(`expected a float, got ${describe(tok)}`);
+}
+function readSignedInt(tok, negative) {
+    if (tok.type !== "int") {
+        throw new Error(`expected an integer, got ${describe(tok)}`);
+    }
+    const value = intTokenToBigInt(tok);
+    return negative ? -value : value;
+}
+function readUnsignedInt(tok, negative) {
+    // Reject any sign for an unsigned field, including "-0": the reflect layer
+    // would silently accept it as 0.
+    if (negative) {
+        throw new Error("an unsigned field does not accept a negative value");
+    }
+    return readSignedInt(tok, false);
+}
+function intTokenToBigInt(tok) {
+    // Octal text keeps its leading "0" (e.g. "0755"), which BigInt would read as
+    // decimal, so it needs the "0o" prefix. Hex ("0x...") and decimal text are
+    // accepted by BigInt as-is.
+    return tok.base === 8
+        ? BigInt("0o" + tok.text.substring(1))
+        : BigInt(tok.text);
+}
+// Concatenate adjacent string literals into a single byte string.
+function concatStrings(first, ctx) {
+    if (ctx.reader.peek().type !== "string") {
+        return first.value;
+    }
+    const parts = [first.value];
+    let length = first.value.length;
+    while (ctx.reader.peek().type === "string") {
+        const tok = ctx.reader.next();
+        parts.push(tok.value);
+        length += tok.value.length;
+    }
+    const bytes = new Uint8Array(length);
+    let offset = 0;
+    for (const part of parts) {
+        bytes.set(part, offset);
+        offset += part.length;
+    }
+    return bytes;
+}
+/**
+ * Skip the value of a reserved field. Like every other nested read, the message
+ * case is guarded by the recursion limit.
+ */
+function skipFieldValue(ctx) {
+    consumeColon(ctx);
+    skipValue(ctx);
+}
+function skipValue(ctx) {
+    readBracketedList(ctx, () => skipSingleValue(ctx));
+}
+function skipSingleValue(ctx) {
+    const tok = ctx.reader.peek();
+    if (tok.type === "{" || tok.type === "<") {
+        skipMessageBlock(ctx);
+        return;
+    }
+    // A leading sign is consumed leniently here: skipping a reserved value should
+    // tolerate "- 5".
+    consumeSign(ctx);
+    const value = ctx.reader.next();
+    if (value.type === "string") {
+        while (ctx.reader.peek().type === "string") {
+            ctx.reader.next();
+        }
+        return;
+    }
+    if (value.type !== "identifier" &&
+        value.type !== "int" &&
+        value.type !== "float") {
+        throw new Error(`expected a value, got ${describe(value)}`);
+    }
+}
+function skipMessageBlock(ctx) {
+    if (++ctx.depth > ctx.recursionLimit) {
+        throw new Error(`recursion limit of ${ctx.recursionLimit} reached skipping a reserved field`);
+    }
+    const close = readMessageOpen(ctx);
+    for (;;) {
+        const tok = ctx.reader.peek();
+        if (tok.type === close) {
+            ctx.reader.next();
+            ctx.depth--;
+            return;
+        }
+        if (tok.type === "eof") {
+            throw new Error(`unexpected end of input, expected "${close}"`);
+        }
+        const nameTok = ctx.reader.next();
+        if (nameTok.type === "[") {
+            ctx.reader.readTypeName();
+        }
+        else if (nameTok.type !== "identifier" && nameTok.type !== "int") {
+            throw new Error(`expected a field name, got ${describe(nameTok)}`);
+        }
+        skipFieldValue(ctx);
+        consumeSeparator(ctx);
+    }
+}
+// Consume an optional "," or ";" that separates fields or list/map elements.
+function consumeSeparator(ctx) {
+    const sep = ctx.reader.peek();
+    if (sep.type === "," || sep.type === ";") {
+        ctx.reader.next();
+    }
+}
+function consumeColon(ctx) {
+    if (ctx.reader.peek().type === ":") {
+        ctx.reader.next();
+        return true;
+    }
+    return false;
+}
+function requireColon(ctx) {
+    if (!consumeColon(ctx)) {
+        throw new Error(`expected ":", got ${describe(ctx.reader.peek())}`);
+    }
+}
+const textFieldCache = new WeakMap();
+/**
+ * Resolve a field by its text format name, mirroring protobuf-go's ByTextName:
+ * group-like fields are addressed by their message type name, with the
+ * lowercase form as an alias; JSON names are not in this table.
+ */
+function fieldByTextName(desc, name) {
+    let byText = textFieldCache.get(desc);
+    if (byText === undefined) {
+        byText = new Map();
+        for (const field of desc.fields) {
+            if (isGroupLike(field)) {
+                setOnce(byText, field.message.name, field);
+                setOnce(byText, field.message.name.toLowerCase(), field);
+            }
+            else {
+                setOnce(byText, field.name, field);
+            }
+        }
+        textFieldCache.set(desc, byText);
+    }
+    return byText.get(name);
+}
+function setOnce(map, key, field) {
+    if (!map.has(key)) {
+        map.set(key, field);
+    }
+}
+function describe(tok) {
+    switch (tok.type) {
+        case "identifier":
+            return `"${tok.value}"`;
+        case "int":
+        case "float":
+            return `"${tok.text}"`;
+        case "string":
+            return "a string";
+        case "eof":
+            return "end of input";
+        default:
+            return `"${tok.type}"`;
+    }
+}

package/dist/esm/txtpb/index.d.ts

@@ -0,0 +1,4 @@
+export { toText } from "./to-text.js";
+export type { TextWriteOptions } from "./to-text.js";
+export { fromText, mergeFromText } from "./from-text.js";
+export type { TextReadOptions } from "./from-text.js";

package/dist/esm/txtpb/index.js

@@ -0,0 +1,15 @@
+// Copyright 2021-2026 Buf Technologies, Inc.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+export { toText } from "./to-text.js";
+export { fromText, mergeFromText } from "./from-text.js";