borsh-js (master) yarn build
yarn run v1.22.18
$ tsc -p ./tsconfig.json
node_modules/@types/prettier/index.d.ts:41:54 - error TS2315: Type 'IsTuple' is not generic.

41 type IndexProperties<T extends

{ length: number }

= IsTuple extends true

node_modules/@types/prettier/index.d.ts:53:6 - error TS2456: Type alias 'IsTuple' circularly references itself.

53 type IsTuple<T> = T extends [] ? true : T extends <span class="error">[infer First, ...infer Remain]</span> ? IsTuple<Remain> : false;
~~~~~~~

node_modules/@types/prettier/index.d.ts:53:65 - error TS2574: A rest element type must be an array type.

53 type IsTuple<T> = T extends [] ? true : T extends <span class="error">[infer First, ...infer Remain]</span> ? IsTuple<Remain> : false;

node_modules/@types/prettier/index.d.ts:53:84 - error TS2315: Type 'IsTuple' is not generic.

53 type IsTuple = T extends [] ? true : T extends [infer First, ...infer Remain] ? IsTuple : false;

node_modules/@types/prettier/index.d.ts:96:5 - error TS2589: Type instantiation is excessively deep and possibly infinite.

96 call<
~~~~

node_modules/@types/prettier/index.d.ts:131:5 - error TS2589: Type instantiation is excessively deep and possibly infinite.

131 each<
~~~~

node_modules/@types/prettier/index.d.ts:165:5 - error TS2589: Type instantiation is excessively deep and possibly infinite.

165 map<
~~~

Found 7 errors.

error Command failed with exit code 2.
info Visit [https://yarnpkg.com/en/docs/cli/run](https://yarnpkg.com/en/docs/cli/run) for documentation about this command.

This issue was automatically created by Allstar.

Security Policy Violation
Security policy not enabled.
A SECURITY.md file can give users information about what constitutes a vulnerability and how to report one securely so that information about a bug is not publicly visible. Examples of secure reporting methods include using an issue tracker with private issue support, or encrypted email with a published key.

To fix this, add a SECURITY.md file that explains how to handle vulnerabilities found in your repository. Go to https://github.com/near/borsh-js/security/policy to enable.

For more information, see https://docs.github.com/en/code-security/getting-started/adding-a-security-policy-to-your-repository.

This issue will auto resolve when the policy is in compliance.

Issue created by Allstar. See https://github.com/ossf/allstar/ for more information. For questions specific to the repository, please contact the owner or maintainer.

I see there is no supported 'f32' or 'f64' types. I am new to typescript so maybe I'm missing something here, but is there a work-around to serialize a number type that has a floating point value? Maybe send as a byte array of u8 that reflects IEEE standard?

e.g. 0.0001

Hello, I'm serializing a BTreeMap using BorshSerialize in rust. How to deserialize this uint8array serialized data to vector/array or BTreeMap in Typescript using borsh-js or any other library? Please help

Below is my rust code

#[derive( Clone, BorshDeserialize, BorshSerialize, Copy)]
pub struct BuyOrderData {
    pub owner_public_key: [u8;32],
    pub receiving_bc_public_key: [u8;32],
    pub amount_qc: u64,
}

#[derive(Default)]
pub struct BuyOrderDataBTreeMap {
    is_initialized: bool,
    bids:  BTreeMap<u128, BuyOrderData>,
}



const INITIALIZED_BYTES: usize = 1;
const BTREEMAP_LENGTH: usize = 4;
const BTREEMAP_STORAGE: usize = 8600;
const ORDER_DATA_HEAP_ACCOUNT_STATE_SPACE: usize = INITIALIZED_BYTES + BTREEMAP_LENGTH + BTREEMAP_STORAGE;


impl BuyOrderDataBTreeMap {
    pub fn set_initialized(&mut self) {
        self.is_initialized = true;
    }

    pub fn add(&mut self, new_key:u128, new_owner_public_key:  [u8;32], new_receiving_bc_public_key:  [u8;32],  new_amount_qc: u64) {
        self.bids.insert(new_key, BuyOrderData {owner_public_key: new_owner_public_key, receiving_bc_public_key: new_receiving_bc_public_key, amount_qc: new_amount_qc});
    }

    pub fn remove(&mut self, key: u128) {
        self.bids.remove(key);
    }

}

impl Sealed for BuyOrderDataBTreeMap {}

impl IsInitialized for BuyOrderDataBTreeMap {
    fn is_initialized(&self) -> bool {
        self.is_initialized
    }
}

#[allow(clippy::ptr_offset_with_cast)]
impl Pack for BuyOrderDataBTreeMap {
    const LEN: usize = ORDER_DATA_HEAP_ACCOUNT_STATE_SPACE;
    fn unpack_from_slice(src: &[u8]) -> Result<Self, ProgramError> {
        let src = array_ref![src, 0, BuyOrderDataBTreeMap::LEN];
        let (is_initialized_src, data_len_src, data_src) = array_refs![src, INITIALIZED_BYTES, BTREEMAP_LENGTH, BTREEMAP_STORAGE];
 
        let is_initialized = match is_initialized_src {
            [0] => false,
            [1] => true,
            _ => return Err(ProgramError::InvalidAccountData),
        };
        let data_len = u32::from_le_bytes(*data_len_src) as usize;
        if data_len == 0 {
            Ok(BuyOrderDataBTreeMap {
                is_initialized,
                bids: BTreeMap::<u128, BuyOrderData>::new(),
            })
        }  else {
            let data_dser =
                BTreeMap::<u128, BuyOrderData>::try_from_slice(&data_src[0..data_len]).unwrap();
            Ok(BuyOrderDataBTreeMap {
                is_initialized,
                bids: data_dser,
            })
        }
    }

    fn pack_into_slice(&self, dst: &mut [u8]) {
        let dst = array_mut_ref![dst, 0, BuyOrderDataBTreeMap::LEN];
        let (is_initialized_dst, data_len_dst, data_dst) = mut_array_refs![dst, INITIALIZED_BYTES, BTREEMAP_LENGTH, BTREEMAP_STORAGE];
        is_initialized_dst[0] = self.is_initialized as u8;
        let orders_store_heap_data = self.bids.try_to_vec().unwrap();
        let data_len = orders_store_heap_data.len();
        data_len_dst[..].copy_from_slice(&(data_len as u32).to_le_bytes());
        data_dst[0..data_len].copy_from_slice(&orders_store_heap_data);

    }
}

The check for if the TextDecoder polyfill is required currently references global, which only exists on Node.js and not in the browser.

With bundlers that don't automatically polyfill global this causes problems and requires hacks to make it work. What you would think of one of the following solutions?

• typeof TextDecoder !== "function" ? encoding.TextDecoder : TextDecoder
• typeof globalThis.TextDecoder !== "function" ? encoding.TextDecoder : globalThis.TextDecoder

globalThis has been available for a while now, but I think the first one is best since it doesn't make any assumptions about the name of the global variable. Any thoughts? I'm happy to submit a PR for this.

While the Borsh spec doesn't officially support bool, the Rust implementation does map bool to u8 as 1/0.

Can we add bool and/or boolean to the JS implementation for convenience?

Hi! I'm trying to use the deserializeUnchecked method but it seems to be failing with this error, any idea what might be going on? Does borsh work in browsers or only node environments?

!

Thank you

Right now only fixed sized byte arrays are supported. According to the spec fixed sized arrays should be supported for any type.

In the readme, the documentation mentions about enum.

{ enum: [{ className1: { struct: {...} } }, { className2: { struct: {...} } }, ... ] } - an enum. The variants of the enum are described by the className1, className2, etc. fields. The variants are structs.

But the current implementation seems not correct.

import 

 {Buffer} 

 from "buffer";
 import 

 {deserialize, serialize} 

 from "borsh";

class Test {
 constructor(data) 

 { Object.assign(this, data); } 

}

const value = new Test(

 { x: 255, y: 20, z: '123', q: [1, 2, 3] } 

);
 const schema = new Map([[Test, 

 { kind: 'struct', fields: [['x', 'u8'], ['y', 'u64'], ['z', 'string'], ['q', [3]]] } 

]]);
 const buffer = serialize(schema, value);

Does not work

I am using borsh-js and getting the error Error: Class Array is missing in schema. I would like to deserialize an array of ChatMessage that comes into my js client code as Buffer of u8 array. Note when I deserialize just one class instance it works.

`
class ChatMessage {
archive_id: string = DUMMY_TX_ID;
created_on: string = DUMMY_CREATED_ON; // max milliseconds in date
constructor(
fields:

{ archive_id: string; created_on: string }

| undefined = undefined
) {
if (fields)

{ this.archive_id = fields.archive_id; this.created_on = fields.created_on; }

}
}
const ChatMessageSchema = new Map([
[
ChatMessage,

{ kind: "struct", fields: [ ["archive_id", "String"], ["created_on", "String"], ], }

,
],
]);
const result = serialize(
ChatMessageSchema,
sampleChatMessages
);
`

How I can deserialize in borsh UnorderedMap?

Now my schema is

 const schema = new Map([
 [
 Record,
 {
 kind: "struct",
 fields: [
 ["x2", 

 { kind: 'map', key: 'string', value: 'u128'} 

],
 ],
 },
 ],
 ]);

Rust struct:

 pub struct Counter 

 { x2: UnorderedMap<AccountId, Balance>, } 

But I have error: "Expected buffer length 2036811841 isn't within bounds"

When I try serialize same struct in js, it dont work too

 const value = new Test({ x2: 

 {"1":"1"} 

 });
 const schema = new Map([
 [
 Test,
 {
 kind: "struct",
 fields: [
 ["x2", 

 { kind: 'map', key: 'string', value: 'string'} 

],
 ]
 }
 ]
 ]);
 const buffer = borsh.serialize(schema, value);
 console.log(buffer.toString('base64'))

Error: The first argument must be of type string or an instance of Buffer, ArrayBuffer, or Array or an Array-like Object. Received type number (0)

I have nested classes but am flummoxed on how to write the schema for it. No matter what I try I get errors...

A simple example anyone?

I haven't found any example to to serialize an enum. Can you please give an example to to do it. For example, here is a Rust enum:

#<span class="error">[derive(BorshSerialize, BorshDeserialize, PartialEq, Debug)]</span>
enum CalcInstruction {
Init,
Plus 

{ value: i32 } 

,
Minus 

{ value: i32, comment: String } 

,
}

In https://borsh.io/, the specification mentions there are unnamed fields,
but don't mention it in the documentation.

Suppose we have type (u64, u64, u64, u8, Option<u8>, u64),
how did we (de)serialize with this type?

Hi I have a class type defined with schema mentioned below

`class CampaignDetails {
name: string
description: string
image_link: string
admin: Buffer
amount_donated: number

constructor()
constructor(
name?: string,
description?: string,
image_link?: string,
admin?: Buffer,
amount_donated?: number
)

{ ;(this.name = name!), (this.description = description!), (this.image_link = image_link!), (this.admin = admin!), (this.amount_donated = amount_donated!) }

static schema: any = new Map([
[
CampaignDetails,

{ kind: 'struct', fields: [ ['admin', [32]], ['name', 'string'], ['description', 'string'], ['image_link', 'string'], ['amount_donated', 'u64'], ], }

,
],
])
}`

but when it deserialize the data the structure of it is changing,

`CampaignDetails
admin: undefined
amount_donated: undefined
description: undefined
image_link: undefined
name:
admin: Uint8Array(32) [245, 127, 181, 215, 162, 110, 87, 142, 194, 140, 167, 145, 235, 30, 49, 124, 184, 88, 251, 208, 221, 249, 65, 163, 2, 122, 182, 245, 58, 188, 96, 142, buffer: ArrayBuffer(32), byteLength: 32, byteOffset: 0, length: 32, Symbol(Symbol.toStringTag): 'Uint8Array']
amount_donated: BN

{negative: 0, words: Array(3), length: 1, red: null}

description: "dogs are the best"
image_link: "https://hips.hearstapps.com/hmg-prod.s3.amazonaws.com/images/dog-puppy-on-garden-royalty-free-image-1586966191.jpg?crop=0.752xw:1.00xh;0.175xw,0&resize=640:*"
name: "Save dogs"
[[Prototype]]: Object
[[Prototype]]: Object`

below is the code for deserializing

` const getAllCampaigns = async () => {
let accounts = await connection.getProgramAccounts(programId)

accounts.forEach( =>

{ let campaignData = deserialize( CampaignDetails.schema, //@ts-ignore CampaignDetails, x.account.data ) console.log(campaignData) }

)
}`

What do you think about allowing classes to provide custom serialize/deserialize methods?

By adding something like this to serializeStruct

function serializeStruct(schema: Schema, obj: any, writer: any) {
  if (typeof obj.borshSerialize === 'function') {
    obj.borshSerialize(writer);
    return;
  }
.....
}
.....

function deserializeStruct(
  schema: Schema,
  classType: any,
  reader: BinaryReader
) {
  if(typeof classType.borshDeserialize) {
    classType.borshDeserialize(reader);
    return;
  }
....
}

we could then do this:

export class PublicKey {
    keyType: KeyType;
    data: Uint8Array;

    static borsheDeserialize(reader: BinaryReader) {
        reader.readU8
    }

    static borshDeserialize(reader: BinaryReader): PublicKey {
        const keyType = reader.readU8();

        let data;
        if(keyType === KeyType.ED25519) {
            data = reader.readFixedArray(32);
        } else if(keyType === KeyType.SECP256k1) {
            data = reader.readFixedArray(64);
        }

        return new PublicKey({ keyType, data });
    }

    borshSerialize(writer: BinaryWriter) {
        writer.writeU8(this.keyType);
        writer.writeFixedArray(this.data);
    }
}

If this looks good I'm happy to implement it and submit a PR.

Example: 'i64' instead of 'u64'

I would like some sort of confirmation that borsh-js (https://github.com/near/borsh-js) can only be used in node? I would like to use it directly in HTML Javascript using script tag. So far, I've not seen anyone using it that way. I tried to make a copy of the index.js (within the git above) and reference directly from script within my HTML. It throws error.

As title

No test is running

When schema is u64 or u128, the input value type can be number, bigint, string, boolean. But when schema is u8, u16, u32, the input value type can only be number, they are not consistent, that is strange.

I think we should make input value type be strict when serialize u64 or u128, I mean ONLY allow bigint.
Alternatively, we should make input type NOT be strict when serialize u8, u16, u32, I mean allow number, bigint, string and boolean

It is hard to define a schema for complex object, should we impl a util like BorshSchema.ts in borsh-js?

This issue is a part of near/near-api-js#252.
Please, check the details there.

Is it possible to have a feature that auto-generates a schema from a plain-old-javascript-object? Or does Borsch not support that?

Need more documentation for Test object. Otherwise unclear what it is and how it should be defined.

const value = new Test(

{ x: 255, y: 20, z: '123', q: [1, 2, 3] } 

);

What is Test? What is its definition?

README says borsh option maps to null or type but code and tests says that it maps to undefined or type instead

Is there a way to serialize a primitive without wrapping it in a struct? Does Borsh support that?

The deserialize method is not implemented as documented with a Class type parameter:

deserialize(schema: Schema, buffer: Uint8Array, class?: Class): any

Instead, it is:

deserialize(schema: Schema, buffer: Uint8Array, validate: boolean): DecodeTypes

Since there is no link to the type implementing Schema, a type assertion is required to assign deserialized values to custom types:

let data = ...;

type MyType = {
    value: string;
};
const MyTypeSchema: Schema = {
  struct: {
    value: "string"
  }
};

// No explicit type
const myTypeInstance_1 = borsh.deserialize(MyTypeSchema, data);
myTypeInstance_1.value; // Unresolved variable value 

// Use DecodeTypes
const myTypeInstance_2: DecodeTypes = borsh.deserialize(MyTypeSchema, data);
myTypeInstance_2.value; // Unresolved variable value

// Use MyType without type assertion
const myTypeInstance_3: MyType = borsh.deserialize(MyTypeSchema, data); // Initializer type DecodeTypes is not assignable to variable type MyType 

// Use type assertion
const myTypeInstance_4: MyType = <MyType>borsh.deserialize(MyTypeSchema, data);
myTypeInstance_4.value; // OK

Is this the way the library is supposed to be used? In my opinion the previous interface was cleaner. In any case, the README file is not in line with the recent changes.

Depending on the documentation, the enum schema should be

{ enum: [ { struct: { className1: structSchema1 } }, { struct: { className2: structSchema2 } }, ... ] }

Suppose we have a schema like

const borsh_schema = {
  enum: [
    { 
      struct: { 
        transfer: { 
          struct: {
            recipient: 'string',
            amount: "u64",
          },
        }
      }
    },
    { 
      struct: { 
        deploy: {
          struct: {
            contract: { array: { type: "u8" } },
            cbi_version: "u32",
          },
        }
      } 
    },
  ],
}

The type of schema will become (this is intended behavior)

const borsh_schema: {
    enum: ({
        struct: {
            transfer: {
                struct: {
                    recipient: string;
                    amount: string;
                };
            };
            deploy?: undefined;
        };
    } | {
        struct: {
            deploy: {
                struct: {
                    contract: {
                        array: {
                            type: string;
                        };
                    };
                    cbi_version: string;
                };
            };
            transfer?: undefined;
        };
    })[];
}

But this will fail the type check when we want to pass schema to (de)serialize

deserialize(borsh_schema, bytes)
// this will failed the type check

Typescript Playground: https://tsplay.dev/WPM05w

My editor didn't pick up an annotation on deserialize so I wondered whether deserialize could be made generic over classType so that it actually returns the thing it is trying to deserializing the buffer into but apparently [it is already generic over classType! ](

I'm not sure if i'm not seeign some obvious blocker to this, but from a cursory look at it – deserialize still always returns any because the deserializeStruct(...which deserialize calls?) [doesn't have a return type](

I'm doing a tutorial on figment and at least that's how their borsh/lib type declarations read, maybe their version is lagging behind?

/// <reference types="node" />
import BN from 'bn.js';
export declare function baseEncode(value: Uint8Array | string): string;
export declare function baseDecode(value: string): Buffer;
export declare type Schema = Map<Function, any>;
export declare class BorshError extends Error 

{ originalMessage: string; fieldPath: string[]; constructor(message: string); addToFieldPath(fieldName: string): void; } 

export declare class BinaryWriter 

{ buf: Buffer; length: number; constructor(); maybeResize(): void; writeU8(value: number): void; writeU16(value: number): void; writeU32(value: number): void; writeU64(value: number | BN): void; writeU128(value: number | BN): void; writeU256(value: number | BN): void; writeU512(value: number | BN): void; private writeBuffer; writeString(str: string): void; writeFixedArray(array: Uint8Array): void; writeArray(array: any[], fn: any): void; toArray(): Uint8Array; } 

export declare class BinaryReader 

{ buf: Buffer; offset: number; constructor(buf: Buffer); readU8(): number; readU16(): number; readU32(): number; readU64(): BN; readU128(): BN; readU256(): BN; readU512(): BN; private readBuffer; readString(): string; readFixedArray(len: number): Uint8Array; readArray(fn: any): any[]; } 

export declare function serialize(schema: Schema, obj: any): Uint8Array;
export declare function deserialize(schema: Schema, classType: any, buffer: Buffer): any;
export declare function deserializeUnchecked(schema: Schema, classType: any, buffer: Buffer): any;

1. 1. Bakground

Should it rather not be up to the user to actually polyfill this dependency when needed?

I tried to setup near-api-js with [snowpack](https://www.snowpack.dev/) , and it fails since the text-encoding-utf-8 library does not use default export for the mjs build but rather a normal [export](https://github.com/arv/text-encoding-utf-8/blob/17484662e1e6881accbd69d5e51f3587c901311e/lib/encoding.lib.mjs#L665), which gives errors like these:


I hijacked the borsh module and removed the imports on text-encoding-utf-8, and stuff still works 👌🏽

1. 1. TextEncoder and TextDecoder Support

The text-encoding-utf-8 library has not been updated in four years (not that it is a bad sign), but the TextEncoder and TextDecoder are largely supported by most browsers now, mostly old Internet Explorer which does not support this and which has now been deprecated and no longer supported.

• https://caniuse.com/?search=textencoder
• https://developer.mozilla.org/en-US/docs/Web/API/TextEncoder
• https://developer.mozilla.org/en-US/docs/Web/API/TextDecoder


1. 1. Proposal
      Remove the import and dependency on text-encoding-utf-8 and rather specify in the README that if one needs support for this one should rather polyfill this manually instead.

noob alert:

I'm trying to serialize a Uint8Array to be sent to Rust WASI expecting a Vec to de-serialize and am not sure of how to declare the model. What I have is:

class AccumulatorSer {
constructor(value: Uint8Array) 

{ this.accumulator = value; } 

public accumulator: Uint8Array;
};

const ACCUMULATOR_SCHEMA = new Map([[AccumulatorSer, 

{ kind: 'struct', fields: [['accumulator', 'Uint8Array']] } 

]]);

// in function
let in_acc = new AccumulatorSer(Uint8Array.from(tracking_account_state.inputAccumulator.slice(8)));

serialize(ACCUMULATOR_SCHEMA, in_acc);

I get this error:

TypeError: writer<span class="error">[capitalizeFirstLetter(...)]</span> is not a function

Also, can I avoid having to wrap the Uint8Array in a class to build the schema?

When we try to serialize a string containing non-ASCII characters, instead of using the multi-byte Unicode representation only the last 8 bits of the char value are written into the buffer.

Precise code where this happens is here.

    encode_string(value: unknown): void {
        this.checkTypes && utils.expect_type(value, 'string', this.fieldPath);
        const _value = value as string;

        // 4 bytes for length
        this.encoded.store_value(_value.length, 'u32');

        // string bytes
        for (let i = 0; i < _value.length; i++) {
            this.encoded.store_value(_value.charCodeAt(i), 'u8');
        }
    }

As a result the output buffer is not compatible with the standard deserialization for UTF-8 Strings.

Solution

Actually encode the character as multi-byte.

Let's add ESM alongside CJS to this project.
Can be a requirement for near/near-api-js#607
cc @itegulov

I have been trying to deserialize solana account info with defined class type and schema but unfortunately I couldnt deserialize it.

let accounts = await connection.getProgramAccounts(programId)

accounts.forEach( =>

{ let campaignData = deserialize( CampaignDetails.schema, new CampaignDetails(), x.account.data ) }

)

error: Argument of type 'CampaignDetails' is not assignable to parameter of type 'new (args: any) => unknown'.
Type 'CampaignDetails' provides no match for the signature 'new (args: any): unknown'.ts(2345)

class CampaignDetails {
name: string
description: string
image_link: string
admin: Buffer
amount_donated: number

constructor()
constructor(
name?: string,
description?: string,
image_link?: string,
admin?: Buffer,
amount_donated?: number
)

{ ;(this.name = name!), (this.description = description!), (this.image_link = image_link!), (this.admin = admin!), (this.amount_donated = amount_donated!) }

static schema: any = new Map([
[
CampaignDetails,

{ kind: 'struct', fields: [ ['admin', [32]], ['name', 'string'], ['description', 'string'], ['image_link', 'string'], ['amount_donated', 'u64'], ], }

,
],
])
}

Any suggestionss ?

Hello,

Based on what's documented in the README:
A buffer containing an option type should deserialize to null if the value is of type None, or to the specific type if the value is Some

However, the implementation in deserializeField seems to be returning undefined instead of null.

Is this an oversight? Be happy to put in a PR for a quick change.

Thanks!

This issue was automatically created by Allstar.

Security Policy Violation
No protection found for branch master

This issue will auto resolve when the policy is in compliance.

Issue created by Allstar. See https://github.com/ossf/allstar/ for more information. For questions specific to the repository, please contact the owner or maintainer.

Would you consider a PR for u256? This would be useful to serialize 256 bit cryptographic keys.

Is it possible to have a property with 2 possible values for the length in a borsh-js(v.1.0.0) schema?

For example, I want to have a class with a publicKey property of type Uint8Array that has a length of 32 or 65 bytes.

Basically what I'm trying to do is this: https://github.com/near/near-api-js/pull/985/files#diff-dfff95256a781858ca8cc28ea405fc0e61dee26055dd6544465040df364a6ea4R183 but I don't see how this can be done with borsh-js v1.0.0.

This is how we are using borsh-js v1.0.0 https://github.com/near/near-api-js/blob/master/packages/transactions/src/schema.ts#L90-L230

Tried removing len: 32 but that didn't help.

in rust

#[derive(Debug, BorshDeserialize, BorshSerialize)]
pub struct ContentStorage {
    pub posts: Vec<Post>
}

#[derive(Debug, BorshDeserialize, BorshSerialize)]
pub struct Post {
    pub post_id: String,            // unique string
    pub image_url: Option<String>,  // ipfs link
    pub text: Option<String>,       // text for curr post
    pub owner: [u8; 32],            // post's owner
    pub likes_count: u32,           // likes of current post
    pub repost_count: u32,          // reposts
    pub comments: Vec<String>      // list of comments
}

in ts

export class ContentStorage {
    posts: Post[]
    static schema: Schema = new Map([
        [ContentStorage,
            {
                kind: 'struct',
                fields: [
                    ['posts', ['structs']],
                ]
            }],
    ]);

    constructor(posts: Post[]) {
        this.posts = posts
    }

    serialize(): Uint8Array {
        return serialize(ContentStorage.schema, this);
    }

    deserialize(data: Buffer): ContentStorage {
        return deserialize(ContentStorage.schema, ContentStorage, data);
    }
}

when I try to run it my result:
TypeError: writer[capitalizeFirstLetter(...)] is not a function

I am getting this error in an Angular14 project while trying to deserialize/decode a Uint8Array array.

ERROR Error: Uncaught (in promise): TypeError: reader[capitalizeFirstLetter(...)] is not a function
TypeError: reader[capitalizeFirstLetter(...)] is not a function
    at deserializeField (index.js:356:20)
    at deserializeStruct (index.js:410:33)
    at deserializeUnchecked (index.js:438:12)

Currently the library throws when the buffer supplied for deserialization is longer than expected:

This is nice when the input schema is of a fixed size. However, there's a use case where you have:

• a schema with option fields; and
• a fixed size storage backend.

under this scenario the trailing bytes are actually useless, and it's impossible to know the length of meaningful bytes without performing the deserialization that throws.

A possible solution would be to add an optional parameter on deserialize() which ignores the excess bytes.