Skip to main content

Scenario Simple Values

We went through the structure of a scenario, and you might have noticed that in a lot of places values are expressed in diverse ways.

The VM imposes very few restrictions on its inputs and outputs, most fields are processed as raw bytes. The most straightforward way to write a test that one could think of would be to have the actual raw bytes always expressed in a simple format (e.g. like hexadecimal encoding). Indeed, our first contract tests were like this, but we soon discovered that it took painfully long prepare them and even longer to refactor. So, we gradually came up with increasingly complex formats to represent values in an intuitive human-readable way.

We chose to create a single universal format to be used everywhere in a scenario file. The same format is used for expressing:

  • addresses,
  • balances,
  • transaction and block nonces,
  • contract code,
  • storage keys and values,
  • log identifiers, topics and data,
  • gas limits, gas costs,
  • ESDT metadata, etc.

The advantage of this unique value format is that it is enough to understand it once to then use it everywhere.

The scenario value format is closely related to the MultiversX serialization format. This is not by accident, Scenarios are designed to make it easy to interact MultiversX contracts and their data.

Exceptions: txId, comment and asyncCallData are simple strings. asyncCallData might be changed to the default value format in the future and/or reworked.

important

It must be emphasized that no matter how values are expressed in scenarios, the communication with the VM is always done via raw bytes. Of course it is best when the value expression and the types in the smart contract match, but this is not enforced.

A note on error messages: whenever we write a test that fails, the test runner tries its best to transform the actual value it found from raw bytes to a more human-readable form. It doesn't really know what format to use, so it tries its best to find something plausible. However, all it has are some heuristics, so it doesn't always get it right. It also displays the raw bytes so that the developer can investigate the proper value.

A note about the value parser and the use of prefixes​

The value interpreter is not very complex and uses simple prefixes for most functions. Examples of prefixes are "str:" and "u32:".

The | (pipe) operator, which we use for concatenation has the highest priority. More about it here.

The arguments of functions start after the prefix (no whitespace) and end either at the first pipe (|) or at the end of the string.

Multiple prefixes evaluated right to left, for instance "keccak256:keccak256:str:abcd" will first convert "abcd" to bytes, then apply the hashing function on it twice.

With that being said, the following sections will describe how to express different value types in scenarios. A full list of the prefixes is at the end of this page.

Empty value​

Empty strings ("") mean empty byte arrays. The number zero can also be represented as an empty byte array. Other values that translate to an empty byte array are "0" and "0x".

Hexadecimal representation​

To provide raw hexadecimal representations of the values, use the prefix 0x and follow it with the base-16 bytes. E.g. "0x1234567890". After the 0x prefix an even number of digits is expected, since 2 digits = 1 byte.

Examples
  • "0x"
  • "0x1234567890abcdef"
  • "0x0000000000000000"

Standalone number representations​

Unprefixed numbers are interpreted as base 10, unsigned. Unsigned numbers will be represented in the minimum amount of bytes in which they can fit.

Examples
  • "0"
  • "1"
  • "1000000".
  • "255" is the same as "0xff"
  • "256" is the same as "0x0100"
  • "0" is the same as ""
tip

Digit separators are allowed anywhere, for readability, e.g. "1,000,000".

Standalone signed numbers​

caution

Only use signed numbers if you absolutely need to. Big signed integer representation has some pitfalls that can lead to subtle and unexpected issues when interacting with the contract.

Sometimes contract arguments are expected to be signed. These arguments will be transmitted as two’s complement representation. Prefixing any number (base 10 or hex) with a minus sign will convert them to two’s complement. Two’s complement is interpreted as positive or negative based on the first bit.

Sometimes positive numbers can start with a "1" bit and get accidentally interpreted as negative. To prevent this, we can prefix them with a plus. A few examples should make this clearer:

Examples
  • "1" is represented as "0x01", signed interpretation: 1, everything OK.
  • "255" is represented as "0xff", signed interpretation: "-1", this might not be what we expected.
  • "+255" is represented as "0x00ff", signed interpretation: "255". The prepended zero byte makes sure the contract interprets it as positive. The + makes sure those leading zeroes are added if necessary.
  • "+1" is still represented as "0x01", here the leading 0 is not necessary. Still, it is good practice adding the + if we know the argument is expected to be signed.
  • "-1" is represented as "0xff". Negative numbers are also represented in the minimum number of bytes possible.

For more about signed number encoding, see the big number serialization format.

Nested numbers​

Whenever we nest numbers in larger structures, we need to somehow encode their length. Otherwise, it would become impossible for them to be deserialized.

The format helps developers to also easily represent nested numbers. These are as follows:

  • biguint: is useful for representing a nested BigUint. It outputs the length of the byte representation, followed by the big endian byte representation itself.
  • u64: u32: u16: u8: interpret the argument as an unsigned int and convert to big endian bytes of respective length (8/4/2/1 bytes)
  • i64: i32: i16: i8: interpret the argument as a signed int and convert to 2's complement big endian bytes of respective length (8/4/2/1 bytes)
Examples
  • "biguint:0" equals 0x00000000
  • "biguint:1" equals 0x0000000101
  • "biguint:256" equals 0x00000020100
  • u64:1 equals 0x0000000000000001
  • i64:-1 equals 0xFFFFFFFFFFFFFFFF
  • u32:1 equals 0x00000001
  • u16:1 equals 0x0001
  • u8:1 equals 0x01

Nested items​

The nested: prefix prepends the length of the argument. It is similar to biguint:, but does not expect a number.

Examples
  • "nested:str:abc" equals 0x00000003|str:abc
  • "nested:0x01020304" equals 0x0000000401020304

Booleans​

The format offers these 2 constants, for convenience:

  • "true" = "1" = "0x01"
  • "false" = "0" = "".
caution

This is the standalone representation. If your boolean is embedded in a structure or list, use u8:0 instead of false.

ASCII strings​

The preferred way of representing ASCII strings is with the str: prefix.

important

The '' and `` prefixes are common in older examples. They are equivalent to str:, but considered legacy. We recommend avoiding them because they clash with the syntax of languages where we might want to embed scenario code (Go and Markdown in particular).

User Addresses​

address: constructs a dummy user address from a word.

Addresses need to be 32 bytes long, so

  • if the word is longer, it gets chopped off at the end
  • if the word is shorter, it gets extended to the right with 0x5f bytes (the "_" character).
Example

"address:my_address" is the same as:

  • "str:my_address______________________" or
  • "0x6d795f616464726573735f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f".

Smart Contract Addresses​

sc: constructs a dummy smart contract address.

On MultiversX, smart contract addresses have a different format than user address - they start with 8 bytes of zero.

important

The format requires that all accounts with addresses in SC format must have non-empty code.

It is forbidden to have accounts with code but an address that doesn't obey the SC format.

Example

"sc:my_address" is the same as:

  • "0x0000000000000000|str:my_address______________" or
  • "0x00000000000000006d795f616464726573735f5f5f5f5f5f5f5f5f5f5f5f5f5f".

Sometimes the last byte of a a SC address is relevant, since it affects which shard the contract will end up in. It can be specified with a hash characher #, followed by the final byte as hex.

Example

"sc:my_address#a3" is the same as:

  • "0x0000000000000000|str:my_address_____________|0xa3" and
  • "0x00000000000000006d795f616464726573735f5f5f5f5f5f5f5f5f5f5f5f5fa3".

File contents​

file: loads an entire file and uses the contents of the entire file as value.

The path of the file is given relative to the current scenario file.

Used in the first place for specifying smart contract code. It can, however, be used for specifying any value, anywhere.

Example

"file:../output/my-contract.wasm"

Example usage:

  • initializing contract code,
  • contract code to deploy,
  • contract code to be passed to another contract as argument for indirect deploy,
  • checking some contract code in storage,
  • any large argument

Hash function​

keccak256: computes the Keccak256 hash of the argument. The result is always 32 bytes in length.

The full list of scenario value prefixes​

The prefixes are:

  • str: converts from ASCII strings to bytes.
  • address: dummy user address
  • sc: dummy smart contract address
  • file: loads the entire contents of a file
  • keccak256: computes the hash of the argument
  • u64: u32: u16: u8: interpret the argument as an unsigned int and convert to big endian bytes of respective length (8/4/2/1 bytes)
  • i64: i32: i16: i8: interpret the argument as a signed int and convert to 2's complement big endian bytes of respective length (8/4/2/1 bytes)
  • biguint: big number unsigned byte length followed by big number unsigned bytes themselves
  • nested: prepends the length of the argument