Reading JSON Output

--format=json writes CCIP integer fields (chainSelector, token amounts, fees, gas limits) as plain JSON numbers. These are uint64 and uint256 values, so they routinely exceed 2^53.

Python, and Go with json.Number, read such values at full precision. JavaScript does not: JSON.parse represents every JSON number as an IEEE-754 double, which holds 53 bits of integer precision, so larger values come back rounded, with no error raised.

JavaScript

JSON.parse('{"chainSelector":11344663589394136015}').chainSelector
// => 11344663589394135000   (expected 11344663589394136015)

This is deliberate: the format targets engines with native big-integer support. The requirement is on the reader.

JavaScript and TypeScript

Use jsonParse from @chainlink/ccip-sdk. It returns large integers as bigint. Pass a type parameter for the shape you expect; it returns unknown by default:

TypeScript

import { jsonParse } from '@chainlink/ccip-sdk'

type ShowOutput = { request: { lane: { sourceChainSelector: bigint } } }

const parsed = jsonParse<ShowOutput>(stdout)
parsed.request.lane.sourceChainSelector
// => 16015286601757825753n

If you're not importing the SDK, lossless-json solves the same problem.

jq

jq 1.7 and later keep large integers as literal text through passthrough and field selection. Any numeric operation converts the value to a double and rounds it. jq 1.6 and earlier round these values even on plain jq ., so check your version first:

Shell

jq --version   # jq-1.7 or later required for the passthrough behavior below

ccip-cli show <tx> --format=json | jq '.request.lane.sourceChainSelector'
# 16015286601757825753

ccip-cli show <tx> --format=json | jq '.request.lane.sourceChainSelector + 0'
# 16015286601757825000

Use | tostring when a shell variable is needed instead of a numeric operation.

Python, Go

Python's json.loads returns arbitrary-precision int and needs no special handling.

In Go, the default decode produces float64. Use json.Decoder.UseNumber(), or unmarshal into a string or *big.Int field:

Go

dec := json.NewDecoder(r)
dec.UseNumber()

The rule

Keep CCIP integer fields as text or bigint in every language. Converting them to a native floating-point number loses precision.

Testing your pipeline

Round values such as 1e18 are exactly representable as doubles and survive JSON.parse unchanged, so a "1 token" test passes while the pipeline is still broken. Test with a non-round value:

JavaScript

JSON.parse('{"amount":1234567890123456789}').amount
// => 1234567890123456800

For the jsonStringify/jsonParse pair and how they round-trip, see the SDK's JSON Integer Precision guide.