assert: add JSONEqBytes by GCrispino · Pull Request #1513 · stretchr/testify

GCrispino · 2023-12-02T02:57:57Z

Summary

This PR adds a new function into the assert package, JSONEqBytes, that does the same as JSONEq, but by receiving []byte values as parameter instead of string ones.

Changes

assert/assertions.go:
The implementation of JSONEqBytes is pretty much what was in JSONEq, but having it receive []byte values and then not needing to convert these values to []byte when calling json.Unmarshal inside it.

Then, the implementation of JSONEq was changed such that it just calls JSONEqBytes by converting its original arguments to []byte values (which it already did when calling json.Unmarshal)

assert/assertions_test.go:
Tests for JSONEqBytes function were replicated from the tests from JSONEq.

Motivation

It is common in Go to store JSON text in []byte variables instead of string values, so having a function that allows to receive these values when doing JSON-equality checks without having to cast them to string can be useful.

Maybe not as important, it would also avoid unnecessary []byte-to-string conversions, which could add up when dealing with big JSON payloads.

Example of usage:

package main

import (
    "fmt"
    "testing"
    "github.com/stretchr/testify/assert"
    "github.com/stretchr/testify/require"
)


func TestJSONFiles(t *testing.T) {
    actualFile := "/path/to/file.json"
    byteValue, err := ioutil.ReadAll(jsonFile)
    require.NoError(t, err)

    expectedContentsFile := "/path/to/expected.json"
    byteValue, err = ioutil.ReadAll(jsonFile)
    require.NoError(t, err)

    assert.JSONEqBytes(t, expected, byteValue)
}

hendrywiranto · 2023-12-12T16:00:56Z

Hi !
this one is great but I think currently maintainers are more focused in fixing bugs rather than expanding the API interface.
Let's wait their response

assert/assertions_test.go

brackendawson

This is a good implementation.

I have questions about whether JSONEq should have ever been added. My preference would always be your API should yield stably formatted output. If JSONEq should have been added then it should have always accepted a []byte rather than a string. I'm not going to merge this without input from other maintainers, any thoughts @dolmen?

It would be useful to see if there is a real world example where the string is prohibitively expensive.

brackendawson · 2024-10-29T13:57:31Z

@GCrispino can you re-base to run the latest version of the tests.

brackendawson · 2024-10-29T13:58:30Z

I note that YAMLEq also accepts a string, that's an even more problematic assertion.

GCrispino · 2024-10-29T22:01:31Z

@brackendawson rebased my branch against master and re ran go generate and go fmt 👍🏼

brackendawson

Still approve, as per my previous comments.

GCrispino · 2024-10-30T01:08:29Z

@brackendawson thanks! Is there something else I should do, or are you only expecting review from another maintainer?

brackendawson · 2024-10-30T11:18:12Z

@brackendawson thanks! Is there something else I should do, or are you only expecting review from another maintainer?

Nothing else needed from you at this time, just wait for another maintainer to opine.

fredbi · 2025-09-27T19:39:25Z

Note to self: push merge conflict resolve

this is one nice addition in a million. The string stuff should never have existed in the first place but let’s be pragmatic and adopt one api that we test writers may use idiomatically, ie without type conversion

GCrispino · 2025-09-28T15:21:44Z

Hi @fredbi !

I have gone ahead and solved the merge conflicts myself. I believe it's all good, as tests are passing.
I've ran go generate ./... and go fmt ./... again but it appears that no changes happened as a result of this.

fredbi

Great job

Resolved merge conflicts with the go-openapi fork, rebase and ensured all tests pass. Signed-off-by: Frederic BIDON <fredbi@yahoo.com>

fredbi · 2025-10-02T18:01:41Z

assert/cpu.out

We probably don't need this cpu.out file.

Sorry, I hadn't realized I had added that file. Pushed a commit removing it 👍🏼

fredbi · 2025-10-02T19:18:09Z

@GCrispino merged in our fork go-openapi/testify

dolmen

We must have special messages for the []byte(nil) cases.

GCrispino · 2025-10-06T00:27:08Z

@dolmen

We must have special messages for the []byte(nil) cases.

Currently if expected == []bytes(nil), the "Expected value ('') is not valid json.\nJSON parsing error: 'unexpected end of JSON input'" message will be used, and in the case where actual == []bytes(nil), the "Input ('') needs to be valid json.\nJSON parsing error: 'unexpected end of JSON input'" one will be used.

I've added a commit that will use the messages of "Expected value is empty" and "Actual value is empty", respectively, in case that the provided byte slices are either nil or empty, if you believe we need special messages in these cases, apart from the current ones complaining that the input is not proper JSON.
Otherwise, I can just revert this commit.

Can you please have a look at it?

ccoVeille · 2025-10-06T11:09:37Z

@GCrispino I think you quoted me while you wanted to quote @dolmen

GCrispino · 2025-10-06T14:43:17Z

@ccoVeille oh yeah sorry, that was the case indeed 😅 , thanks for the heads up

dolmen added enhancement pkg-assert Change related to package testify/assert labels Dec 7, 2023

suqin-haha reviewed Oct 15, 2024

View reviewed changes

assert/assertions_test.go Show resolved Hide resolved

brackendawson previously approved these changes Oct 29, 2024

View reviewed changes

brackendawson requested a review from dolmen October 29, 2024 13:57

GCrispino force-pushed the jsoneq-bytes branch from c1830b9 to 528f908 Compare October 29, 2024 21:55

GCrispino dismissed brackendawson’s stale review via 29a3e3b October 29, 2024 21:59

GCrispino force-pushed the jsoneq-bytes branch from 528f908 to 29a3e3b Compare October 29, 2024 21:59

brackendawson previously approved these changes Oct 29, 2024

View reviewed changes

dolmen changed the title ~~Add function for supporting JSON equality checks using byte slices~~ assert: add JSONEqBytes May 28, 2025

ccoVeille previously approved these changes Sep 28, 2025

View reviewed changes

Gabriel Crispino added 3 commits September 28, 2025 12:14

feat: add JSONEqBytes function

e9a2742

test: JSONEqBytes

9e7dbc5

doc: update documentation comment in JSONEqBytes

6f0c3fc

GCrispino dismissed stale reviews from ccoVeille and brackendawson via df6d47d September 28, 2025 15:16

GCrispino force-pushed the jsoneq-bytes branch from 29a3e3b to df6d47d Compare September 28, 2025 15:16

chore: run go generate and go fmt

aa741a8

GCrispino force-pushed the jsoneq-bytes branch from df6d47d to aa741a8 Compare September 28, 2025 15:19

fredbi approved these changes Sep 29, 2025

View reviewed changes

fredbi added a commit to go-openapi/testify that referenced this pull request Oct 2, 2025

merged stretchr#1513

d425f21

Resolved merge conflicts with the go-openapi fork, rebase and ensured all tests pass. Signed-off-by: Frederic BIDON <fredbi@yahoo.com>

fredbi reviewed Oct 2, 2025

View reviewed changes

delete file added by mistake

df3025c

dolmen requested changes Oct 5, 2025

View reviewed changes

GCrispino added 2 commits October 5, 2025 21:08

test: fix wrong test names

ae727a2

use special error message in the case where byte slices are nil or empty

176bd69

GCrispino requested a review from dolmen October 6, 2025 00:27

Conversation

GCrispino commented Dec 2, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Motivation

Uh oh!

hendrywiranto commented Dec 12, 2023

Uh oh!

Uh oh!

brackendawson left a comment

Choose a reason for hiding this comment

Uh oh!

brackendawson commented Oct 29, 2024

Uh oh!

brackendawson commented Oct 29, 2024

Uh oh!

GCrispino commented Oct 29, 2024

Uh oh!

brackendawson left a comment

Choose a reason for hiding this comment

Uh oh!

GCrispino commented Oct 30, 2024

Uh oh!

brackendawson commented Oct 30, 2024

Uh oh!

fredbi commented Sep 27, 2025

Uh oh!

GCrispino commented Sep 28, 2025

Uh oh!

fredbi left a comment

Choose a reason for hiding this comment

Uh oh!

fredbi Oct 2, 2025

Choose a reason for hiding this comment

Uh oh!

GCrispino Oct 2, 2025

Choose a reason for hiding this comment

Uh oh!

fredbi commented Oct 2, 2025

Uh oh!

dolmen left a comment

Choose a reason for hiding this comment

Uh oh!

GCrispino commented Oct 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ccoVeille commented Oct 6, 2025

Uh oh!

GCrispino commented Oct 6, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

GCrispino commented Dec 2, 2023 •

edited

Loading

GCrispino commented Oct 6, 2025 •

edited

Loading