If == matcher fails with empty Diff, error notice should be more helpful. #123
Comments
|
Using == we get this: Using eq we get this: These have the same information, just formatted differently. What is it that you would like to see? |
|
Why it fails. ;) Please read this again with a 100000 miles perspective: Expected A What should this mean? Whats the failure? If one expects that A equals A and it fails with: "Hey you expected A but got A with a difference of none" its not very helpful, isn't it? ;) I think a little bit more explaination would be helpful.
|
|
I understand the goal and the motivation. I was asking you for a recommendation on wording. How about "A == A returns true, but the objects do not report any difference." |
|
Call me stupid, but I think I really don't get it. The test fails, because "A == A returns true, but the objects do not report any difference."? What am I missing? |
|
That should have been "A == A returns false ..." |
|
Phew, and I thought I am loosing my mind. :) What about: "A == A haven't any difference, but returns false. Are you comparing new instances of ActiveRecord Models? Try to save them first." Or is this too specific? |
|
Right track but yes, too specific. How about something like: "A.==(B) returned false even though the diff between A and B is empty. Check the implementation of A.==." |
|
It's ok for me. But maybe some more people could share their thoughts on this. |
That just slows things down. We can always iterate on @dchelimsky's suggestion after it's released. P.S. I think @dchelimsky suggestion is plenty good enough. |
|
I like @dchelimsky's suggestion, too. |
|
@tvdeyen would you like to implement this? |
|
Ok, why not. But you have to guide me the right way. |
|
I'm having the same failure message in a spec and I would like to know what it exactly means. |
|
@Agis-, it's likely due to an object that overrides |
|
These are the actual specs: https://gist.github.com/3744048 |
|
Ah, that's ActiveRecord. If two ActiveRecord objects are being compared and both of their tldr: it doesn't look like |
|
FWIW, the message that was added here gets printed in some cases where it makes no sense. See #143. |
|
I see, thanks @alindeman. |
|
Same problem here. Inside rspec [hash] == [hash] returns false. In terminal it returns true. Looks like the implementation is working, but the spec test says otherwise. Terminal: race = Race.first
teams = race.teams.map(&:attributes)
RaceExportation.new(race).to_hash[:teams] == teams # trueInside rspec require 'spec_helper'
describe RaceExportation do
let(:race) { category.race }
let(:team) { create(:team, race: race) }
describe "#to_hash" do
subject { RaceExportation.new(race).to_hash }
it "sets :teams key" do
subject[:teams].should == [team.attributes]
end
end
end
|
|
Turns out it is something related to Time. As my time stubbing didnt work I did that workaround (im not very proud of it). describe RaceExportation do
def remove_times(object)
if object.class == Array
object.map{|e| remove_times(e)}
else
object.except('created_at', 'updated_at')
end
end
# let(){}..
it "sets :teams key" do
remove_times(subject[:teams]).should == remove_times([team.attributes])
end
end |
|
Sorry to reopen this issue but I am facing the empty diff message again... I even checked that the created_at and updated_at times where identical to the millisecond. |
|
Are the times equal in ruby? Regardless of clock equalness. |
|
Not sure if I understand what you asked for but this returns true : json_1 being the json created with the as_json method of stat_1. stat_1 is a mongoid entity I create at the beginning of the test. |
|
Same for updated_at and gist how you're comparing json_1 with stat_1? |
|
Here is the gist with the prints in a |
|
I'd guess id is a string in one and a Bson::ObjectId in another... |
|
No because the method is as follows : As you may see here, I give to the method |
|
If the diff is empty but they are not equal it's a symptom of something being printed to_s in the diff thats not in the object, time is the usual culprit but there are others. |
|
Well if I fail to remove the Bson id from the Stat object in the test, I get an error but with an explicit diff that I can use : So I guess the culprit is not in the BSON id this time, no ? |
|
I was thinking maybe the diff is not empty but is actually some white spaces but I have no idea what could create that and how to check that. |
|
Getting something similar at the moment... `Failure/Error: expect(ExpandedCompany.find(@company.id)).to eq(@expanded_company) ` Not sure why these objects are not == ? Edit: I'm using SimpleDelegator, if that could be the issue. |
|
@ryanwanger you have to specifically delegate 2.2.3 :001 > require 'delegate'
=> true
2.2.3 :002 > object = Object.new
=> #<Object:0x007fc8921de8b0>
2.2.3 :003 > klass = Class.new(SimpleDelegator)
=> #<Class:0x007fc89221c700>
2.2.3 :004 > klass.new(object) == klass.new(object)
=> false
2.2.3 :005 > klass.new(object).inspect
=> "#<Object:0x007fc8921de8b0>" |
|
It seems to depend on what is delegated: > file = File.open('/etc/passwd')
=> #<File:/etc/passwd>
> file == file
=> true
> SimpleDelegator.new(1) == SimpleDelegator.new(1)
=> true
> SimpleDelegator.new(file) == SimpleDelegator.new(file)
=> false This seems like a flaw in SimpleDelegator.new(file).__getobj__ == SimpleDelegator.new(file).__getobj__
=> true |
…cts is empty. Fixes rspec/rspec-expectations#123 This commit was imported from rspec/rspec-expectations@e12a2d9.
Comparing two objects via == that are not equal, but have no differences make rspec fail. The error notice shown is:
It would be helpful if an explanation of what's the problem shows up. Like the comparing with eq notice!
The text was updated successfully, but these errors were encountered: