Is there a way to not "inspect" the expected output when it fails? #1188
Comments
|
Sorry for the delay, it took me a while to get to the response form through that log. Would you agree to make a slight change in your spec? - expect(output).not_to include "bad yarn binstub"
+ expect(output.include?("bad yarn binstub")).not_to be trueIf it's not possible due to strict styling rules, you may give |
|
👋 @schneems if turning off truncation (by setting |
Literally the shortest amount of time i've waiting on an OSS issue asking about a random thing. Thanks a ton! The problem with: Is then I get no ouput. It would say something like "expected true to be false". I want all the output, but instead of
I would like the ability to tell rspec to give me the raw output rather than I'm already using some other pretty deeply hooked plugins for this project like https://github.com/Shopify/ci-queue which have tended to interact with other plugins in a not-always-predictable-way. If a custom formatter wasn't too hard I would probably adopt that approach.
Ahh, that's a good tip. I was setting it to Thanks for all your work here! |
|
Ah, sorry, I've misread your original problem. For a spec: RSpec.describe do
it do
long = <<~LONG
not here
neither here
LONG
expect(long).to include 'rare'
end
endI'm getting this failure locally: which includes an expectation, a formatted output and a diff. I've intentionally set a low limit for the output: RSpec.configure do |config|
config.expect_with :rspec do |expectations|
expectations.max_formatted_output_length = 20
end
endIf you change - expect(long).to include 'rare'
+ expect(long * 100).to include 'rare'the diff gets longer as well. Do you have the diff part in your output? |
|
If nothing else works you may pass the expect(long).to(include('rare'), long)
# or
expect(output).not_to(include("bad yarn binstub"), output) |
|
While reading the code of I found this snippet that may help you Richard: def assert_execute(*command)
stdout, stderr, status = Open3.capture3(*command)
command_to_show = command.map { |c| c.gsub(Dir.pwd, '.') }.shelljoin
on_failure = -> {
# Show output directly since RSpec truncates long output
puts "\n\n#{'=' * 100}"
puts "Failed to execute:\n#{command_to_show}"
puts "\nstdout:\n```\n#{stdout}```\n\nstderr:\n```\n#{stderr}```\n\n\n"
"Failed to execute: #{command_to_show}"
}
expect(status.success?).to eq(true), on_failure
if ENV.key?('VERBOSE')
puts "\n```\n$ #{command_to_show}\n#{stdout}```\n\n"
end
unless stderr.empty?
$stderr.puts "stderr:\n```\n#{stderr}```"
end
end |
|
@schneems Have managed to solve the problem? Do you think there's anything we can help you with? |
|
I guess we'll close this for now. |
I'm using rspec to test build deploys which are very verbose. I want the full output when it fails because it's very valuable for debugging, but when an expectation fails it looks like it is "inspect"-ed so it shows up with newlines like this and is largely unreadable:
Is there a config for disabling calling inspect on the input?
The text was updated successfully, but these errors were encountered: