It appears that the code AFTER the yield in my sidekiq client middleware is not executing. No exception was raised (I'm trapping that), so I can't imagine how the code is being skipped (no return inside the lambda, either). Can anyone explain why? Here's the middleware code:
class SidekiqClientWorkless
def initialize(options = nil)
end
def call(worker, msg, queue)
Log.create!(task: "testing", message: "This message gets logged")
begin
yield
rescue Exception => e
require 'ruby-debug'
debugger
end
Log.create!(task: "testing", message: "This message does not get logged")
end
end
Sidekiq.configure_client do |config|
config.redis = { :url => 'redis://user:pwd#barb.redistogo.com:9725/',
:namespace => 'mynamespace' }
config.client_middleware do |chain|
chain.add SidekiqClientWorkless, :foo => 1, :bar => 2
end
end
Have you tried using ensure as per the example from the docs?
I'm honestly not sure why it would be required, but it seems to be...
Related
With the migration from Rails 2 to Rails 3 validation errors were moved from ActiveRecord::Error to ActiveModel::Errors.
In rails 2 the validation error had a type and a message (among other things) and you could check the type of the validation error by doing something like the following:
rescue ActiveRecord::RecordInvalid => e
e.record.errors.each do |attr, error|
if error.type == :foo
do_something
end
end
end
But with Rails 3 it seems everything but the invalid attribute and message has been lost. As a result the only way to determine the type is to compare the error message:
rescue ActiveRecord::RecordInvalid => e
e.record.errors.each do |attr, error|
if error == "foobar"
do_something
end
end
end
Which is not at all ideal (eg. what if you have several validations which use the same message?).
Question:
Is there a better way in rails 3.0 to determine the type of validation error?
Check for added? on ActiveModel::Errors:
https://github.com/rails/rails/blob/master/activemodel/lib/active_model/errors.rb#L331
That allows you to do this:
record.errors.added?(:field, :error)
I needed it not only for test purposes, but also for API. I've ended up with monkey patch:
module CoreExt
module ActiveModel
module Errors
# When validation on model fails, ActiveModel sets only human readable
# messages. This does not allow programmatically identify which
# validation rule exactly was violated.
#
# This module patches {ActiveModel::Errors} to have +details+ property,
# that keeps name of violated validators.
#
# #example
# customer.valid? # => false
# customer.errors.messages # => { email: ["must be present"] }
# customer.errors.details # => { email: { blank: ["must be present"] } }
module Details
extend ActiveSupport::Concern
included do
if instance_methods.include?(:details)
fail("Can't monkey patch. ActiveModel::Errors already has method #details")
end
def details
#__details ||= Hash.new do |attr_hash, attr_key|
attr_hash[attr_key] = Hash.new { |h, k| h[k] = [] }
end
end
def add_with_details(attribute, message = nil, options = {})
error_type = message.is_a?(Symbol) ? message : :invalid
normalized_message = normalize_message(attribute, message, options)
details[attribute][error_type] << normalized_message
add_without_details(attribute, message, options)
end
alias_method_chain :add, :details
def clear_with_details
details.clear
clear_without_details
end
alias_method_chain :clear, :details
end
end
end
end
end
# Apply monkey patches
::ActiveModel::Errors.send(:include, ::CoreExt::ActiveModel::Errors::Details)
I'm creating an API in Rails and I've run into a situation where I'd like to alert the user that something bad happened by passing them some JSON with an error message. However, I'd also like to re-raise the exception so that Airbrake (formerly Hoptoad) will still catch the error and notify us so that we can look into the problem a bit more.
I'm currently catching the error like so:
begin
if #myobject.update_attributes(:foo => "bar")
render :json => #myobject, :status => :ok
else
render :json => #myobject.errors, :status => :unprocessable_entity
end
rescue Exception => e
render :json => { :errors => { :message => "Uh oh! Something went wrong." } }
raise e
end
The problem is that my client never gets the JSON message since the raise e stops it from rendering and sends it a generic 500 error.
How should I fix this?
[My Solution]
As suggested by Jordan below, I simply call notify_airbrake(ex) in my code any time that I catch the exception. However, I abstracted it slightly by adding the following to my ApplicationController so that I can easily change from Airbrake to something else in the future:
class ApplicationController < ActionController::Base
...
def notify_exception_service(ex)
notify_airbrake(ex)
end
...
end
So, instead of notify_airbrake(ex) I just call notify_exception_service(ex).
From the Airbrake gem documentation:
If you want to log arbitrary things which you've rescued yourself from a controller, you can do something like this:
rescue => ex
notify_airbrake(ex)
flash[:failure] = 'Encryptions could not be rerouted, try again.'
end
The #notify_airbrake call will send the notice over to Airbrake for later analysis. While in your controllers you use the notify_airbrake method, anywhere else in your code, use Airbrake.notify.
You don't have to re-raise the exception in order to log it with in Airbrake.
As I've said on chat, I think you can't becouse of how rails render things.
When you call render, a #_something_render_variable is set, but the page is not directly render, there is still additional call. Raising an exception obviusly block this flow, actually breaking the render of the webpage.
Changing this behaviour is really hard, you must alias the render method and work on it, I had a similar problem.
When I run this in practice it works, but I can't seem to write a working test for my route constraint with rspec.
When the test runs the constraint is triggered, but the request params are empty, thus it does not validate and the test fails.
I am running Rails 3.0.9, rspec-rails 2.6.1 and rspec 2.6.0.
config/routes.rb
match ":param1-unique-:param2" => "controller#index",
:constraints => ParamConstraint.new
lib/param_constraint.rb
class ParamConstraint
def matches?(request)
#request ||= request
valid_param1? && valid_param2?
end
def valid_param1?
#request.params[:param1] == "lorem"
end
def valid_param2?
#request.params[:param2] == "ipsum"
end
end
spec/routing/param_constraint_spec.rb
require 'spec_helper'
describe "param constraint routing" do
it "recognizes route for param1 and param2" do
{ :get => "/lorem-unique-ipsum" }.
should route_to(
:controller => "controller",
:action => "index",
:param1 => "lorem",
:param2 => "ipsum"
)
end
end
Update
If I inspect the request in the constraint I get the following output:
#<ActionDispatch::Request:0x007fee140ff910 #env={
"rack.version"=>[1, 1],
"rack.input"=>#<StringIO:0x007fee1446da48>,
"rack.errors"=>#<StringIO:0x007fee1446e768>,
"rack.multithread"=>true,
"rack.multiprocess"=>true,
"rack.run_once"=>false,
"REQUEST_METHOD"=>"GET",
"SERVER_NAME"=>"example.org",
"SERVER_PORT"=>"80",
"QUERY_STRING"=>"",
"PATH_INFO"=>"/lorem-unique-ipsum",
"rack.url_scheme"=>"http",
"HTTPS"=>"off",
"SCRIPT_NAME"=>"",
"CONTENT_LENGTH"=>"0"
}>
I ran into this same issue today, searching for an answer brought me to this page's question. For what it's worth, I had to resort to writing a request spec instead.
context "passing params that satisfy ParamConstraint" do
before do
visit "/lorem-unique-ipsum"
end
it "should serve up a page with content" do
# replace this with some assertion that gets satisfied by
# pages served up when ParamConstraint.new.matches? returns true
page.should have_selector("html body div#foo")
page.should_not have_selector("html body div#bar")
end
end
context "passing params that DO NOT satisfy ParamConstraint" do
before do
visit "/other-unique-other"
end
it "should serve up a page with different content" do
# replace this with some assertion that gets satisfied by
# pages served up when ParamConstraint.new.matches? returns false
page.should_not have_selector("html body div#foo")
page.should have_selector("html body div#bar")
end
end
This doesn't answer your question, which I take to be "how to test routing constraint", as the proper way would be via a routing spec. But given this gap in how request.params works when you use "should route_to", this is a workaround. A request spec, as opposed to a routing spec, will fill request.params correctly.
Same issue exists years later, with rspec-core 3.4.4, rspec-rails 3.4.2, rails 4.2.6. Don't have time to dig into exactly why...
You can use a request spec as suggested above, but don't use it to test the page contents. Instead, replicate a routing test (route_to) by checking the conversion of URL paths to request params:
RSpec.describe 'routes', type: :request do
describe '/:slug' do
it 'routes correctly' do
get '/test-product-slug'
expect(request.params).to eq(
'controller' => 'product',
'action' => :index,
'slug' => 'test-product-slug'
)
end
end
end
The rails 3 appis is using the create! method inside the create of a controller.
Sometimes it works, and sometimes it does not.It fails consistently with always the same use case, however i have checked and rechecked and cannot understand why it fails.
The create! method fails silently, there is no indication on the logs of the problem. How can I make the create! methode more verbose?
Code :
class NotificationPaiementsController < ApplicationController
protect_from_forgery :except =>[:create]
skip_before_filter :authorize, :only => [:create]
def create
logger.debug "params is #{params}"
logger.debug "invoice is #{params[:invoice]}"
logger.debug "payment_status is #{params[:payment_status]}"
logger.debug "txn_id is #{params[:txn_id]}"
#notification_paiement = NotificationPaiement.create!(:params => params,
:cart_id => params[:invoice],
:status=> params[:payment_status],
:transaction_id => params[:txn_id])
logger.debug "notification_paiement is #{#notification_paiement}"
render :nothing=>true
end
end
EDIT:
Thx for your answers, it would have been faster to catch exception, but i managed to identify the problem using new and savevia the console. At the save i had an error about UTF-8 encoding : ArgumentError: invalid byte sequence in UTF-8.
Paypal was changing "molière" in "moli\xE8re" and the error was never displayed.
The create! constructor raises an exception if it fails:
Creates an object just like Base.create but calls save! instead of save so an exception is raised if the record is invalid.
So, if you're going to use create!, you should wrap it in exception handling:
begin
#notification_paiement = NotificationPaiement.create!(...
rescue ActiveRecord::RecordInvalid => e
# Deal with your errors.
end
You can temporarily remove the backtrace silencers in config/initializers/backtrace_silencers.rb in case exception info is being swallowed.
I have a pretty standard authenticate method
private
def authenticate_user
#current_user = User.find_by_authentication_token(params[:token])
unless #current_user
error = { :error => "Invalid token." }
respond_with(error, :status => 401 )
end
end
I am calling the API to ensure the authenticate fails.
I get an error stating
ArgumentError (Nil location provided. Can't build URI.):
app/controllers/api/v1/base_controller.rb:13:in `authenticate_user'
What am I doing wrong?
By the specific flavor of your error, I am guessing that "authenticate_user" is called as part of a "create" action.
If that is the case, I believe the answer I provided here will help you as well.
Assuming, however, that this is part of creating an authenticated session, meaning there is no actual location for the newly created "resource", I would supply nil for the response location, as in:
...
respond_with(error, :status => 401, :location => nil)
...
That will make more sense once you have a look at the linked answer. If it still doesn't make sense, I'll be happy to clarify.
I changed respond_with to render and it worked:
render json: { success: false, message: "an error" }, status: 500