Prefix the method name with the class name for validation errors #17754
Conversation
|
|
|
LGTM, even the test failures show who this affects and they all seem like legitimate things to change in the tests. |
|
cc @mzazrivec |
bdunne
force-pushed
the
better_validation_errors
branch
from
e385a95 to
c70bf90
Compare
| @@ -23,4 +23,8 @@ class ApplicationRecord < ActiveRecord::Base | |||
| def self.display_name(number = 1) | |||
| n_(model_name.singular.titleize, model_name.plural.titleize, number) | |||
| end | |||
|
|
|||
| def self.human_attribute_name(attribute, *args) | |||
| "#{name}: #{super}" | |||
There was a problem hiding this comment.
I wonder if this should be display_name ?
There was a problem hiding this comment.
I don't know. @jrafanie and I talked about it briefly and we're leaning toward no, but ¯\_(ツ)_/¯
There was a problem hiding this comment.
I don't know, that seems to be masking the real model name failing validation.
There was a problem hiding this comment.
display_name would get you a nice looking model name, but at the same time the nice looking model name would be translated into current locale (shrug).
There was a problem hiding this comment.
Right, 👎 to using display_name, I don't think this should be translated.
| @@ -0,0 +1,13 @@ | |||
| describe ApplicationRecord do | |||
| context ".human_attribute_name puts the class name in the validation error" do | |||
There was a problem hiding this comment.
Might as well reword this to not upset the puts checker 😆
instance.errors.full_messages will now produce something like: ["SomeClass: The method can't be blank"]
bdunne
force-pushed
the
better_validation_errors
branch
from
c70bf90 to
8e53499
Compare
|
Checked commit bdunne@8e53499 with ruby 2.3.3, rubocop 0.52.1, haml-lint 0.20.0, and yamllint 1.10.0 |
There's only one way to find out. We get burned by indirect validation failures often enough that we should just fix anything that this change affects. |
ManageIQ/manageiq#17754 added a prefixed model name in a number of exceptions which is causing a few specs to fail which were matching on the raised exception.
instance.errors.full_messages will now produce something like:
["SomeClass: The attribute can't be blank"]
Full disclosure... I have no idea what else this will effect (especially in the UI) ¯\_(ツ)_/¯