Jonathan Hefner

That could work too. Though branching on `arity == 1` effectively creates two "classes" of callables — one where the return value matters and one where it doesn't. Also, branching...

> Maybe inverting it makes it simpler? > > ```ruby > Rails.application.config.filter_parameters += [ > lambda { |key, value| throw :allow, key.match?(/(\Aid|_id)\z/) } > ] > ``` Syntactically, I like...

Thank you for attacking this issue! :pray: :tada: Unfortunately, there is a counter-case for this approach which appears in #34408. In #34408, the last page of results changes from `[7,...

@alipman88 Interesting... > on large collections, the first and last ID approach takes about 50% longer than the current approach, and the MD5 checksum approach takes about 290% longer than...

@kaspth #34408 is a good summary and illustration of the issue. I think this PR is the latest and greatest approach to solving it. A less complicated alternative approach is...

@aried3r That is an interesting example! Would it be possible, in your case, to summarize a user's permissions into a single value, and make that part of the key passed...

(Preface: I use Sprockets in all of my Rails *apps*, and I greatly appreciate all the work that has gone into it.) I recently encountered two use cases for disabling...

@fatkodima is correct. `normalizes` ["declares a normalization for one or more attributes"](https://api.rubyonrails.org/v7.1/classes/ActiveRecord/Normalization/ClassMethods.html#method-i-normalizes), and `store` accessors are not implemented as attributes. However, as he mentions, the store itself _is_ an attribute,...

Thank you, @connor4312! :smiley:

This seems useful! I have a few questions: * In your browser automation examples, what would be the `content` accompanying the `message`? * You specify that "This message should not...