ankane / archer Goto Github PK

If I put archer-rails in my Gemfile and then try to use heroku run rails console, I get the following exception:

Traceback (most recent call last):
        8: from /app/vendor/bundle/ruby/2.7.0/gems/archer-rails-0.2.1/lib/archer/engine.rb:11:in `block (2 levels) in <class:Engine>'
        7: from /app/vendor/bundle/ruby/2.7.0/gems/archer-rails-0.2.1/lib/archer.rb:47:in `save'
        6: from /app/vendor/bundle/ruby/2.7.0/gems/archer-rails-0.2.1/lib/archer.rb:70:in `quietly'
        5: from /app/vendor/bundle/ruby/2.7.0/gems/activesupport-6.1.0/lib/active_support/logger_silence.rb:18:in `silence'
        4: from /app/vendor/bundle/ruby/2.7.0/gems/activesupport-6.1.0/lib/active_support/logger_thread_safe_level.rb:52:in `log_at'
        3: from /app/vendor/bundle/ruby/2.7.0/gems/activesupport-6.1.0/lib/active_support/logger_silence.rb:18:in `block in silence'
        2: from /app/vendor/bundle/ruby/2.7.0/gems/archer-rails-0.2.1/lib/archer.rb:71:in `block in quietly'
        1: from /app/vendor/bundle/ruby/2.7.0/gems/archer-rails-0.2.1/lib/archer.rb:49:in `block in save'
/app/vendor/bundle/ruby/2.7.0/gems/archer-rails-0.2.1/lib/archer.rb:58:in `history_object': uninitialized constant Archer::Readline (NameError)
Did you mean?  Reline

Strangely, this didn't happen the first time I brought up a Heroku Rails console after installing archer-rails following the instructions, but only the second time and all times after (I wanted to check that the command history was working).

I've also tried having archer-rails included for all groups instead of just production, as well as using heroku run bundle exec rails console just in case, but I get the same error. I've added/run the migrations as per the instructions.

This is with Rails 6.1.0 under Ruby 2.7.2 with Bundler 2.2.4.

Hi there!

I had been looking at doing something similar to this project in order to aggregate IRB history into our observability platform for any executions that might happen within Heroku Dynos via Rails console, so we can have historical record of the activity and a general audit trail incase of incidents. Would you be open to accepting a pull request(s) that would extend this project in the following ways:

1. Adds a hook so that the history log entry could be sent or processed elsewhere (in our use case this would be New Relic)

2. Adds the ability to define a user callback method that gets executed to determine the default user to send. We use Heroku and we're looking at using the Heroku CLI to determine the authenticated user as a default.

I'm happy to submit a PR if you think those two pieces would be suitable for the project.

In some cases you might not want a session to be saved in the DB as clear text. Is there a way to maybe run a command before exiting that would let archer know not to save this session?

Multi-line commands seem to be broken up into their individual lines by Archer, which makes it hard to re-run them. Contrived example:

[1,2,3].each do |item|
  puts item
end

Pressing up within the same console session restores the entire 3-line command, but upon starting a new console and pressing up, I can only select a single line, ex:

irb(main):001* [1,2,3].each do |item|
irb(main):002*

However, pressing the up arrow on the second line simply moves up to the first line, instead of letting me choose the puts item line from the history to even attempt to piece the multi-line command back together.

Hi,
Thanks for this awesome gem.

Is there a way I can set a new user for every session so that it doesn't keep updating the same record? It could be something like current date time.

How can I achieve that?