allocated memory issue #204
Comments
Because it's added to derailed in development. So when you're running in production for real (not just simulating it with derailed) you also won't have dead_end.
I've described the issue here ruby/syntax_suggest#124 (comment). The TLDR is that any gem that patches into If it's causing confusion we could perhaps introduce an env var into dead_end that would optionally disable the monekypatching behavior. We could also move it from a general derailed require to a per-task require. |
|
Also that output is from memory_profiler. One question to ask is: why isn't this a problem for bootsnap? Did get it special cased somehow? |
|
Just ran into this myself, and was admittedly confused when the single largest consumer of "allocated memory by gem" by a long way was (indirectly) from the package I just installed to examine the memory itself (derailed_benchmark). Thankfully this GitHub issue came up relatively quick from searching, but it's not a particularly great first-time experience to immediately get led on a (short) wild goose chase inflicted by this very gem. If there is a way to suppress derailed's own dependencies from the output (when they are not otherwise used by the host application of course), I think I'm in favor of that. |
Currently dead_end works by monkey patching require. This causes confusion and problems as other tools are not expecting this. For example zombocom/derailed_benchmarks#204 and #124. This PR utilizes the new SyntaxError#detailed_message as introduced in ruby/ruby#5516 that will be released in Ruby 3.2. That means that developers using dead_end with Ruby 3.2+ will experience more consistent behavior. ## Limitations As pointed out in #31 the current version of dead_end only works if the developer requires dead_end and then invokes `require`. This behavior is still not fixed for Ruby 3.2+ ``` $ ruby -v ruby 3.2.0preview1 (2022-04-03 master f801386f0c) [x86_64-darwin20] $ cat monkeypatch.rb SyntaxError.prepend Module.new { def detailed_message(highlight: nil, **) message = super message += "Monkeypatch worked\n" message end } # require_relative "bad.rb" # Note that i am commenting # out the require, but leaving # in the monkeypatch ⛄️ 3.2.0 🚀 /tmp $ cat bad.rb def lol_i-am-a-synt^xerror ⛄️ 3.2.0 🚀 /tmp $ ruby -r./monkeypatch.rb bad.rb bad.rb:1: syntax error, unexpected '-', expecting ';' or '\n' def lol_i-am-a-synt^xerror ``` Additionally we are still not able to handle the case where a program is streamed to ruby and does not exist on disk: ``` $ echo "def foo" | ruby ``` As the SyntaxError does not provide us with the contents of the script. ``` $ echo "def foo" | ruby -:1: syntax error, unexpected end-of-input def foo ```
Currently dead_end works by monkey patching require. This causes confusion and problems as other tools are not expecting this. For example zombocom/derailed_benchmarks#204 and #124. This PR utilizes the new SyntaxError#detailed_message as introduced in ruby/ruby#5516 that will be released in Ruby 3.2. That means that developers using dead_end with Ruby 3.2+ will experience more consistent behavior. ## Limitations As pointed out in #31 the current version of dead_end only works if the developer requires dead_end and then invokes `require`. This behavior is still not fixed for Ruby 3.2+ ``` $ ruby -v ruby 3.2.0preview1 (2022-04-03 master f801386f0c) [x86_64-darwin20] $ cat monkeypatch.rb SyntaxError.prepend Module.new { def detailed_message(highlight: nil, **) message = super message += "Monkeypatch worked\n" message end } # require_relative "bad.rb" # Note that i am commenting # out the require, but leaving # in the monkeypatch ⛄️ 3.2.0 🚀 /tmp $ cat bad.rb def lol_i-am-a-synt^xerror ⛄️ 3.2.0 🚀 /tmp $ ruby -r./monkeypatch.rb bad.rb bad.rb:1: syntax error, unexpected '-', expecting ';' or '\n' def lol_i-am-a-synt^xerror ``` Additionally we are still not able to handle the case where a program is streamed to ruby and does not exist on disk: ``` $ echo "def foo" | ruby ``` As the SyntaxError does not provide us with the contents of the script. ``` $ echo "def foo" | ruby -:1: syntax error, unexpected end-of-input def foo ```
I have the following in my Gemfile
However, when I run
bundle exec derailed bundle:objects production, I see the followingand when I run
bundle exec gem dep -RI see that dead_end is only used by derailed_benchmarks
I'm a bit confused why:
productionAny thoughts?
I'm not sure if it's related to this issue action_mailbox/engine still loads even after not requiring it in application.rb where @andrehjr says it's derailed_benchmark's use of
require "action_controller/railtie"The text was updated successfully, but these errors were encountered: