-
Notifications
You must be signed in to change notification settings - Fork 203
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Native version doesn't build since version 1.6.6 and Ruby 3.1.4 #411
Comments
Hi, @sirux88, thanks for opening this issue. I've reproduced what you're seeing on my development machine. Will spend some time on a fix this week. |
Thanks @flavorjones for investigating on this. |
@sirux88 Ah, ok, I figured out what this is.
My advice is to update your system/container to Rubygems 3.4.9 or later to work around this! Sorry for the inconvenience. I'm planning to get some work in on first-class musl support later this month, you can follow rake-compiler/rake-compiler-dock#75 for the upstream work that #372 is blocked on. |
Thank you @flavorjones for digging into this. I'll check how to upgrade to rubygems >=3.4.9 and if this is ok on my project |
* Implements Honeybadger.event by sync log call * API changes, add timestamp * fix #event: Use Hash() and reverse merge order. * feat: implement simple debug backend endpoint for events (#513) * Implement simple debug backend endpoint for events This currently is missing a queue and calls the backend directly from the agent. Should I implement an events_worker within this PR or in the PR that adds the server backend? * Refactor signature of events backend to take only one argument * WIP: Add worker * WIP start of worker spec * Worker spec successfully duplicated * Implement timeout mechanism using separate thread Given that the worker relies on the Queue as the main scheduling mechanism I saw no other way than to start a second thread that occasionally throws a message into the queue to check if the timeout is reached. This seems to work in testing. * Remove one timeout check, namespace config * Remove unused code * Add events worker to agent stop/flush commands * Fix debug message in events worker --------- Co-authored-by: Joshua Wood <josh@joshuawood.net> * Slightly bump sleep values in test to fix jruby tests There seems to be a slight difference in how sleep works in jruby so the timeouts in the tests did not hit predictably. * install sqlite dev package for rails tests * use sudo * Okay, sqlite problem seems to be based on rubygems issue sparklemotion/sqlite3-ruby#411 * I have no idea what I'm doing * feat: http(s) backend implementation for events (#520) * Implement simple debug backend endpoint for events This currently is missing a queue and calls the backend directly from the agent. Should I implement an events_worker within this PR or in the PR that adds the server backend? * Refactor signature of events backend to take only one argument * WIP: Add worker * WIP start of worker spec * Worker spec successfully duplicated * Implement timeout mechanism using separate thread Given that the worker relies on the Queue as the main scheduling mechanism I saw no other way than to start a second thread that occasionally throws a message into the queue to check if the timeout is reached. This seems to work in testing. * Remove one timeout check, namespace config * Remove unused code * Add server back end functionality for events This adds a minimal set of tests to ensure API conformance I've tested the code manually against "the real thing(tm)" * Add events worker to agent stop/flush commands * Fix debug message in events worker --------- Co-authored-by: Joshua Wood <josh@joshuawood.net> * Support Hash as first argument to Honeybadger#event (#521) This enables both signatures: # With event type as first argument (recommended): Honeybadger.event("user_signed_up", user_id: 123) # With just a payload: Honeybadger.event(event_type: "user_signed_up", user_id: 123) * Don't memoize events config The config is initialized after the agent is created (when the app loads). * Lazy initialize events worker This results in less change for current users—if you aren't using insights, the extra threads don't need to run. We could change this back in the future. --------- Co-authored-by: Joshua Wood <josh@joshuawood.net>
When trying to use the native version e.g.
gem 'sqlite3', force_ruby_platform: true'
in gemfile the build is not successfull.Steps to reproduce:
gem install sqlite3:1.6.6 --platform=ruby
results in
Details
I did some testing and it seems this is tied to psych version >4.0.x
Two sidemarks:
The text was updated successfully, but these errors were encountered: