<fidothe>
I'm doing some stuff with a FunctionInterface and I'm not sure I understand how to make the magic add-a-method-any-method java aspect work without block passing, which may account for some of that
<headius[m]>
Hmm
<headius[m]>
You are seeing those errors when trying to implement an interface?
<headius[m]>
fidothe
<fidothe>
Yup
<fidothe>
Will make a minimal repro case after small child is fed and asleep
<headius[m]>
Great thank you
subbu is now known as subbu|lunch
nirvdrum has quit [Remote host closed the connection]
<jeremyevans>
Now that the uri lib is on GitHub, do you think you could submit a pull request for what you want?
<headius[m]>
I never got back to upstreaming anything and we still maintain our own fork of uri/generic
<jeremyevans>
I don't think it should be a problem to have RUBY_ENGINE checks in the library, as long as CRuby backwards compatibility is maintained
<headius[m]>
ok sure, I will see what I can put together
<jeremyevans>
headius[m]: OK. Technically, uri is maintained by akira, but he doesn't seem to be active. I can review and merge after the PR is submitted.
<headius[m]>
sure, looks like our diff is pretty small
<headius[m]>
jeremyevans: the only thing JRuby-specific this touches is our "java env" ENV_JAVA... would it be best to defined?(ENV_JAVA) or use RUBY_ENGINE?
<headius[m]>
for example I don't think that Truffleruby has this constant, but they might want to add it to reuse this functionality
<jeremyevans>
headius[m]: RUBY_ENGINE == 'jruby' is probably the best way
<headius[m]>
I suppose we can change it later if that happens
<byteit101[m]>
headius: Looks good, easy merges, but I realized that 9ccc321a1180fabaad86307fae2c87600c898ff1 is a really nasty merge for my PR
<byteit101[m]>
9ccc is enebo's arity->signature code
<byteit101[m]>
(from remove_block_stuff)
<headius[m]>
Ahh yeah that is a bit more invasive
<headius[m]>
jeremyevans I wonder if it is worth keeping the bind_call with the overhead of respond_to
ur5us_ has quit [Ping timeout: 240 seconds]
<headius[m]>
All bind call gets us is skipping the bound Method object
<jeremyevans>
headius[m]: The way I implemented it, there is no cost at runtime, since it defines different methods
<jeremyevans>
headius[m]: Purely switching bind_call to bind.call would make things slower. Granted, inspect is not a fast path :)
<headius[m]>
Oh right, did was not great on mobile
<headius[m]>
Diff
<headius[m]>
Yeah seems fine them
<jeremyevans>
Great. I guess either merge it or approve it and I'll merge it.
<headius[m]>
Ok gimme one min
<jeremyevans>
headius[m]: The CI was already setup for Ruby 2.4-2.7, so it looks like it was expected to work on Ruby 2.4+, so I think setting the required_ruby_version to 2.4 is fine.
<headius[m]>
Yeah and I saw you added tests as well 👍