<headius[m]>
and init from Java can follow same path
<headius[m]>
yeah
<byteit101[m]>
yea, thats the benefit
<byteit101[m]>
and ratains dynamism vs hard jit
<byteit101[m]>
*retains
<headius[m]>
so I have come full circle then... basically this splits the initialize such that pre and post super can be called independently with the super becoming actual java super and fully init of java obj
<byteit101[m]>
It's been a fun show, my popcorn is almost empty :-D
<byteit101[m]>
> so I have come full circle then..
<headius[m]>
heheh
<headius[m]>
and slow on a new keyboard
<byteit101[m]>
ah yes! How's it feeling?
<headius[m]>
ok so conceptually I guess I like where this is now
<headius[m]>
it feels good but really drives home the fact that I kinda float rather than staying on home row
<headius[m]>
Tyo
<headius[m]>
eek
<headius[m]>
Typing prose feels ok but programming characters will take some gettting used to
<headius[m]>
no strain on wrists or thumbs though so that goal is happening
<byteit101[m]>
Nice!
<headius[m]>
what other input do you need from me right now
<byteit101[m]>
Ok, so to clarify: JIT: maybe better? should I investigate?. Split+super: current favorite?
<headius[m]>
I think split super is the best non-jit option right now
<byteit101[m]>
And for questions: the Design> Reify Options (ruby api)
<byteit101[m]>
cool, I'll start some cleanup in that domain then
<headius[m]>
but jit option would not change user side really at all if we do it later
<byteit101[m]>
yes
<headius[m]>
so what you have maintains current use case and user expectation, 👍
<headius[m]>
what concerns do you have about the current code?
<byteit101[m]>
Thanks, I was going for that. Only real change is the Reify Options sections for the ruby api
<byteit101[m]>
currently? needs some cleanup, plus two questions under implementation on the PR (end of big comment)
<byteit101[m]>
Lots of special cases and todo's litter the code with minor things for a demo, major things for a release (corner cases, etc)
<byteit101[m]>
Oh, and where to put code, but that's always present :-)
<byteit101[m]>
like right now it blows up if you don't have initialize or an initialize with a super :-O
<headius[m]>
ah well there's that 😀
<headius[m]>
looking at questions
<byteit101[m]>
Thanks, and thank you for your time answering my questions so far!
<headius[m]>
don't worry about perf right now, we can add caching any time
<headius[m]>
interface impl should be using indy too so we can do that in one sweep later
<byteit101[m]>
cool!
<headius[m]>
oh god, interface impl still uses RuntimeCache object
<headius[m]>
so basically using CachingCallSite for invokes
<headius[m]>
so that will be nicer with indy and the interface calls from Java will actually inline the Ruby method
<byteit101[m]>
Oh right: you said to rope enebo in on IR transforms after the election?
<headius[m]>
yeah he may have ideas on how to make it cleaner
<byteit101[m]>
I hope so, right now it's quite sketchy feeling :-)
<byteit101[m]>
Do you have a comment pending on that PR or shall I make a summary of this coversation?
<headius[m]>
go ahead with that... I have to move on to something else for the evening
<byteit101[m]>
Sounds good, thanks for your input!
<headius[m]>
will be around some tomorrow and then rest of the week no problem
<byteit101[m]>
I probably won't have much time to clean it up this week until friday, but sounds good
<headius[m]>
no worries
ur5us has quit [Remote host closed the connection]
ur5us has joined #jruby
_whitelogger has joined #jruby
_whitelogger has joined #jruby
ur5us has quit [Ping timeout: 264 seconds]
<fzakaria1>
Can I issue `require` in parallel to improve startup ?
<fzakaria1>
(i'm trying to find out a way to hack it to go faster)
<havenwood>
fzakaria1: Does it work for you to just require in separate Threads?
<headius[m]>
Nothing should limit that within JRuby itself