<enebo[m]>
byteit101: yeah new issue as I am guessing you did (trying to get some coding done before I get mired in email/social media so I have not checked issues/mail yet)
<byteit101[m]>
oh nice, I'll check that out this evening
<enebo[m]>
byteit101: I have three commits on this tree. The last commit has instructions for how to hook this up and since I think you work different hours I thought I would share progress
<byteit101[m]>
Sounds good!
<enebo[m]>
Two unimplemented things beyond actually using it: 1) not going past startup interp so it does not magically break after 50 calls 2) Adding the jump/branch detection logic to error saying we do not support super in nested conditions
<byteit101[m]>
> Update for import/style consistency
<byteit101[m]>
Ah, I was going to do that later, I was keeping some weird constructs around to make breakpoints/etc easier. Not a problem though
<enebo[m]>
yeah in retrospect I should not have done that first :)
<enebo[m]>
and I did not even do all the files
<byteit101[m]>
Yea, that's on my to do list once everything is working
<enebo[m]>
but as a style teething commit you should see most of the style things we have in our codebase
<byteit101[m]>
There is a LOT of weird constructs and cleanup I want to do
<byteit101[m]>
yup
<enebo[m]>
Not sure if you want to try consuming that last commit in makeConcreteConstructorSwitch
<enebo[m]>
or try to hook it up
<enebo[m]>
I am almost done for today
<enebo[m]>
Conceptually you ask for the interpretercontext and if it is Exitable you can then ask for a state instance and just call interpret with it twice
<byteit101[m]>
Sounds good
<enebo[m]>
once for first section and a second time for the section after the super
<enebo[m]>
I had one question about return value of super but it should always be nil from Java and our Ruby should not really use it
<enebo[m]>
but if it does use it and it is null I believe we protect and return nil
<enebo[m]>
Still won't know until it is hooked up or I bother reading some more code
<enebo[m]>
I had also started to go down the path of using our 'full' interpreter since we will have a flow graph
<enebo[m]>
I backed off since I feel it can be done later
<enebo[m]>
but that will give us the ability to do better analysis for whether super is valid within nested regions
<enebo[m]>
It also would put it one step closer to making a jitted version but as I said this is simpler for now
<byteit101[m]>
Nice!
<byteit101[m]>
Unsure if I'll get around to hooking that up this evening, but over the weekend I created a series of jrubyfx tests so I should be able to test a few "real world" things
<enebo[m]>
tomorrow I will try and fix the two other problems I mentioned above