01:27 <headius[m]> G. Gibson: ok well let us know

01:27 <GGibson[m]> will do

16:09 <boc_tothefuture[> The way the system works (this part is out of my control) is that there will be multiple files that get executed, parsed, etc. Multiple threads will be executing within those files.

16:09 <boc_tothefuture[> Morning everyone. Have a JSR223 question.. I have been playing around with the various options here: https://github.com/jruby/jruby/wiki/RedBridge and can't quite get to the behavior that I think my users will expect..

16:09 <boc_tothefuture[> What I am trying to work through is instance variables. The users of this system will expect instance variables to be scoped to the file they are in. However, because when coming into a file it is just 'Object' the instance variables seem to get shared between files. Is there anyway for me to fix that and scope the instance variables to just the current file?

16:13 <boc_tothefuture[> If it helps illustrate it, this cucumber test is failing: https://github.com/boc-tothefuture/openhab-jruby/blob/ec6eafa5a907ff64d3583a69f9a9074e8571152f/features/ruby.feature#L41

16:17 <slonopotamus[m]> I observe that on JRuby+Windows, Timeout.timeout fails to interrupt IO.read (I'm reading stderr from a spawned process). Interrupting does work on JRuby+Linux though. Is it a known bug?

16:30 <headius[m]> slonopotamus: unfortunately that is the case, on Windows we still use the jvm classes for IO and in some cases they are difficult to interrupt

16:30 <headius[m]> Process control is one of those cases

16:31 <slonopotamus[m]> <headius[m] "slonopotamus: unfortunately that"> What are the cases that can be interrupted? I possibly could reorganize my code to avoid hitting unsupported cases.

16:57 <headius[m]> slonopotamus: I think most non-process IO should be ok

16:57 <headius[m]> perhaps this is one last project I should try to finish for 9.3

18:26 <boc_tothefuture[> Hey headius did you see my question above?

18:28 <headius[m]> oh 223 question

18:29 <headius[m]> instance variables can't be scoped like that unfortunately

18:30 <headius[m]> they are stored on the object and there is no internal indirection to do them per file... you would need to use separate containers with separate objects to keep the instance state for different users or files separate

18:35 <boc_tothefuture[> Ok, bummer... is there anyway for me to figure out I am in different files?

18:40 <headius[m]> well FILE will always be whatever file was parsed to get the code

18:40 <headius[m]> ugh `__FILE__`

18:40 <boc_tothefuture[> except.. FILE in this case will just says <script>

18:41 <boc_tothefuture[> because its just getting a stream...

18:41 <headius[m]> ahh hmm... through our JRuby-specific embedding API there are ways to give it a filename

18:41 <boc_tothefuture[> I guess, I was looking for a way in JRuby to get some sort of unique ID of that stream.

18:41 <headius[m]> I am not sure if that is possible through 223 API

18:41 <boc_tothefuture[> yeah, the javax.script.filename property. I have a PR in place to get that added to the base project.

18:42 <headius[m]> I am not familiar with that one, looking

18:42 <boc_tothefuture[> Hmm.. I think you told me about it? but it was a while ago.

18:44 <boc_tothefuture[> Details with link here: https://github.com/openhab/openhab-core/pull/1885

18:45 <boc_tothefuture[> But, no worries.. when that is merged, I could then use that filename.

18:45 <boc_tothefuture[> I guess I was hoping there was a way to get a hash of the stream or something.

18:45 <headius[m]> aha

18:46 <headius[m]> certainly possible I forgot about this but it doesn't seem familiar

18:47 <headius[m]> that looks like it will work but I think it will have threading issues

18:47 <headius[m]> it is just a single value for the whole container it seems

18:49 <boc_tothefuture[> It is synchronized when it loads them and changes the value before each manager.loadScript

18:49 <boc_tothefuture[> https://github.com/openhab/openhab-core/blob/00a6f3679d238f10b837e752b957dd79c1c8ee6b/bundles/org.openhab.core.automation.module.script.rulesupport/src/main/java/org/openhab/core/automation/module/script/rulesupport/internal/loader/ScriptFileWatcher.java

18:50 <boc_tothefuture[> So I think it shouldn't have thread safety problems in this use?

18:50 <headius[m]> ah yeah might be ok

18:50 <headius[m]> unfortunate that the 223 API doesn't provide a way to give a filename with the stream

18:50 <boc_tothefuture[> Although.. this does create a set of resources that can be run at anytime.. so those would hit the issue.

18:50 <headius[m]> directly I mean

18:51 <boc_tothefuture[> Ok, thanks.. I will have to ponder how to handle this.

18:52 <headius[m]> yeah I know you are leaning on this framework so that limits your options but feel free to brainstorm in here