[jruby-dev] Binding methods: configuration of per-call logic

2007-07-11 Thread Charles Oliver Nutter
I'm going to refactor the method-binding logic in RubyModule to be simpler and use fewer DynamicMethod types, by having it accept a configuration object per method that says what sorts of per-call logic it needs. For example.. defineMethod("times", CallFrame.FULL, CallScope.NONE, BindType.INST

[jruby-dev] Making the compiler smarter: optional scoping

2007-07-11 Thread Charles Oliver Nutter
Attached is a broken, experimental patch that makes the compiler aware of scoping method calls. In cases where no scope is needed, it will not create a dynamic scope and will instead use Java local variables for Ruby local variables. The performance increase is extremely impressive. ## before

Re: [jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Thomas E Enebo
All the recent changes removed an attribute called cref (a singly linked list) with a direct RubyModule parent reference. I would expect this to screw up applying a patch (or svn update) but not to screw up your current work (other than if you marked cref transient, then you should mark parent t

Re: [jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Alan McKean
I checked out the trunk again and everything seems to be okay. Except that there is a lot in RubyModule that wasn't there when I did the first version. I'm having trouble restoring the dispatcher when a persistent object faults back in. It seems to be an anonymous class. Is there a good way

Re: [jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Thomas E Enebo
This error was happening for 3292-3297. So you probably have one of those checkouts. -Tom On 7/11/07, Alan McKean <[EMAIL PROTECTED]> wrote: It must be my environment. Now it even happens when I try to build the jar from the unmodified trunk code. This is the same stack trace that I get when I

Re: [jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Charles Oliver Nutter
Alan McKean wrote: It must be my environment. Now it even happens when I try to build the jar from the unmodified trunk code. This is the same stack trace that I get when I run any of the samples. Yeah, that's pretty core, and certainly doesn't appear to be broken here. Update to current trun

Re: [jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Alan McKean
It must be my environment. Now it even happens when I try to build the jar from the unmodified trunk code. This is the same stack trace that I get when I run any of the samples. generate-method-classes: [touch] Creating /Users/amckean/Desktop/jruby/jruby/build/ __empty.rb [java] Exce

Re: [jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Charles Oliver Nutter
Alan McKean wrote: I checked out the trunk and get a NullPointerException in StaticScope.determineModule() when I run anything. Any suggestions? I am hoping to roll my changes in for a diff so I can send them along for potential checkin, but I can't get the head revision of the trunk to work.

[jruby-dev] NullPointer in StaticScope.determineModule()

2007-07-11 Thread Alan McKean
I checked out the trunk and get a NullPointerException in StaticScope.determineModule() when I run anything. Any suggestions? I am hoping to roll my changes in for a diff so I can send them along for potential checkin, but I can't get the head revision of the trunk to work. --

[jruby-dev] [jira] Created: (JRUBY-1201) TCPSocket.new(dest_adrr, dest_port, src_addr, src_port) missing from JRuby

2007-07-11 Thread Alex D (JIRA)
TCPSocket.new(dest_adrr, dest_port, src_addr, src_port) missing from JRuby -- Key: JRUBY-1201 URL: http://jira.codehaus.org/browse/JRUBY-1201 Project: JRuby Issue Type: