Index

Show enters and exits. Hide enters and exits.

00:01:28	brixen	now 0ef372 will have special significance
00:01:34	rue	What kind of behaviour should sexps have?
00:01:39	Defiler	I'm going to take a dinner break
00:01:48	rue	Is that even Code As Data compliant? :)
00:03:31	brixen	we definitely have a problem calling private/protected methods
00:03:55	rue	Bah, I can stop any time I want
00:04:37	brixen	I can't, because Hash won't work without it
00:05:05	brixen	Class#initialize is trying to call a protected instance method on Class
00:05:29	brixen	which wraps a slot accessor
00:17:50	evan	and that's not working?
00:18:35	brixen	evan: oh, got a sec?
00:18:39	evan	sure
00:18:41	brixen	I'm in send_method
00:18:55	brixen	and I need docs
00:19:19	evan	what do you need?
00:19:23	brixen	well, I believe it is send method based on the opcode print out
00:19:30	brixen	what is task->call_flags?
00:19:36	brixen	what is msg.priv?
00:19:47	evan	msg.priv indicates if private methods can be considered
00:19:55	brixen	k
00:19:57	evan	it's not related to protected
00:20:05	evan	task->call_flags is how msg.priv is set
00:20:08	brixen	and that's set to task->call_flags & 1
00:20:12	evan	yep.
00:20:18	brixen	and then task->call_flags is set to 0
00:20:19	evan	there is a set_call_flags instruction
00:20:26	evan	yes, call_flags is reset to 0
00:20:40	brixen	so, calling a protected method, what should msg.priv be?
00:20:44	evan	doesn't matter.
00:20:54	evan	it's not considered for protected methods
00:20:57	brixen	k
00:21:26	evan	look at line 96 of sendsite.cpp
00:21:33	brixen	well, this isn't a dep thing
00:21:35	evan	thats where protected methods are validated.
00:21:44	brixen	k
00:22:57	brixen	this is common/class.rb:64 calling to bootstrap/class.rb:39
00:23:26	brixen	I don't understand the "not being called from the same module" comment in sendsite.cpp:97
00:24:16	evan	where is needs_cleanup marked protected?
00:24:22	brixen	common/class.rb
00:24:27	evan	ok, i see.
00:24:54	evan	the logic for protected is that the current self is a kind_of where the method is located
00:25:03	brixen	k
00:25:27	evan	you should back track
00:25:34	evan	it's possible that current_self is not being set properly
00:25:42	brixen	and current_self is?
00:26:15	evan	the current self
00:26:16	evan	:)
00:26:19	brixen	vs recv?
00:26:22	brixen	in msg
00:26:23	evan	yes
00:26:26	evan	totally different
00:26:29	brixen	got it
00:26:34	evan	because if I do
00:26:38	evan	blah.woo.go
00:26:49	evan	sweet = self
00:26:53	evan	sweet is current_self
00:26:57	evan	the return value from woo is recv
00:26:59	evan	when calling go
00:27:08	brixen	hah
00:27:11	evan	thats in the wrong order, but you get it
00:27:14	brixen	I'll try to parse that
00:27:15	brixen	heh
00:28:25	evan	msg.current_self should be set to task->active->recv
00:28:29	evan	by a higher layer
00:34:50	evan	all right, i'm headed out for a bit.
00:57:32	tarcieri	headius: you work at Sun?
01:00:13	headius	tarcieri: yes
01:01:59	tarcieri	headius: I knew a bunch of people who used to work at Sun's Broomfield campus
01:02:09	tarcieri	emphasis on used to... guess you're sticking around?
01:02:21	tarcieri	do they pay you to do JRuby stuff?
01:02:25	headius	jruby pretty much lives autonomously
01:02:27	headius	yes
01:02:30	tarcieri	noice
01:02:35	headius	full time, no strings attached so far
01:02:51	tarcieri	yeah I suppose everyone I knew who worked there was a server monkey
01:03:30	headius	still lots of them around...sun's going through a major tansition right now what with everything being open-sourced
01:03:37	headius	betting the company that it's the right thing to do
01:04:06	tarcieri	yeah, kinda seems like too little too late at this point
01:04:30	tarcieri	depends on what, I suppose
01:04:37	headius	at the very least it will guarantee solaris and java live on
01:04:46	tarcieri	yeah
01:04:48	headius	whether that means sun will live on, anyone's guess :)
01:05:03	tarcieri	I used to be a Solaris administrator... now I want nothing to do with it
01:05:25	headius	still seems more stable/solid to me than linux, and they finally realized that usability is kinda important
01:05:39	headius	the solaris 11 stuff is pretty nice
01:05:59	tarcieri	JDS was... uggghhh
01:06:15	tarcieri	when that came out most of my users chose to stick with CDE
01:06:37	tarcieri	I was all psyched for ZFS and it didn't exactly deliver what I was hoping
01:06:54	tarcieri	using lvm2 + xfs
01:07:55	headius	solaris guys are taking a lot of hints from debian and ubuntu
01:08:05	headius	lots of power struggles between the new and old way
01:08:17	tarcieri	I tried running Nexenta several times
01:08:19	headius	new is winning
01:08:24	tarcieri	good
01:09:29	drbrain	hrm, why is gsub! finding Kernel#gsub not String#gsub
01:11:08	brixen	drbrain: method tables seem a little wonky or something
01:11:24	brixen	so far, no methods on an instance of Hash are resolving
01:12:00	drbrain	I hooked up some config stuff, and now with vm/vm plus.rb it pukes out when $_ is nil
01:12:10	drbrain	because it does Kernel#gsub instead of String#gsub :(
01:12:13	brixen	this is what I see trying to send :initialize form Hash.new http://pastie.org/261372
01:12:29	brixen	s/form/from
01:12:57	drbrain	where is frame 6?
01:13:07	drbrain	oh, send_prim
01:13:41	drbrain	yours looks like a stack problem
01:13:54	drbrain	something didn't get filled in in msg properly
01:14:09	brixen	hmm
02:05:24	drbrain	ha! prim_tty
02:09:13	brixen	heh
02:14:27	drbrain	why do we have both Rubinius::RBX_VERSION and RBX_VERSION (toplevel)?
02:20:37	CIA-21	* Move Rubinius::Terminal constant to kernel, wire up IO#tty? to support it. Environment::set_rubinius_constants() is now gone, as it is not needed. - ...; c11533d - Eric Hodel
02:20:50	drbrain	well, we don't now :)
02:52:40	drbrain	sweet! now I get two backtraces!
02:54:06	drbrain	which is as it should be, I guess
02:56:41	drbrain	something seems to be off in the stack again
02:57:11	seydar	gratz!
02:58:54	drbrain	not gratz
02:59:01	drbrain	gratz sounds like a good thing
02:59:06	drbrain	this is not good
02:59:10	seydar	but you got 2 backtraces. i take it you didn't get them before?
02:59:59	seydar	btw, i've racked up a study plan for compilers, so in a month i'll have racked up 40 hrs. expect me to be mildly useful then!
03:00:00	drbrain	I'm getting them because the stack seems off
03:00:14	drbrain	Rubinius is jumping from String#gsub! to Kernel#gsub
03:00:19	drbrain	awesome
03:00:20	seydar	the stack is only used, in rubinius, for method arguments, right?
03:00:24	drbrain	what are you studying?
03:00:47	seydar	just compilers. i've been snagging dartmouth course info, and just going through papers
03:00:48	drbrain	well, the VM's stack seems messed up
03:00:53	drbrain	not the C stack
03:00:56	drbrain	cool
03:01:07	seydar	why do you reimplement a stack?
03:01:13	drbrain	and the VM stack is used for all kinds of stuff
03:01:30	seydar	do proceed good sir
03:01:34	drbrain	having a separate stack from the C stack makes a bunch of things easier
03:01:42	drbrain	like running multiple VMs in the same process
03:02:40	seydar	do you use the C stack at all?
03:02:50	drbrain	makes the GC a little simpler too, since it doesn't have to guess about the contents of the C stack
03:03:05	drbrain	the VM runs on the C stack
03:03:13	drbrain	and ruby code runs on the VM stack
03:03:29	seydar	but for running ruby code, no?
03:03:31	drbrain	but ruby C extensions would run on the C stack, of course
03:03:37	seydar	sounds icky
03:03:54	drbrain	no
03:03:55	seydar	doesn't that raise hell for you?
03:04:03	seydar	how do you get the two stacks to play nicely?
03:04:27	drbrain	the VM stack is just a chunk of memory, so it's no problem
03:04:38	drbrain	weren't you working on a scheme implementation?
03:04:42	seydar	yea, i am
03:04:59	seydar	got distracted on a brainfuck one, practicing simple optimizations of parsing
03:05:06	drbrain	you have something in there that is recognizable as a stack
03:05:30	seydar	bf or scheme? the scheme one has like, 20 lines of code in it atm
03:05:37	drbrain	scheme
03:05:50	drbrain	every time you apply a lambda, you push onto the call stack
03:06:09	seydar	and for rubinius, you push onto the VM stack
03:06:09	drbrain	rather, every time you apply a lambda in a non-tail-call position
03:06:29	drbrain	yes, but you push on for a bunch of other stuff too
03:06:42	drbrain	since it's a lower-level stack
03:06:44	seydar	the arguments?
03:07:06	drbrain	or for if
03:07:12	seydar	ah
03:08:06	seydar	dammit i wish i knew more
03:08:21	drbrain	you're young! it'll happen
03:08:29	seydar	but the world will pass!
03:08:41	drbrain	it'll still be there
03:08:50	seydar	when i FINALLY take a compilers course in spring '09, where will rubinius be? significantly further
03:09:00	seydar	and i'll be plaing catchup for a whiule
03:09:08	drbrain	there's plenty of important stuff to do, like making out with girls (or boys, should that be your thing)
03:09:56	seydar	full of wisdom ^^^^^^^
03:10:09	drbrain	if you're writing a scheme or bf interpreter, you're already far enough along
03:14:11	seydar	then i need to learn c++
03:14:22	drbrain	not really
03:14:30	seydar	plus, i'm following a tutorial on the interpreter for scheme
03:14:35	drbrain	I barely know enough to be functional
03:14:38	seydar	hehe
03:14:53	seydar	what can i do to help with the VM then?
03:14:55	drbrain	that's fine, if you were taking a class, you'd be doing roughly the same thing
03:15:16	seydar	im am so thirsty to struggle with something interesting
03:15:16	drbrain	in a couple days we'll be back to running specs
03:15:34	drbrain	then you can try to figure out why things work in shotgun, but not in the VM
03:15:41	seydar	yess
03:15:47	seydar	"here, have a cookie"
03:16:29	seydar	sweet
03:16:39	seydar	im psyched now
03:16:49	seydar	i've got 15 minutes to kill, need me to do something/
03:17:22	drbrain	nothing's coming to mind, since everybody is eating or whatever
03:17:30	drbrain	I'm trying to figure out why the stack seems messed up
03:17:46	seydar	pastie it
03:18:02	drbrain	ok
03:18:24	drbrain	so, in vm/llvm/instructions.rb in VMMethod::resume there's a #if 0
03:18:33	drbrain	if you switch it to a 1, a bunch of crap gets dumped out
03:19:07	drbrain	so if I run RBX_RUNTIME=runtime vm/vm -e 'puts 3 + 4'
03:19:17	drbrain	I get those exceptions you were talking about
03:19:32	drbrain	when built with the dump-crap mode, I get this output:
03:19:38	drbrain	http://rafb.net/p/pi3Qnl29.html
03:19:46	drbrain	so there are two possible problems here
03:20:05	drbrain	oh, let me back up a step
03:20:27	drbrain	when I add a yield_gdb self in Kernel#gsub, I see that self is a Regexp, not a String
03:20:38	drbrain	if I add yield_gdb in String#gsub!, I see that self is a String
03:20:45	drbrain	so, the possible problems:
03:21:28	drbrain	a) the stack is off by one, and op_push_self ends up putting self (the String) out of range, and the first argument, pattern becomes self for the lookup
03:21:34	drbrain	b) op_push_self is broken (unlikely)
03:21:42	drbrain	c) op_send_stack_with_block is broken (unlikely)
03:22:08	drbrain	yeah, op_send_stack_with_block is called a bunch of times, so its probably ok
03:22:13	drbrain	same for op_push_self
03:22:46	seydar	so where could an OB1 error happen?
03:23:42	drbrain	OB1?
03:23:49	seydar	off by 1
03:23:58	drbrain	that's hard to say
03:24:15	drbrain	I'm not familiar with the compiler or instructions.rb yet
03:25:47	seydar	ok. you know the setup of files better than i do, at least. lets start at the top
03:26:20	drbrain	well, first I was going to yield_gdb self in gsub!
03:26:25	drbrain	String#gsub!
03:26:38	drbrain	then breakpoint on op_send_stack_with_block to see if it was all set up right
03:28:21	drbrain	(I don't know if there's much you can do beyond listening, though :)
03:28:24	drbrain	:(
03:28:52	seydar	its fine, my dad tells me stories of when he would debug other people's code by talking to them
03:29:45	drbrain	ok, so I did that
03:29:56	seydar	whats the class of self?
03:30:05	drbrain	and I looked in the Message at the recv slot (which should be my String)
03:30:10	drbrain	it should be String
03:30:26	seydar	but what is it really?
03:30:33	drbrain	but instead I get: KERN_INVALID_ADDRESS at address: 0x54890824
03:31:05	drbrain	ok, this is really bad
03:31:22	drbrain	task->state is 0x1002400, msg.state is 0xe4458bc2
03:31:32	drbrain	and everything in this Message looks like it has a bogus address
03:31:43	seydar	sounds like you're on the edge of your memory
03:31:59	seydar	if i remember this one tangent a prof went on once about memory space and OSes
03:32:17	drbrain	oh, I was looking at the wrong one
03:32:27	drbrain	ok, now I've got something correct
03:32:35	drbrain	msg hadn't been assigned to yet
03:32:47	seydar	what file are you working in atm?
03:32:49	drbrain	ok, my recv is right
03:33:31	drbrain	send_stack_with_block in vm/instructions.rb
03:35:01	drbrain	oh, ha, I need to walk along further
03:35:07	drbrain	its reusing the current message
03:35:59	drbrain	so it still is all filled in for yield_gdb
03:37:16	drbrain	ok, so yeah, the stack is off by one
03:38:03	seydar	what is the equivalent of -> using .? doesn't -> pass pointers around?
03:38:33	drbrain	-> is used on pointers, . is used on non-pointers
03:38:51	drbrain	so if you see Object* foo, you would foo->method()
03:38:57	drbrain	for Object foo, foo.method()
03:39:55	drbrain	IIRC, Message& msg means msg is a reference to some other piece of memory, and has no local storage
03:40:26	seydar	gotcha
03:40:28	seydar	thanks
03:42:01	seydar	so what is calling op_push_self?
03:42:14	drbrain	the bytecode told it too
03:42:56	seydar	which bytecode? is it in send_stack_with_block?
03:43:05	seydar	wait, nvm
03:43:12	seydar	i understand what you're saying
03:43:13	drbrain	it's in runtime/core/string.rbc
03:44:19	drbrain	ok, so all these values are off by one, and there seems to be an extra nil on top of the stack
03:44:44	seydar	and is self not even on there?
03:45:03	seydar	You expected 'String -> Regexp -> String', and its 'Regexp -> String -> Nil'?
03:45:04	drbrain	self is down one further than it should be
03:45:33	seydar	if self were to be up one slot, would the nil still be there?
03:45:39	drbrain	at the start, the stack should have been (from the top)
03:46:04	drbrain	block, replacement, regexp, self
03:46:13	drbrain	instead it is nil, block, replacement, regexp, self
03:46:21	drbrain	yeah
03:46:30	seydar	bah humbug
03:47:15	seydar	can you take a look at what's below self in the correct run vs. the incorrect run?
03:47:40	drbrain	right now there's no correct run, so I'm mostly guessing what that would be
03:47:49	drbrain	below self is nil
03:48:02	seydar	that would complicate things
03:48:31	drbrain	its probably from some operation we're going to resume upon return from gsub!
03:48:37	drbrain	so it's really none of my business
03:48:56	seydar	so you think #gsub has the bug, not #gsub!?
03:49:47	drbrain	no, it's not in #gsub or #gsub!
03:50:03	drbrain	its in vm/instructions.rb *or* the compiler
03:50:20	Defiler	ooh someone said instructions.rb
03:50:21	seydar	but the compilter, isn't that the same from shotgun?
03:50:33	Defiler	It has been modified to suit the new VM
03:50:34	drbrain	since the same gsub, gsub! code works the same in shotgun
03:50:39	Defiler	and there could be problems with that transition
03:50:40	drbrain	no, there have been changes
03:50:51	drbrain	Defiler: I've got an off by one error
03:50:52	Defiler	we changed to left-to-right evaluation order
03:50:57	drbrain	try vm/vm -e 'puts 3 + 4'
03:51:05	drbrain	end up with it puking on $_
03:51:11	Defiler	dude vm supports -e now?
03:51:15	Defiler	That is frikkin awesome
03:51:22	seydar	damn right it is
03:51:24	drbrain	since it called Regexp#gsub instead of String#gsub like it was supposed to
03:51:29	drbrain	it tries, but fails :)
03:51:49	Defiler	OK, let me gdb this bad boy
03:51:59	drbrain	self (the String) ends up one further down the stack
03:52:07	seydar	grab a monster and a bag of m&ms - you guys got this
03:52:11	drbrain	the easy way is to add yield_gdb to String#gsub!
03:52:26	seydar	later homefries
03:52:27	drbrain	then add a breakpoint on op_send_stack_with_block when you get there
03:52:32	drbrain	later
03:52:43	drbrain	a j over the raise will get you there
03:53:28	Defiler	aha
03:54:35	drbrain	so when you get there, and n down to at least msg.stack
03:54:44	drbrain	you can p *(js->stack - 2) then rp $whatever
03:54:53	drbrain	and find that its a Regexp instead of a String
03:54:59	drbrain	and that the String is at -3
03:56:12	drbrain	oh, also, js->stack is nil, and js->stack - 1 is the block that should have been popped
03:56:19	drbrain	... so then I've got two extra values?
03:57:57	drbrain	I'm going to start again from the top
03:58:43	Defiler	Why is there a 0x0 on the stack?
03:58:51	Defiler	at stack - 3?
03:59:42	drbrain	I have a Regexp
04:00:16	Defiler	OK I should see the problem if I step from here
04:00:21	drbrain	http://rafb.net/p/lQZruo44.html
04:00:29	Defiler	I think I have it in my sights
04:01:11	drbrain	that's before the msg.block = stack_pop()
04:01:27	Defiler	Ok, so this is the _get_field call for 'def env' on BlockContext
04:01:45	drbrain	maybe I didn't push that
04:01:47	Defiler	I think it is a name conflict in C++
04:01:57	Defiler	because this is a blockcontext
04:02:03	Defiler	but it was showing MethodContext* in gdb
04:02:11	CIA-21	* Wire up BlockContext::env to a primitive so we can fetch it - http://is.gd/1ZkC; 1f5c22a - Eric Hodel
04:02:22	drbrain	ok, try now
04:02:51	Defiler	I'm going to finish stepping to test my hypothesis first
04:03:05	drbrain	if it wasn't bitching about $_, you weren't in the same place
04:03:33	Defiler	it was bitching
04:03:46	drbrain	ok
04:03:50	Defiler	OK, so the top two items on the stack are reversed
04:03:52	Defiler	that is the problem
04:03:57	Defiler	so I vote compiler
04:04:01	Defiler	but let's see
04:04:42	Defiler	ok, not compiler
04:05:15	Defiler	damn, phone call
04:05:29	drbrain	me too, since this seems pretty straightforward for the instructions leading up to it
04:10:38	drbrain	oh, its one of these that comes after
04:10:46	drbrain	between 36 and 48
04:32:22	brixen	hm, so attempting to find Class.allocate from a call to super() in Hash.allocate walks MT's in Hash, Enumerable, Object, and Kernel
04:32:28	brixen	so the same ancestors as MRI
04:32:41	brixen	but where is it supposed to look in Class?
04:33:57	drbrain	... that sounds wrong
04:34:11	drbrain	shouldn't it be Hash, Class, Module, Object, Kernel?
04:34:40	brixen	well, dunno
04:34:48	brixen	le'me try to trace it in shotgun
04:35:01	drbrain	since you're calling Hash::allocate, instead of Hash#allocate
04:35:06	brixen	yeah
04:47:30	Defiler	Is it just me or is the stack null here?
04:47:36	Defiler	I'm looking at context_get_field
04:47:40	drbrain	it is
04:47:52	drbrain	oh, I don't know about that
04:47:52	Defiler	That seems bad. :)
04:48:05	drbrain	but, there's a pass through BlockPass' bytecode
04:48:11	drbrain	and this is obviously bad
04:48:28	drbrain	dup_top is_nil goto_if_true
04:48:46	Defiler	p js->stack is 0x0
04:49:07	brixen	there was an sassert to prevent NULL from getting into the stack
04:49:12	drbrain	I'm not seeing that, but we're probably looking in different spots
04:49:57	brixen	in the define of stack_push
04:49:58	drbrain	no, this gsub! is called without a block
04:50:11	brixen	but not in stack_set_top
04:50:15	Defiler	msg.task->js->stack is 0x0 in rubinius::Task::send_message
04:50:23	Defiler	on the message that was built for this send
04:50:29	mass	is doing a restore wipe on his iphone
04:50:35	brixen	Defiler: perhaps add an sassert to stack_set_top
04:50:40	Defiler	That might be OK for all I know. just seems odd
04:50:41	mass	there is a crash report on it which causes the crash reporter to crash
04:50:55	drbrain	haha
04:51:40	mass	anytime I want a core file all I have to do is plug in a USB cable
04:52:21	mass	I wiped the crash report directory clean, plugged in the iphone, zipped up everything in the directory afterwards and attached it to a bug report.
04:53:01	brixen	according to ruby programming language, method lookup goes into included modules and ancestor modules, and then to superclasses
04:53:35	brixen	we're invoking method missing after looking in the modules
04:53:46	drbrain	still, there should be a Class andModule in that list
04:53:54	brixen	right
04:54:00	brixen	those would be superclasses
04:56:10	brixen	hmm, but class method resolution is a bit different
04:56:28	drbrain	ugh, this mismatch between compiler and instructions.rb is a PITA
04:56:34	drbrain	I can't find shit
04:56:36	Defiler	It would have to be, or it would be recursive, right?
04:56:39	drbrain	because none of the names match
04:57:01	Defiler	Maybe not
04:58:32	Defiler	How do I dump the gdb commands I have typed in this session so far to a file?
04:58:40	Defiler	or make a file that replays them?
04:58:53	drbrain	it should be in ~/.gdb_history
05:25:18	Defiler	oh hoh
05:25:37	Defiler	obj->reference_p() should always be true for Strings I assume?
05:25:37	Defiler	heh
05:26:17	Defiler	http://rafb.net/p/dU1BPV40.html
05:26:32	drbrain	should be true
05:28:52	Defiler	it is false for this one
05:29:38	brixen	what was the shortcut command for __show__(obj) ?
05:29:43	Defiler	rp
05:29:46	drbrain	rp
05:29:55	drbrain	see .gdb_init
05:30:03	drbrain	err, .gdbinit
05:30:04	Defiler	you can see me using it in that rafb paste as well
05:30:18	brixen	thanks
05:31:14	Defiler	huh. it appears to only be false in gdb, but the program flow is correct
05:31:30	Defiler	that's comforting ha ha
05:31:54	Defiler	what no it did it wrong wtf
05:32:40	Defiler	OK, what the hell is going on here? http://rafb.net/p/gUOaUa36.html
05:32:43	drbrain	I gave it up, it was too twisty
05:33:08	brixen	http://pastie.org/261459
05:33:14	brixen	that's what I'm talking about
05:33:18	brixen	I'll push in a sec
05:33:25	brixen	should make inspecting LT's much nicer
05:33:37	Defiler	brixen: Cool
05:34:04	Defiler	Is it possible to see what value T::type has taken on in gdb?
05:36:13	Defiler	OK, I need to disable optimizations
05:36:18	Defiler	do we have an env var for it?
05:36:21	Defiler	or do I just hack the rakefile
05:37:34	drbrain	I don't think we do
05:38:01	Defiler	Yeah, no -O at all
05:38:08	Defiler	but gdb says "static field type optimized out"
05:43:01	evan	yeah
05:43:06	evan	const static == gone entirely
05:47:10	CIA-21	* Teach LookupTable how to show itself in gdb. - http://is.gd/1ZrY; 1bad43c - Brian Ford
05:47:46	drbrain	sounds naughty
05:47:50	brixen	heh
05:47:54	Defiler	unified_load is so hideous
05:47:59	brixen	indeed
05:48:09	brixen	on both counts
05:51:00	Defiler	$LOAD_PATH is [0]
05:51:04	Defiler	Fun.
05:52:05	drbrain	it is?
05:52:10	Defiler	Yeah
05:52:15	drbrain	it was ['.'] on Monday
05:53:20	Defiler	$LOAD_PATH.unshift(*ENV['RUBYLIB'].split(':'))
05:53:23	Defiler	it is that line that does it
05:53:59	drbrain	awesome
05:54:46	Defiler	oh sorry
05:54:55	Defiler	it is this one
05:54:55	Defiler	$LOAD_PATH.insert($LOAD_PATH.index('.'), *additions)
05:58:08	Defiler	haha!
05:58:10	brixen	evan: got some spare cycles?
05:58:15	Defiler	Unable to load default compiler: No method 'ffi_get_field' on File::Stat::Struct (Class)
05:58:29	Defiler	FFI::Struct(File::Stat::Struct)#[] at kernel/platform/ffi.rb:604
05:58:47	Defiler	Array#insert is hosed up
05:58:54	Defiler	but I don't have time to look at why right now
05:59:08	Defiler	Probably it is Array#[]= since insert doesn't do much work
06:01:36	Defiler	OK, so what do gsub! and Array#insert have in common
06:02:10	evan	brixen: sure, wassup
06:02:17	brixen	evan: http://pastie.org/261473
06:02:26	evan	Defiler: ah, yes. the ffi_get_field/ffi_set_field aren't hooked up
06:02:35	evan	Defiler: they're all written in MemoryPointer
06:02:40	evan	need to be wired out primitives
06:02:57	brixen	evan: I'm trying to figure out why searching for .allocate from super() in Hash.allocate doesn't find Class
06:03:09	brixen	but it does in Rubinius::RubyConfig
06:03:13	brixen	for example
06:03:19	brixen	see top of pastie
06:03:31	Defiler	evan: How does $_ wire up in cpp?
06:03:36	brixen	I br on sendsite.cpp:72 and wait for :allocate to be called
06:03:40	evan	brixen: RubyConfig doesn't use super
06:03:44	brixen	yes
06:03:51	evan	brixen: so you mean why does going directly work
06:03:55	evan	but via super not
06:03:59	brixen	right
06:04:23	evan	ok
06:04:27	evan	there are a couple bugs I see
06:04:30	evan	in instructions.rb
06:04:36	evan	in all the send_super_* methods
06:04:39	Defiler	aha!
06:04:44	evan	msg.priv should be = true
06:04:47	evan	not using call_flags
06:04:51	brixen	k
06:04:56	evan	super calls by default can access private methods
06:05:03	evan	an no call_flags are emitted for supers
06:05:13	evan	so msg.priv was always false in a send_super_*
06:05:15	Defiler	We definitely have some kind of stack management bug here
06:05:27	evan	Defiler: there is no globals in C++
06:05:30	Defiler	Coercion error: Proc.to_str => String failed
06:05:32	Defiler	etc
06:05:43	evan	ruby globals
06:05:44	Defiler	$_ is the wacky last match thing
06:05:50	Defiler	and it needs a field, right
06:05:50	evan	right
06:05:52	evan	no
06:05:54	evan	it does not.
06:06:03	Defiler	I thought we had it on MethodContext before?
06:06:03	evan	it didn't in shotgun
06:06:05	drbrain	no, $_ is the last thing read via Kernel#gets
06:06:05	evan	nope.
06:06:20	drbrain	but in the $_ case, it jumps from String#gsub to Kernel#gsub!
06:06:26	drbrain	because recv is wrong
06:06:26	evan	which is wrong
06:06:33	evan	it's actually right
06:06:37	evan	i was debugging that earlier today
06:06:39	Defiler	Also, we init it to nil
06:06:45	evan	i'm not sure why it didn't use String#gsub
06:06:46	Defiler	but it has a guard against nil
06:07:11	evan	Defiler: thats expected
06:07:12	evan	try in MRI
06:07:15	Defiler	Yeah
06:07:21	evan	>> p $_ # nil
06:07:25	evan	>> gsub
06:07:28	evan	TypeError: ...
06:07:28	Defiler	but that's why that error was confusing
06:07:39	Defiler	because it was nil due to not being set correctly
06:07:46	evan	no
06:07:48	evan	thats not true
06:07:52	evan	it's never been set
06:07:55	evan	thats why it's nil.
06:07:59	evan	why do you think it should be set to something?
06:08:00	Defiler	That's what I just said
06:08:05	drbrain	it shouldn't be set, we never should have called Kernel#gets
06:08:12	drbrain	err, we never called Kernel#gets
06:08:23	Defiler	gsub you mean right?
06:08:28	evan	no
06:08:35	evan	if Kernel#gets is not called
06:08:37	evan	$_ is nil
06:08:43	evan	thats MRI behavior
06:08:52	evan	if you call Kernel#gsub and $_ is nil
06:08:55	Defiler	Oh.. look at that.. Kernel(Regexp)#gsub in the backtrace
06:08:56	evan	you get a TypeError
06:08:58	evan	Defiler: yes
06:08:59	Defiler	that should never happen, right?
06:09:02	evan	exactly
06:09:11	evan	you're getting confused by something COMPLETELY different
06:09:13	Defiler	Sorry, it is late here.
06:09:16	evan	that code is calling the wrong gsub method.
06:09:21	evan	it should be calling String#gsub
06:09:28	evan	but for some reason, it finds Kernel#gsub
06:09:34	drbrain	yeah, somehow there's an extra thingy on the stack
06:09:36	evan	I was going to track that a bit tonight
06:09:50	Defiler	It would be nice to add a locate_method test for this
06:09:59	drbrain	evan: the pattern you called String#gsub! with ends up as self for the next call
06:10:12	drbrain	it's not a problem looking up methods, it's that recv is wrong
06:10:33	evan	you mean the " str = gsub(pattern, replacement, &block)
06:10:35	evan	line right?
06:10:45	evan	the receiver of the gsub message is something weird
06:10:47	evan	NOT self
06:10:53	evan	thats what you're assuming yes?
06:10:56	drbrain	right
06:11:01	evan	right
06:11:05	evan	i wasn't going to assume that
06:11:07	evan	i was going to verify it
06:11:30	drbrain	I walked around in gdb when it's setting up the #gsub call, and the stack is off by one
06:11:47	evan	oh! got a op trace run?
06:12:14	drbrain	http://rafb.net/p/IOPhCs80.html
06:14:30	evan	ok, one sec.
06:14:35	evan	i'm comparing it to a compiler stream output
06:14:47	evan	perhaps block_pass is going the wrong place on the stack
06:15:04	drbrain	maybe, there doesn't seem to be a block to pass in this code
06:15:09	evan	sure there is
06:15:14	evan	gsub(pattern, replacement, &block)
06:15:17	evan	the &block
06:15:17	drbrain	the call in __split_path__ above it is:
06:15:27	drbrain	path.gsub!(/\.(so|bundle|dll|dylib)$/, ".#{Rubinius::LIBSUFFIX}")
06:15:32	drbrain	should be nil
06:15:36	drbrain	that's all
06:15:38	evan	huh?
06:15:49	evan	i guess we're looking 2 different places
06:16:01	evan	i'm just trynig to figure out why gsub! isn't calling the right gsub
06:16:08	evan	how it got there doesn't directly concern me
06:16:14	drbrain	ok
06:16:27	evan	because this is pretty basic functionality
06:16:41	drbrain	in my particular case of analyzing this, it's called from __split_path__ when I do vm -e 'puts 3 + 4'
06:16:57	drbrain	so for this run through the opcode stream, the block should be nil
06:17:13	evan	ah! ok.
06:17:16	evan	i see why you bring that up
06:17:17	evan	ok
06:17:19	evan	thats good
06:17:23	evan	it's secondary confirmation
06:17:52	Defiler	I'm running a conditional breakpoint for when the message name is the symbol :gsub!
06:17:56	Defiler	hopefully that will work
06:18:15	evan	the compiler output is pretty clearly fucked up
06:18:35	evan	http://rafb.net/p/BSnxGP22.html
06:18:45	evan	thats 'gsub(1,2,&3)'
06:19:00	evan	see the weird [:find_const, 0] stuck in the stream?
06:19:04	evan	thats the problem
06:19:10	Defiler	Whoa
06:21:36	Defiler	are we sure class Alias in bytecode.rb is right?
06:21:54	Defiler	and if so, is VAlias just under it wrong?
06:21:56	evan	pretty sure
06:23:56	evan	Defiler: whats wrong with them?
06:24:33	Defiler	reversed args on the stack
06:24:35	Defiler	just wondering
06:25:02	evan	check Globals.add_alias
06:25:10	evan	perhaps it's expecting that order
06:25:19	evan	which probably should be changed
06:25:21	evan	if it does.
06:26:35	evan	yep
06:26:36	Defiler	This looks like blockpass bytecode
06:26:38	evan	class BlockPass was wrong
06:26:44	evan	it as leaving the Proc constant on the stack
06:26:47	evan	if the block was nil
06:26:49	evan	fixed
06:26:50	evan	testing now.
06:26:52	Defiler	oh, I see it
06:26:55	evan	yeah
06:26:59	evan	thats just a fuck up my be
06:27:05	evan	when rotating the stack
06:27:32	evan	yep
06:27:34	evan	that fixed it.
06:27:37	evan	i'll push
06:27:45	Defiler	cool
06:28:35	CIA-21	* Fix BlockPass from leaving crap on the stack - http://is.gd/1Zuw; 0eb736a - Evan Phoenix
06:28:58	evan	ok
06:29:03	evan	now $LOAD_PATH is getting Fixnums in it
06:29:05	evan	it seems
06:29:12	Defiler	I have a fix for that here
06:29:20	Defiler	but it presumably is caused by another problem we actually need to fix
06:29:27	evan	oh?
06:29:32	evan	wassat
06:29:34	Defiler	It was happening in Array#insert though, and it was too complicated to want to debug this with
06:29:47	Defiler	in loader.rb there is a call to insert that I switched to be an unshift
06:29:49	Defiler	and that made it happy
06:30:03	evan	ah, so #insert is busted
06:30:35	brixen	my turn
06:30:46	evan	Defiler: ok, did that
06:30:46	Defiler	Try it with that
06:30:51	Defiler	(pushed)
06:30:51	evan	see it blow up on ffi_get_field
06:30:56	CIA-21	* Use unshift instead of insert in loader.rb - http://is.gd/1ZuI; 83f5d4d - Wilson Bilkovich
06:30:58	Defiler	Yeah, cool.
06:31:38	evan	brixen: ok go!
06:31:52	evan	YAY!
06:31:52	brixen	how do we add assertions for msg.priv = true in these instructions tests?
06:31:59	evan	got 15 new HD channels today!
06:32:20	brixen	and I only see two send_super_* instructions, one already had msg.priv = true
06:32:26	brixen	changing the other didn't help
06:32:27	drbrain	I might consider getting cable if I could get only HD channels
06:32:49	CIA-21	* Adjust field widths to be more sane - http://is.gd/1ZuP; 57c23e6 - Eric Hodel
06:32:53	brixen	I don't see how send_xx is the problem when it's not looking in the right modules
06:33:15	evan	brixen: ok
06:33:19	evan	well, it was a shot in the dark
06:33:21	evan	and it was wrong
06:33:22	mernen	hey, I
06:33:25	mernen	whoops
06:33:55	mernen	hey, I've got a couple patches to make the test suite compile under gcc 4.2. They don't pass, but well, it's a start
06:34:11	mernen	just tested, it still passes cleanly on osx
06:34:24	mernen	any problem if I push them?
06:34:37	evan	mernen: what OS?
06:34:40	evan	we're all on gcc 4.2
06:34:41	evan	on OS X
06:34:54	mernen	ubuntu 8.10
06:35:04	brixen	I'm on 4.0.1 on os x
06:35:13	evan	oh!
06:35:14	Defiler	evan: are ffi_get_field and memorypointer_get_field different?
06:35:15	evan	i guess so am I
06:35:15	dbussink	4.0.1 is still the default on os x
06:35:17	evan	der.
06:35:20	mernen	yeah, I'm on 4.0.1 here too
06:35:27	evan	mernen: yeah, go ahead and push
06:35:50	evan	Defiler: all ffi_get_field should become memorypointer_get_field
06:35:56	Defiler	OK, thought so
06:35:59	evan	ditto for ffi_set_field
06:36:32	evan	brixen: alrighty
06:36:43	evan	a quick run down
06:36:53	evan	calling super() in Hash.allocate doesn't find Class.allocate
06:36:55	evan	correct?
06:36:59	brixen	yep
06:37:16	evan	I'm wondering if Hash's metaclass isn't hooked up right
06:37:22	evan	that could cause the problem too
06:37:24	brixen	me too
06:37:30	brixen	but Hash is pure ruby
06:37:44	evan	but the way Hash, the class, is created, isn't
06:37:44	brixen	so, not sure where to start tracking that
06:37:54	evan	wow, WAY too many commas in that sentence
06:37:58	evan	that was like comma crack
06:38:04	brixen	heh
06:38:10	Defiler	evan: Right now, ffi_get_field (now memorypointer) is hooked up via an attach_function call
06:38:18	Defiler	Shouldn't I make a method that wraps the primitive in the kernel
06:38:20	Defiler	instead?
06:38:25	CIA-21	* Use unsigned literals to make stricter compilers (gcc 4.2) happier. - http://is.gd/1Zv5; 8ca001a - Daniel Luz
06:38:27	CIA-21	* Declare test string buffers as char[] to ensure their mutability. Required for gcc 4.2. - http://is.gd/1Zv7; e1cce51 - Daniel Luz
06:38:29	CIA-21	* Cast comparison value to native_int to make stricter compilers happy. - http://is.gd/1Zv8; 32bc147 - Daniel Luz
06:38:38	evan	Defiler: it should be a primitive instead, yes.
06:39:32	brixen	evan: I don't understand what ya mean by Hash, the class, the way it is created, is not pure ruby
06:39:52	evan	brixen: you have
06:39:53	evan	class Hash
06:39:54	evan	at the top
06:40:03	evan	how does that go off and create a Hash?
06:40:16	evan	it uses the open_class instruction
06:40:21	evan	which is not in ruby
06:40:27	brixen	right
06:40:32	evan	thats what I mean.
06:40:35	brixen	k
06:40:44	brixen	there's no special constructor for it, is what I mean
06:40:47	brixen	by pure ruby
06:40:49	evan	right
06:40:51	evan	gotcha
06:41:02	evan	but the metaclass should be hooked up correctly by open_class
06:41:06	evan	so we should look in that direction
06:41:09	brixen	k
06:41:19	evan	the test_open_class should verify our assumptions
06:41:33	evan	i'm going to go watch project runway a bit
06:41:41	evan	but here's the test params
06:42:07	evan	the new class's metaclass should have a superclass that points to the metaclass of the classes superclass
06:42:15	evan	does that make sense?
06:42:19	brixen	yep
06:42:30	evan	so, write a test for that assumption
06:42:33	evan	and we'll go from there
06:42:35	brixen	k
06:45:20	evan	let me know how that goes
06:45:41	brixen	k
06:46:09	Defiler	hell yeah got it
06:46:30	Defiler	no such file to load -- compiler/init
06:46:36	Defiler	is the new error
06:47:01	evan	cool
06:47:04	evan	call it a night
06:47:09	evan	go play
06:47:10	evan	:D
06:47:44	Defiler	http://rafb.net/p/21pyzI13.html
06:47:47	Defiler	that looks OK, right?
06:48:17	evan	sure
06:48:30	evan	def memorypointer_get_field
06:48:33	evan	is quite a mouthful
06:48:34	Defiler	I guess the wrapper could be called 'set_field' but that name is so scary heh
06:48:34	evan	but fine.
06:48:47	Defiler	I can easily rename it
06:48:54	evan	I should probably banish 'field' from my method naming vocab
06:49:11	Defiler	How about 'set_at_offset' and 'get_at_offset'?
06:49:19	evan	great
06:52:05	Defiler	When we get mspec up and running I am going to take a chainsaw to kernel/common/compile.rb
06:52:13	Defiler	Chainsaw Maid
06:52:36	evan	it's been quite a pain
06:52:42	evan	but yeah,
06:52:47	evan	the impl is crazy.
06:53:08	Defiler	Not touching it until the require tests are passing though heh
06:53:20	evan	good plan
06:55:55	brixen	evan: did you explain this before? what happened to push_encloser?
06:56:16	evan	it was pre-StaticScope
06:56:26	brixen	ok, opcode docs are out of date
06:56:30	evan	so I finally eliminated it
06:56:36	Defiler	It turned out to be crazy hard to try to manage the scopes at that level
06:56:37	evan	add_scope is used now
06:56:51	brixen	k
06:56:55	CIA-21	* Avoid creating an extra BlockContext on every require - http://is.gd/1Zwz; fba6b2c - Wilson Bilkovich
06:56:59	evan	as the first instruction of a module/class body
06:56:59	CIA-21	* Rename memorypointer_*_field prims to _*_at_offset and wire them. - http://is.gd/1ZwA; 47ef8d3 - Wilson Bilkovich
06:59:22	CIA-21	* Correct the primitive name in ffi.rb (oops) - http://is.gd/1ZwE; 68af39d - Wilson Bilkovich
07:00:29	Defiler	http://rafb.net/p/3bbwr670.html
07:00:33	Defiler	leaving it there for the night
07:00:59	evan	Defiler: looks good
07:01:00	evan	!
07:03:12	drbrain	"compiler/init..bundle"
07:03:15	drbrain	ha!
07:03:55	Defiler	You probably still have that gsub commented out?
07:04:31	drbrain	nope
07:04:39	drbrain	it seems broken
07:05:11	drbrain	(the gsub)
07:05:11	Defiler	this code uses a ton of masgn
07:05:17	Defiler	Any chance we still have masgn problems?
07:05:23	drbrain	could be LIBSUFFIX or path
07:05:27	drbrain	err, gsub
07:05:41	drbrain	I'm seeing File.file? return false for a file that exists
07:05:59	Defiler	that sounds fruitful
07:07:29	drbrain	haha, to_s(2) doesn't work
07:11:31	drbrain	heh, this file is regular and a character device!
07:11:38	evan	hehe
07:13:55	drbrain	and something else
07:14:45	drbrain	I'm too tired to look at this any further tonight
07:14:49	drbrain	'night!
07:14:51	evan	nite!
07:27:48	evan	RAWR.
07:28:01	evan	64bit stat strikes again!
07:35:17	CIA-21	* Fix so ffi_stat calls the 64bit version of stat - http://is.gd/1Zz6; 0144441 - Evan Phoenix
07:35:20	CIA-21	* Fail hard if compiler/init.rbc is missing - http://is.gd/1Zz7; 5dd02c5 - Evan Phoenix
07:42:46	Defiler	badass
07:42:46	dbussink	evan: hmm, shotgun used compiler flags for those 64 bit fs support
07:42:53	Defiler	Called unbound or invalid primitive from: load_from_file
07:43:05	evan	Defiler: yep
07:43:08	evan	fixing that now
07:43:16	evan	we've got a pure ruby unmarshaler now
07:43:22	evan	just got to wire it in
07:43:32	Defiler	awesome
07:43:32	evan	dbussink: yeah, I know
07:43:34	evan	dbussink: thats a hack
07:43:54	dbussink	maybe we want a central header that handles this kind of stuff?
07:44:10	dbussink	so it sets up all defines for various platforms
07:44:32	evan	probably eyah
07:45:13	Defiler	I guess we need loader.rb to figure out whether it is running from an install or from a checkout
07:45:33	Defiler	Looking at that todo in there that wants to use the configure-based lib path
07:45:53	Defiler	If we turned that on we would end up loading installed libs when we wanted libs in the checkout I think
07:46:01	evan	well
07:46:06	evan	configure-based is compile time only
07:46:09	evan	it's not dynamic
07:47:36	Defiler	That's what I mean
07:47:49	Defiler	if we pull it in in the loader, it will be the same value every time, even if we are running from checkout
07:48:42	Defiler	There should really be a __DIR__ method like the __FILE__ method
07:48:58	Defiler	since so many people do File.dirname(__FILE__) blah
07:51:45	headius	evan: I had an idea
07:51:55	evan	gets the media on the phone
07:51:57	evan	:D
07:52:27	headius	actually, it's fading now
07:52:46	headius	I was thinking duby could target rubinius bytecode also but that wouldn't really gain you anything without static-typed opcodes
07:52:54	headius	it would basically just erase the types
07:53:32	headius	trying to think of a better way to build the native (C/ASM/whatever) backend
07:54:01	headius	dumping out C code is gross and problematic because where C requires explicit exit points Ruby allows implicit exit points
07:54:09	evan	yep
07:54:23	evan	so, what do ya get by dumping out rubinius bytecode?
07:54:25	headius	dumping JVM bytecode or rbx bytecode is trivial because you just return the last value
07:54:48	headius	since all bodies eventually leave something on the stack
07:54:56	headius	well, I guess not a whole lot
07:55:11	headius	the benefit of duby is in the inferred types
07:55:45	headius	I suppose if you added a send operation that could be statically bound at load time, it could make use of that
07:56:23	evan	hm
07:56:24	headius	argument types aren't really important, but if there were a static_send with a type path declaration, duby could emit that
07:56:59	headius	the protocol would be that it binds to the method present when that bytecode is loaded
07:57:11	headius	forever
07:58:00	headius	similar to JVM's invoke bytecodes but without requiring a static signature as well
08:00:34	headius	anyway
08:00:36	headius	back to work
08:02:13	evan	headius: how late do you work?
08:02:29	headius	until I can't keep my eyes open anymore
08:02:41	headius	llvm isn't stack-based, is it
08:02:45	evan	nope
08:02:49	headius	darn
08:02:49	evan	SSA
08:02:59	evan	it's super easy to make it stack based though
08:03:16	headius	that might be the easiest native backend for duby
08:03:17	evan	you just assign everything to a variable
08:03:22	evan	and return the name of the variable you assign to
08:03:29	evan	so you basically have a pointer to the top of the stack
08:03:36	headius	I'll have to look into that later
09:50:22	evan	HOT.
09:50:33	evan	got lib/compiler/init.rbc loaded.
10:43:21	dbussink	evan: party time!
14:53:48	rue	Good morning
17:13:31	rue	Evening
17:15:17	dbussink	rue: howdy
17:21:11	brixen	can I get another pair of eyes on this: http://pastie.org/261822
17:21:19	brixen	or eye if that's all you can spare ;)
17:21:31	brixen	the rbx section is from shotgun
17:21:36	brixen	not cpp
17:22:22	brixen	looks like we're not setting up F.metaclass.superclass correctly
17:22:54	brixen	I'm going to try to figure out what cpp does now
17:23:15	Defiler	check out attach_metaclass probably
17:24:38	brixen	Defiler: but would you agree that we appear to be doing it wrong in shotgun?
17:25:08	brixen	what I believe evan told me was that: F.metaclass.superclass == F.superclass.metaclass
17:25:15	brixen	which is true in shotgun, but false in MRI
17:25:25	brixen	so, I'm trying to figure out if evan is confused or I am :)
17:25:57	brixen	alternatively, parse this line:
17:25:58	brixen	22:42 evan >> the new class's metaclass should have a superclass that points to the metaclass of the classes superclass
17:26:29	brixen	Guest76243: cloking?
17:26:31	brixen	heh
17:26:51	brixen	or cloaking if you're a native speller :P
17:27:44	rue	Actually
17:29:09	rue	The superclass of an object's metaclass should be the metaclass of the object's class' superclass
17:29:11	rue	I think
17:29:22	brixen	rue: use ruby please :P
17:30:15	brixen	did you mean? F.metaclass.superclass == F.class.superclass.metaclass
17:30:22	brixen	because that's false in MRI
17:32:26	Defiler	what is Class.metaclass in rbx?
17:32:57	Defiler	Class itself, right?
17:33:14	brixen	MetaClass Class
17:33:30	rue	class A; end; class B < A; end; b = B.new; b.class => B; b.class.superclass => A; b.metaclass => MetaB; b.metaclass.class => (Metaclass or Class)
17:33:44	rue	Remember, MRI does not have MetaClass
17:33:52	brixen	it does
17:34:01	brixen	it just doesn't have a nice way to get it
17:34:16	brixen	it doesn't have a class MetaClass
17:34:21	rue	Correct
17:34:21	brixen	it's anonymous
17:34:37	rue	The class of a MRI metaclass is Class
17:34:49	brixen	same with rbx
17:34:49	rue	The class of an rbx metaclass is MetaClass
17:34:53	brixen	MetaClass < Class
17:35:06	rue	Not the same, you have an extra step in the chain
17:35:21	brixen	hmm
17:36:22	brixen	rue: so, you're saying this is correct in rbx: F.metaclass.superclass == F.superclass.metaclass # => true
17:39:24	Defiler	Is there code that isn't working, or does it just seem odd?
17:40:04	brixen	Defiler: super() is not working
17:40:34	Defiler	Got some sample code?
17:40:39	brixen	so, in rbx, F.superclass.metaclass == MetaObject, F.superclass.metaclass == MetaObject
17:40:41	Defiler	Or is it any use of super at all?
17:40:57	brixen	class Hash; def self.allocate; super(); end; end
17:40:58	Defiler	you have one of those reversed
17:41:04	brixen	not in rbx
17:41:06	brixen	try it
17:41:12	Defiler	You wrote the same line twice
17:41:17	brixen	erg
17:41:33	brixen	F.metaclass.superclass == MetaObject
17:41:33	Defiler	Does super work in instance methods?
17:41:36	brixen	in the second
17:42:09	brixen	in MRI: F.superclass.metaclass == MetaClass:Class (<Class:Class>)
17:42:18	brixen	F.metaclass.superclass == MetaObject
17:43:00	brixen	Defiler: I haven't tried instance methods
17:43:41	Defiler	I get "Called unbound or invalid primitive from: open_with_mode"
17:43:54	Defiler	when I try to run a script with an instance method call to super
17:44:28	brixen	k
17:44:33	evan	morning
17:44:37	brixen	morning
17:44:46	brixen	nice work on loading compiler/init :)
17:44:50	evan	thanks!
17:44:57	evan	went on a bit of a tear
17:45:00	evan	last night
17:45:03	brixen	no doubt
17:45:05	Defiler	Yeah, saw that. Heh
17:45:08	brixen	surprised you're awake
17:45:33	Defiler	brixen: What local changes do you have?
17:46:15	brixen	nothing that would affect this
17:46:36	Defiler	OK, so what happens when you try to run RBX_RUNTIME=runtime vm/vm some.rbc
17:46:44	Defiler	where some.rb is just (3 + 4).__show__ ?
17:47:08	rue	brixen: I have not quite decided yet :P
17:47:12	brixen	rue: k
17:47:19	evan	brixen: you can't look at the proper superclass in MRI
17:47:22	evan	to test that
17:47:26	evan	it skips it.
17:47:35	brixen	evan: ok, so shotgun is right?
17:47:45	evan	so you can't ask F.metaclass.superclass and get the correct answer
17:47:51	evan	yeah, shotgun is correct
17:47:53	brixen	k
17:48:04	brixen	I need help writing the test in open_class to show this
17:48:37	brixen	I tried this: http://pastie.org/261852
17:48:55	brixen	also, I'm still on page 0 of "how to write instruction tests" :P
17:49:04	rue	Hm
17:50:01	brixen	Defiler: I can't load all of runtime because I have a Hash class that expects to call super() in .allocate
17:50:13	brixen	and other kernel stuff needs to instantiate a Hash
17:50:30	evan	brixen: that seems fine
17:50:40	Defiler	Yeah, it looks OK to me too
17:50:50	brixen	evan: excellent, cus it fails :)
17:50:56	evan	bingo
17:51:20	Defiler	so the superclass of the attached metaclass is being set to what?
17:51:31	evan	oh wait.
17:51:32	Defiler	Can you add a local variable there and print it?
17:51:33	evan	hold on.
17:51:45	evan	test_open_class isn't asking for a subclass of TrueClass
17:51:50	evan	which is what you're testing
17:51:54	Defiler	ohhh yeah
17:51:56	brixen	yeah, that's my problem
17:51:58	Defiler	it is just a constant on TrueClass
17:52:10	evan	the 'task->push(Qnil)"
17:52:16	evan	replace the Qnil with an existing class
17:52:28	brixen	why wouldn't that be open_class_under?
17:52:31	evan	no
17:52:37	evan	open_class_under is for the syntax
17:52:39	evan	class A::B
17:52:39	Defiler	replace it with G(true_class)
17:52:44	Defiler	instead of Qnil
17:52:47	Defiler	and see what happens
17:52:47	brixen	k
17:52:49	evan	under refers to where the constant is located
17:52:51	evan	not it's superclass
17:52:59	brixen	and why is SS set to G(true_class)
17:53:05	brixen	I can't find docs on that either :P
17:53:25	Defiler	We are opening class 'C' inside TrueClass
17:53:34	Defiler	so the sendsite is on trueclass
17:53:38	evan	yes
17:53:40	brixen	k
17:53:45	evan	thats testing that open_class uses the current lexical scope
17:53:56	evan	the StaticScope stuff sets up the lexical scope
17:54:25	brixen	k
17:54:28	Defiler	It is a little confusing because the scenario being set up in this test is unlikely in real life
17:54:40	Defiler	but the same functions are being called at least
17:54:49	brixen	I basically understood that, but translating it into code is the problem
17:54:56	brixen	we need docs on how to write instruction tests
17:54:58	mass	good day everyone
17:55:03	mass	&yawn;
17:55:03	brixen	hides
17:55:06	Defiler	brixen: I will do some
17:55:15	brixen	Defiler: you will be my hero
17:55:16	Defiler	since I've got more of them to do today
17:55:17	brixen	again :)
17:55:18	mass	I nominate brixen
17:55:42	Defiler	Actually, I do have one question about them that evan can probably answer
17:55:57	brixen	ok, that test passes if task->push(G(true_class))
17:55:58	Defiler	When we declare the 'stream' opcode array in the test glue..
17:55:58	evan	yessum?
17:56:14	Defiler	Is it OK to fill that with zeroes first?
17:56:21	evan	fill what?
17:56:24	evan	your lunch box?
17:56:25	Defiler	the memory
17:56:38	evan	replace that with a real noun
17:56:53	brixen	right, so you don't have to do stream[1] = (opcode)0; all the time
17:56:58	brixen	or stream[n]
17:56:59	Defiler	In memory, 'stream' starts with the opcodes we are going to run for this test
17:57:04	Defiler	No, you would still need that
17:57:16	brixen	if they were all 0 to begin with?
17:57:20	evan	so whats 'that'?
17:57:29	Defiler	but when you ask gdb to print the stream, it has a bunch of values after the section that this test will run()
17:57:34	rue	Hm, no, I do not think that `B.metaclass.superclass == B.superclass.metaclass` should be true
17:57:50	Defiler	that I have been assuming are just previous values/objects residing in that region of memory. am I wrong about that?
17:58:19	Defiler	brixen: 0 isn't arbitrary.. when you see instruction tests setting it that way, it is because the opcode they are going to run takes arguments, and we are passing it a zero
17:58:33	Defiler	It just so happens that the tests often use 0 for their index, etc.. but it could also be 11
17:58:36	brixen	ok
17:58:42	evan	Defiler: i still don't get what you're asking
17:58:46	brixen	can we ditch run() as well
17:58:49	Defiler	Let me paste a gdb session
17:58:51	evan	brixen: no
17:58:59	Defiler	I don't see how, since you need to be able to have multiple per test method
17:59:00	evan	it has to be there
17:59:05	evan	otherwise, how is the code run?
17:59:09	brixen	why does it have to be a macro
17:59:15	brixen	why not the real code?
17:59:33	evan	because it abstracts the tests from knowing their setup
17:59:45	evan	which has saved me from having to edit every damn test a number of times
17:59:47	brixen	k, well with docs it will be better
17:59:52	brixen	right now it's just confusing IMO
18:00:13	evan	why?
18:00:15	brixen	run() // magic happens here
18:00:18	evan	so we can document it
18:00:18	Defiler	sec. someone at the door
18:00:29	evan	run() means "perform the opcode you're testing"
18:00:45	evan	perhaps a better name is in order
18:01:04	rue	perform_the_opcode_being_tested()
18:01:26	brixen	I'm just saying, looking at the tests was like my first month of learning spanish, I understood every few words but no sentences
18:01:31	evan	perhaps a bit verbose
18:02:03	evan	brixen: thats true of any framework you don't know.
18:02:04	brixen	probably 20 lines of docs at the top would do
18:02:30	evan	if you need a rails analogy, run == render
18:02:48	brixen	I don't need an anology
18:02:59	brixen	I need a connection between run() (empty meaning) and task->
18:03:03	brixen	which does stuff
18:03:11	Defiler	evan: http://rafb.net/p/VypFvj18.html
18:03:13	brixen	once I looked at the gen'd test file it made more sense
18:03:30	Defiler	That's what I mean.. 29 is the opcode we are testing, and that line is about to set the open_class index to zero..
18:03:42	Defiler	..but the stuff after it isn't going to be run just now in this test, no?
18:03:47	evan	Defiler: oh.
18:04:13	evan	Defiler: sure
18:04:21	Defiler	brixen: #define run(val) task->execute_stream(stream)
18:04:22	evan	just add a memset() in the setup code
18:04:30	Defiler	OK, cool. Doing that
18:04:41	brixen	Defiler: yeah, got that
18:04:53	Defiler	So we could just replace the 'runs' with calls to that
18:05:37	brixen	Defiler: if you could just doc a test, we'll see if that helps others with the confusion
18:06:30	Defiler	Will do
18:06:49	brixen	thanks
18:07:39	rue	This is a bit complicated, actually
18:09:21	Defiler	(gdb) p stream
18:09:22	Defiler	$1 = {0 <repeats 100 times>}
18:09:23	Defiler	hell yeah
18:10:05	rue	Because an MRI metaclass' superclass is always a metaclass
18:10:16	Defiler	god damn that is so much easier to read
18:14:06	drbrain	what's our status this morning?
18:14:38	evan	lib/compiler/init.rbc is loading
18:15:49	Defiler	I miss our old checkin bots
18:15:52	Defiler	CIA is terrible
18:16:24	evan	yeah
18:16:40	drbrain	I end up in kernel/delta/marshal.rb:13
18:17:38	evan	yep
18:17:39	evan	me too
18:17:49	evan	thats where we are
18:20:42	drbrain	args.last is broken :(
18:20:47	drbrain	I ran vm plus.rbc
18:20:51	drbrain	I'll poke at it
18:20:58	evan	k
18:21:01	drbrain	err, Array#last
18:24:48	drbrain	ah, at(-1) isn't working
18:27:18	brixen	oh man, I love rp
18:27:49	brixen	fixing up MetaClass and others to print nice info
18:27:59	brixen	like our rbx #inspect did
18:29:58	drbrain	how do I indicate a primitive failure?
18:30:06	drbrain	Array::aref should barf on -1
18:30:10	drbrain	it doesn't
18:30:42	brixen	raise
18:30:56	brixen	ArgumentError::raise e.g.
18:31:00	evan	Array::aref needs to throw a primitive failure
18:31:02	evan	in that case.
18:31:13	brixen	oh yeah PrimitiveFailed
18:31:18	evan	so the ruby code can handle it
18:31:31	evan	it's just eating the out of bounds indexes and returning nil
18:31:38	drbrain	ok, now I've got an example
18:31:47	drbrain	evan: I know, I wrote it :)
18:31:52	evan	ah!
18:31:58	evan	well then we've got the man for the job
18:33:38	drbrain	urg, TS_ASSERT_THROWS is backwards
18:33:43	drbrain	expected comes first!
18:34:22	brixen	not in OO, obj.should_throw something
18:34:50	drbrain	it's a unit test framework, so expected comes first
18:35:42	brixen	hm, perhaps it's the first unit test framework with correct ordering ;)
18:42:37	rue	By the by, I think the Github repo could use a bit of tightening
18:44:35	evan	tightening?
18:44:41	evan	a new belt perhaps?
18:45:11	rue	Corset? I was thinking git-prune (or whatever it was called)
18:47:41	evan	not sure what ya mean
18:48:05	Defiler	Are you saying we have unreachable objects to get rid of?
18:48:19	Defiler	oh, so we do
18:48:33	evan	github runs it automatically
18:48:38	evan	you can run it locally
18:48:41	evan	and should once in a while
18:48:57	Defiler	hoh dang that made things faster
18:49:03	evan	yeah
18:49:08	rue	Pruning, packing?
18:49:26	evan	git am (and thus git rebase) runs git gc automatically at certain times
18:50:08	rue	Ah yes, git-gc
18:50:20	Defiler	brixen: How about I comment up the first test in the file? Or should I do a sample one as a big-ass comment at the top?
18:50:26	Defiler	(re: instructions.rb)
18:50:39	brixen	Defiler: I think straight docs at the top
18:50:46	brixen	but do what works for ya
18:50:51	brixen	I'll give you feedback
18:51:00	rue	I was trying to think about what they named the GC command in git.. *facepalm
18:51:26	brixen	yeah, one of the more clearly named commands
18:56:33	evan	man
18:56:38	evan	iphone really makes my speakers freak out
18:56:44	evan	these computer speakers are really old
18:56:50	evan	i guess I need some newer, shielded ones
18:57:09	tarcieri	evan: its awesomely unshielded GSM?
18:57:26	tarcieri	i've had that problem with every AT&T/Cingular phone I've ever owned
18:57:47	drbrain	damn you CIA! where's my commits?
18:58:00	evan	github is looking at it
18:58:05	evan	their hook to send to CIA is busted.
18:58:07	evan	it seems
18:58:10	drbrain	I love how some cars in GTAIV have speakers that freak out
18:58:20	evan	yeah, when your phone rings?
18:58:27	drbrain	yeah
18:58:57	drbrain	I have the run_ruby task setting RBX_RUNTIME now
18:59:15	evan	k
18:59:17	drbrain	and vm/vm plus.rbc is now at the same spot as -e
18:59:43	evan	pulling in new code
19:00:02	drbrain	(or rake run_ruby[plus.rb]
19:01:45	evan	drbrain: still on marshal.rb:13, yes?
19:02:50	drbrain	yup
19:03:00	evan	you're working on that then?
19:03:14	drbrain	I thought somebody else was
19:03:20	drbrain	if not, I will
19:03:26	evan	go for it
19:03:33	drbrain	ok
19:09:10	evan	ooh, 'git log --graph' is cool.
19:09:36	rue	I think two years from now we will still find yet-undiscovered git secrets
19:10:18	tarcieri	I take it that's a git 1.6 thing
19:10:54	evan	might bee
19:10:55	evan	be
19:11:02	evan	it's like an ascii gitk
19:12:12	drbrain	gah, this is like walking in a minefield
19:16:39	drbrain	in IO#seek, how did prim_seek(-@buffer.size, SEEK_CUR) ever work?
19:16:46	drbrain	I don't see a #size on IO::Buffer in shotgun
19:17:13	drbrain	the problem seems to be we've read one too many lines from the rbc
19:17:19	drbrain	but IO#lineno is worthless for this
19:17:28	evan	oh
19:17:36	drbrain	oh, it was a String subclass
19:17:41	evan	it worked because IO::Buffer used to be a String subclass
19:17:42	evan	yeah
19:17:51	evan	.size should be replaced with .unused
19:17:58	drbrain	thanks
19:18:43	drbrain	@used ?
19:18:54	drbrain	or a new method that's @total - @used?
19:19:42	evan	it's already there
19:19:57	evan	IO::Buffer#unused
19:20:12	drbrain	oh, heh
19:20:23	drbrain	it was invisible, I swear!
19:21:15	evan	heh
19:27:25	drbrain	no, unused is wrong
19:27:30	drbrain	since shotgun has it
19:29:22	drbrain	looks like it should be @used
19:29:26	evan	hm.
19:29:30	evan	ah
19:29:33	evan	yeah, I see.
19:29:44	evan	that wants to rewind the current size of the buffered data
19:29:56	drbrain	yeah
19:30:05	evan	@used is how much data the Buffer contains
19:30:07	evan	gotcha
19:30:07	drbrain	and I'm getting EINVAL right now
19:30:13	drbrain	(I wired it up to FFI instead)
19:30:29	evan	calling seek?
19:30:35	drbrain	yes
19:30:39	evan	yeah
19:30:41	evan	you probably can't
19:30:42	drbrain	lseek
19:30:44	evan	yeah
19:30:48	evan	lseek is like stat
19:30:49	drbrain	since that what the io_seek prim did
19:30:54	evan	yeah
19:30:57	evan	if you use FFI
19:30:59	evan	you get the 32bit version
19:31:10	evan	but everything else is using 64bit stuff
19:32:49	evan	brb.
19:41:55	drbrain	is native_int big enough to hold an off_t?
19:42:22	Defiler	Platform-specific answer
19:42:41	drbrain	I'm thinking its never true
19:42:41	Defiler	if the platform makes off_t 64bit even with 32bit pointers, no
19:42:45	drbrain	err, not true for 32 bit
19:42:58	Defiler	the ML2N macro in shotgun could handle it though
19:43:05	Defiler	I can't remember what it is called in VM
19:43:11	drbrain	to_long_long()
19:43:18	Defiler	probably yeah
19:43:47	Defiler	brixen: OK, check that out
19:43:53	brixen	k
19:46:58	brixen	Defiler: can you pull and walk through how we'd test msg.priv = true in that commit?
19:47:48	brixen	also, could we clarify: * A Task for with that CM activated.
19:48:00	Defiler	Yeah, I am making some edits. sec
19:48:01	brixen	otherwise, those docs are super helpful
19:48:07	brixen	k
19:49:13	drbrain	damn, my primitive still failed
19:50:50	drbrain	is there a C++ way to do std::string + int ?
19:50:55	drbrain	I'm using std::string + sprintf
19:50:59	Defiler	brixen: OK. We can check this by having a second run() that changes the send we are doing to be non-private
19:51:47	brixen	drbrain: you could use a strstream << int I believe
19:52:03	Defiler	but we will have to figure out what to call to create this method hierarchy
19:52:19	brixen	Defiler: a second call to run() doesn't seem like it'd test what we want
19:52:33	Defiler	Well, the first run() would be what is there now
19:52:50	brixen	k, but when I changed that to = true, the test did not fail
19:53:00	Defiler	and then we would replace the sendsite
19:53:02	brixen	so, it must have been set up that way
19:53:03	Defiler	I guess
19:53:07	Defiler	and run it again
19:53:47	Defiler	msg.send_site = as<SendSite>(task->literals->field[index]);
19:53:59	Defiler	whatever we shove into the literals list for the test there
19:54:08	Defiler	will be the sendsite used by send_message
19:54:26	rue	drbrain: Yes, strstreams are best when constructing like that
19:55:39	Defiler	brixen: Check out vm/builtin/sendsite.cpp and look at the ::resolve functions, if you haven't
19:55:40	brixen	Defiler: I want to test that msg.priv = true before the task->send_message(msg) runs
19:55:54	Defiler	why?
19:56:10	brixen	because that is a precondition for me
19:56:22	Defiler	The message is not exposed by the instruction
19:56:25	brixen	send_message(msg) should have certain behavior based on what is in msg
19:56:28	Defiler	It just creates it locally and passes it into send_message
19:56:36	Defiler	You can test that in the send_message test
19:56:46	Defiler	but this is the opcode test
19:56:58	Defiler	and checking the internal state in the middle of the operation seems weird to me
19:57:10	brixen	the send_super_stack_with_splat is setting up a particular msg and invoking send_message
19:57:11	Defiler	You are emitting an instruction and asking it to do something for you on the stack
19:57:51	brixen	a good test to me is confirm msg is set right, and that results in a particular behavior
19:57:54	Defiler	This doesn't seem like the right place to test that
19:58:04	brixen	well, it's being done in that instruction
19:58:04	Defiler	Everything in this call path is now hard-coded to true
19:58:10	brixen	so I guess I'm confused
19:58:33	Defiler	msg.priv = true is done in the method_missing branch
19:58:39	Defiler	and in the instruction body you have msg.priv = TRUE
19:58:51	brixen	true == TRUE
19:58:55	Defiler	I don't think we need to test that the compiler knows how to set something to TRUE
19:59:12	Defiler	So I think what you are describing is a send_message test, not a send_super_stack_with_splat test
19:59:26	Defiler	Here, we should test that we can make a super send off the stack with a splat
19:59:34	brixen	well, one send_super_* had true and one had task->call_flags & 1
19:59:46	Defiler	Right. and we could have written a failing test for that
19:59:56	Defiler	because a properly-constructed send would fail
20:00:00	brixen	those aren't equiv according to evan
20:00:09	Defiler	what aren't equiv?
20:00:09	brixen	so I want to see a test that makes one fail and the other not fail
20:00:13	evan	because call_flags can be 0
20:00:17	evan	and for super calls
20:00:23	evan	you ALWYS look for private methods
20:00:29	drbrain	is strstream even standard anymore?
20:00:31	Defiler	Yeah I know why it is hard-coded
20:00:33	Defiler	That is correct
20:00:49	Defiler	What I'm saying is that testing that it is true in a function that has it set literally like that seems wrong
20:00:53	Defiler	and silly
20:01:04	Defiler	Here, we should test that the send works
20:01:13	brixen	here's my question: "Why didn't the test fail when I changed msg.priv = task->call_flags & 1 to msg.priv = true?"
20:01:14	rue	drbrain: The header is sstream, the classes are ({o,i})stringstream
20:01:36	evan	brixen: which test?
20:01:37	evan	i'll look.
20:01:40	Defiler	Why would it fail?
20:01:51	brixen	git show 5a4b58c
20:02:03	brixen	Defiler: because if they are equivalent, I didn't need to change anything
20:02:10	brixen	why wouldn't it fail?
20:02:12	evan	they're not equiv
20:02:18	Defiler	The method you are trying to call isn't private
20:02:24	evan	because the method isn't private
20:02:35	brixen	then that line of code is not tested
20:02:36	Defiler	msg.priv means it allows private sends
20:02:41	brixen	I got that part
20:02:41	evan	brixen: yes, thats correct.
20:02:46	brixen	I want a test!
20:02:47	Defiler	That is correct, and is what I was trying to tell you
20:02:48	rue	drbrain: istringstreams allow pretty useful scanf-type reading. Plus you can make your own types work with them by providing the << functions
20:02:49	evan	this is not testing that msg.priv = true
20:02:53	evan	brixen: so write one!
20:02:54	Defiler	because we need two run() calls to test it
20:02:55	drbrain	hrm, I'm getting EBADF
20:02:57	evan	make the super method private
20:03:20	rue	And by << I mean >>
20:03:23	Defiler	leave the test as is, and then under it, twiddle the super method to be private, make the stack look the way you want, and run again
20:03:39	brixen	thanks
20:03:50	Defiler	that test will pass, though, since this code looks right
20:03:55	Defiler	(now)
20:04:13	Defiler	but you can change msg.priv = FALSE and see it fail
20:04:13	evan	and you can verify it by setting msg.priv = false
20:04:15	evan	and seeing it fail
20:04:18	evan	then change it back to true
20:04:19	evan	and it should pass
20:04:24	Defiler	*highfive*
20:04:31	evan	*highfive*
20:04:34	brixen	yay
20:05:57	drbrain	lunches
20:06:30	evan	dito
20:06:32	evan	ditto
20:06:41	evan	just gave fog a bath
20:06:42	evan	that was fun.
20:08:31	mernen	ugh, bad news... strnstr is a BSD-only function
20:09:20	mernen	some say it's flawed on osx 10.4 too
20:09:53	evan	mernen: ack.
20:10:00	evan	well, you can replace it with strstr()
20:10:07	evan	ust be sure that the buffers are null terminated.
20:10:15	mernen	is that safe?
20:10:23	evan	if they're null terminated, yeah.
20:10:25	evan	though, actually
20:10:28	evan	now that I think abuot it
20:10:34	evan	strstr has the wrong logic
20:10:37	mernen	I mean, can you actually be sure they're null-terminated strings
20:10:40	evan	because we allow our buffers to have nulls
20:10:45	mernen	exactly
20:10:51	evan	so strstr is going to fail when it should pass
20:10:52	evan	in some cases
20:10:55	mernen	strnstr seems to also be a bad choice
20:10:59	evan	yeah
20:11:07	evan	sounds like it needs to be a hand coded loop
20:11:09	mernen	there's memmem, but guess what... GNU-only
20:11:10	evan	or something else
20:11:18	evan	ah, hrm.
20:11:20	evan	well, i have to run to lunch.
20:12:04	mernen	I'll see if I can write something later
20:12:11	mernen	a simple dumb scanning loop would do anyway
20:18:03	brixen	evan: *ignoring* included modules, is this correct so far? http://pastie.org/261964
20:20:59	Defiler	What is the MetaClass (Object) part?
20:21:03	Defiler	I didn't realize there were two links there
20:22:58	brixen	F.metaclass.superclass == Object.metaclass
20:23:06	brixen	or, IOW
20:23:11	Defiler	oh really
20:23:17	Defiler	oh I remember this yeah
20:23:19	brixen	F.metaclass.superclass == F.superclass.metaclass
20:23:55	Defiler	well, that explains why Object.metaclass.superclass has to be Class
20:24:02	Defiler	and not some other metaclass
20:24:17	Defiler	otherwise you would jump over it
20:25:29	brixen	grabbing some food and ibuprofen for this killer headache
20:25:34	Defiler	=(
20:25:39	Defiler	diagram looks good though
20:27:09	brixen	btw, we need more explicity criteria for test completeness
20:27:19	brixen	that previous test was apparently correct, but incomplete
20:27:28	brixen	too bad we don't have heckle for cxxtest
20:27:30	Defiler	Sure. Once we have tests for everything, we should do another pass and look for holes
20:27:37	Defiler	haha imagine how long that would take to run
20:27:44	brixen	heh, true
20:28:23	brixen	also, opcode docs need overhaul
20:28:31	brixen	e.g. no push_encloser
20:29:57	Defiler	Yeah
20:30:07	Defiler	some don't even have docs
20:31:46	rue	So by convention the metaclass hierarchy can be made parallel
20:32:05	Defiler	I like seeing it in the order it will really be checked
20:33:57	rue	In MRI, F.metaclass.superclass == F.superclass.metaclass.superclass
20:34:05	goodney	evan: you around?
20:34:08	Defiler	He's at lunch
20:34:15	goodney	a
20:34:16	goodney	h
20:34:19	goodney	so i see
20:58:58	rue	MRI breaks down badly
21:07:05	brixen	cpp has the classes set up the same as that diagram, which I'm pretty sure is correct
21:07:14	brixen	so lookup must be broken somewhere
21:08:29	brixen	and :allocate is in Class.method_table as a CM, not a MV, so it must be public
21:10:37	brixen	/* If this was a private send, then we can handle use
21:10:37	brixen	* any method seen. */
21:10:40	drbrain	why is lseek() giving me EBADF?
21:10:55	drbrain	I don't see where the fd gets closed
21:10:56	brixen	drbrain: does it work from C++, not FFI?
21:11:08	drbrain	it does not work from either
21:11:17	brixen	in that comment above, what is a "private send"? that's not ruby terminology
21:11:29	drbrain	tries some things
21:11:35	rue	Essentially the method lookup works so that if B < A; mb = B.new.metaclass; mb < mB < mA. I think
21:11:45	brixen	rue: I want to word it better
21:12:24	brixen	"If we're allow to look at private methods, use any method that matches +name+" ?
21:12:40	brixen	allowed
21:13:02	Defiler	brixen: private send means a send without a receiver
21:13:11	brixen	Defiler: ah, thanks
21:13:17	Defiler	without a lexical receiver, I should say
21:13:21	brixen	right
21:13:29	Defiler	or, I guess, one with literal self as the receiver
21:13:41	Defiler	but I think the sexps are a little different. doesn't effect us here
21:13:49	Defiler	affect
21:14:05	brixen	we should have a glossary
21:14:21	brixen	I shall make one right this minute
21:14:58	Defiler	awesome
21:16:36	brixen	I guess we should rdoc it
21:16:50	brixen	what is a definition list in rdoc?
21:17:02	Defiler	I have no idea how to use rdoc =(
21:17:02	drbrain	use foo::
21:17:12	drbrain	ri RDoc::Markup
21:17:36	Defiler	Nothing known
21:17:52	drbrain	you must not have RDoc 2
21:17:55	drbrain	sec
21:17:55	Defiler	main RDoc page has some stuff though
21:18:26	drbrain	http://rdoc.rubyforge.org/classes/RDoc/Markup.html
21:18:28	Defiler	god this is huge
21:19:09	Defiler	cool though
21:19:15	Defiler	I didn't know about foo::
21:20:56	Defiler	Wow rdoc just uses regexps to parse
21:21:20	drbrain	depends
21:21:35	drbrain	it's got a son of irb for ruby parsing
21:21:42	drbrain	for everything else, regexp
21:22:48	brixen	drbrain: would you mind reformatting if I push? then we have an example
21:23:13	drbrain	ok
21:25:17	brixen	drbrain: ok, pushed. feel free to revise however
21:25:50	brixen	we should be throwing words/phrases into that whenever, regardless of whether we have a good definition
21:25:52	Defiler	so CIA is totally jacked up I guess
21:26:02	brixen	it's not CIA
21:26:04	brixen	it's github
21:26:07	drbrain	which file am I looking at?
21:26:13	brixen	doc/glossary.txt
21:26:16	Defiler	yeah, saying zargle:: wtf is this? would be an awesome patch
21:27:02	drbrain	I'll make these changes, but also point them out as I go
21:27:14	brixen	that blurb on private send should probably cross-ref "lexical" blah
21:27:18	brixen	k
21:27:22	brixen	that helps
21:27:30	drbrain	to continue a <pre> block, you need to add whitespace on the empty lines matching the indent
21:27:43	brixen	ahh, I stripped
21:28:18	drbrain	and the "A method call" line was indented, which would turn that into a <pre>, probably not what you wanted
21:28:28	brixen	right
21:28:53	drbrain	try that
21:28:57	drbrain	I pushed
21:29:00	brixen	k
21:29:43	brixen	cool, thanks
21:30:02	drbrain	RDoc is semi-smart about huge <dl>s, but it's eaiser to just use <h1>, <h2> usually
21:30:20	drbrain	you have to maintain indentation all over the place and it gets to be a hassle
21:30:28	brixen	k
21:34:34	drbrain	wait, what? it was closed!
21:37:33	drbrain	oh, hah, I'm a dummy
22:07:07	brixen	ok, pushed more glossary, please check it
22:11:57	john	just curious, what are the plans for rubinius when ruby 1.9 comes out? Are you going to make a separate version that has compatibility with 1.9?
22:12:26	zenspider	not that I know of... not yet planned that way at least.
22:12:42	brixen	probably have to cross that when we know what 1.9 will actually be
22:12:52	brixen	for now, pure speculation
22:12:54	zenspider	I'd rather see us provide a gem or something that shims 1.9 functionality onto rubinius
22:13:07	brixen	which would be pretty easy
22:13:27	Defiler	We can do that now that the parser is in ruby
22:13:31	john	ah, just wondering, didn't they already release their release schedule for 1.9? I thought they had it set for december 08, i read that somewhere on the redmine.ruby-lang.org site wiki
22:13:35	Defiler	Because 1.9 has syntax changes
22:14:37	zenspider	Defiler: yup. it'd be easier if we switch to a subclassable PEG parser, but that is gonna be a while off.
22:15:05	jbarnette	zenspider: have you been looking at any specific PEG parsers?
22:15:27	jbarnette	zenspider: or thinking about rolling your own?
22:18:58	zenspider	<zenspider> I've started a port of ometa, but haven't had time to go back and
22:19:09	zenspider	finish it off. [14:16]
22:19:20	zenspider	<zenspider> I might be doing it wrong, as a straight pull port, instead of
22:19:31	zenspider	doing a push port. [14:17]
22:19:34	zenspider
22:19:39	zenspider	stupid cafe network
22:19:44	john	thats where i read it: http://www.rubyinside.com/ruby-1-9-0-3-and-drops-support-for-9-platforms-970.html
22:19:50	john	Further to the above, Ruby 1.9.0-4 is due on August 25, 1.9.0-5 on September 25, 1.9.1 RC1 on October 25 (this will be quite a significant test - put it in your calendar), 1.9.1 RC2 on November 25, and 1.9.1 proper on December 20, 2008.
22:19:52	zenspider	the real problem is that ometa doesn't have good tests, so it is harder to port
22:22:10	jbarnette	zenspider: ah, right, I remember you working on that a few months ago
22:22:33	jbarnette	zenspider: i've been dicking around with Treetop, but I've been very frustrated by its lack of diagnostics
22:22:42	brixen	Defiler: I'm using __allocate__ for now and going back to finish debugging Hash
22:22:49	brixen	Defiler: then I'll look at super() some more
22:22:54	zenspider	it is a good system, I just think it is a bit much
22:23:06	zenspider	teh thing about ometa is it seems as stripped down as possible
22:23:13	jbarnette	absolutely
22:23:56	brixen	jbarnette: have you talked to nathan about it's lack of diagnostics?
22:24:05	brixen	(not sure what that is)
22:24:16	jbarnette	brixen: i just checked out the code last night; it's on my list :)
22:24:20	brixen	k
22:24:30	brixen	nathan is fun to talk to
22:27:28	rue	john_: Traditionally Christmas Day is the release day in the winter
22:28:00	rue	s/is/has been/
22:28:07	brixen	is there a way to do the equiv of bin/rbx describe foo.rb ?
22:28:18	john	rue: just saying what i read, thats all
22:35:14	zenspider	btw, I'm renaming :fixnum nodes to :int
22:36:09	rue	john_: I was mainly wondering why they were changing it :)
22:36:12	goodney	evan: you back from lunch?
22:36:20	evan	yep
22:36:23	rue	brixen: lib/bin/describe.rb ?
22:36:38	goodney	we're in a lockdown! I'm locked in OHE122 learning about databases!
22:36:46	rue	Uh-oh
22:36:47	evan	what?!
22:36:54	evan	was there a bomb threat?
22:36:54	rue	Is it _hierarchical_ databases?
22:36:57	john	rue: maybe they wanted to give themselves a few days of a buffer, /shrug but right now they released 1.9.0-4 2 days early but that doesn't say they will stay ahead of schedule
22:37:06	rue	john_: Hehe, might be
22:37:13	goodney	bank robbery, shootout => manhunt
22:38:08	evan	wow!
22:38:55	zenspider	rue: 1965 or so.
22:40:48	brixen	evan: k, pushed that diag in glossary
22:41:16	evan	coolness
22:41:18	brixen	evan: btw, Hash is setup exactly like that diagram, so it must be in method lookup
22:41:29	evan	alright
22:41:31	brixen	I've put in __allocate__ for now so I can get hash in
22:41:32	evan	want me to take a look at it?
22:41:50	brixen	well, I'd like to find it, but I'm sure you'll find it much faster
22:42:18	brixen	there's other problems with method calls it appears
22:42:36	evan	I don't want to styme your learning
22:42:40	evan	so just let me know
22:42:56	brixen	e.g. http://pastie.org/262107
22:43:17	brixen	I'd like to get Hash in instead of rebasing these 7 commits endlessly :)
22:43:32	brixen	perhaps I should just squash them
22:43:50	Defiler	or just push it and we can all try to fix it
22:43:59	Defiler	It isn't like anyone is using cpp for production right now :)
22:44:13	brixen	heh, true
22:44:26	brixen	well, there are two problems, the one I just pastied and super
22:44:32	Defiler	this is open source we do what we want *thug life*
22:44:40	brixen	so, we can work in parallel
22:44:54	evan	brixen: just use __allocate__ then
22:44:58	evan	agardiner: HEY!
22:45:04	agardiner	hi there!
22:45:05	brixen	agardiner: wb mate :)
22:45:15	agardiner	hehe! thanks!
22:45:24	agardiner	been a while...
22:45:28	Defiler	The season turns and agardiner is reborn once again
22:45:29	brixen	evan: __allocate__ gets me to http://pastie.org/262107
22:45:41	evan	hm, ok.
22:45:43	agardiner	Defiler: :-D
22:45:43	evan	let me look
22:46:11	agardiner	so, been trying to keep abreast of things via IRC logs...
22:46:12	evan	brixen: oh! probably a compiler bug.
22:46:14	brixen	self shouldn't be nil
22:46:20	evan	i'll bet the stack isn't reversed properly for +=
22:46:25	Defiler	oh hoh
22:46:26	brixen	should be instance of Hash
22:46:47	evan	op_asgn2 o/~ suuuucks o/~
22:46:53	evan	let me take a peek
22:47:25	Defiler	hrm
22:47:34	Defiler	this is OpAssign1 right?
22:47:42	evan	a.b += 1 is op_asgn2
22:47:49	Defiler	oh right
22:48:05	zenspider	Defiler: look at pt_testcase.rb
22:48:13	zenspider	tons of examples for all that stuff
22:48:16	evan	yep
22:48:36	evan	zenspider: can I make a change to op_assign_spec.rb?
22:48:40	evan	i may need to
22:49:54	Defiler	going to left-to-right sure got rid of a lot of swaps. heh
22:50:28	evan	yep.
22:50:34	evan	i see the problem
22:50:38	evan	the sexp form is different
22:50:41	rue	agardiner: Hi!
22:50:49	agardiner	hiya rue
22:51:18	agardiner	watcha up to?
22:51:49	agardiner	still wrestling with apache?
22:52:47	rue	agardiner: C++ branch now
22:53:00	agardiner	ah, all hands on deck, huh?
22:53:15	evan	yep
22:53:15	agardiner	sounds like things are getting close...?
22:53:37	drbrain	hrm, my IO::Buffer is not getting filled in with anything
22:53:52	zenspider	evan: just lemme know when you make a change there... I'm focusing on simple right now...
22:54:01	evan	ok
22:54:01	zenspider	you do some really nasty stuff to the sexps :(
22:54:02	drbrain	or something
22:54:15	evan	zenspider: it's really a tiny change
22:54:21	evan	i'm removing one element from a sexp
22:54:54	zenspider	how much farther is that from the UR sexp?
22:55:05	evan	let me check.
22:55:18	evan	s(:op_asgn2, s(:vcall, :a), :b=, :^, s(:lit, 8))
22:55:20	evan	versus
22:55:26	evan	s(:op_asgn2, s(:call, nil, :a, s(:arglist)), :b=, :^, s(:lit, 8))
22:55:47	zenspider	you're just doing the vcall rewrite?
22:55:53	evan	no
22:55:57	evan	it was
22:56:03	evan	s(:op_asgn2, s(:call, nil, :a, s(:arglist)), :b=, :^, s(:lit, 8))
22:56:04	evan	ack
22:56:11	evan	s(:op_asgn2, s(:call, nil, :a, s(:arglist)), :b, :^, :b=, s(:lit, 8))
22:56:14	evan	it was ^^ that
22:56:22	evan	i'm making it the PT version
22:56:37	zenspider	ah. ok. yeah... anything that gets us closer to UR is cool
22:56:41	evan	yeah
22:56:44	evan	it's closer
22:56:50	evan	i'm not going to make any changes that get further
22:56:52	evan	from UR
22:56:55	zenspider	that's essentially what I'm doing for all these errors
22:57:02	evan	thats just counter to what we're trynig to do
23:00:17	zenspider	evan: http://rafb.net/p/EFamUI27.html that's what I've got currently. it already diverges from your changes in that you didn't actually map some stuff
23:00:22	zenspider	but it seems right
23:01:05	evan	cool
23:01:07	evan	looks right
23:01:29	zenspider	you don't have :symbol, do you?
23:01:34	zenspider	I haven't seen it.
23:01:38	evan	no
23:01:42	evan	we don't
23:01:45	zenspider	you just seem to use them bare instead (see :alias)
23:01:58	evan	there are few places where they're bare, yes
23:02:12	evan	a literal symbol shows up in [:lit]
23:02:55	zenspider	I thought you said yesterday that there was no reason to use lit anymore?
23:03:01	zenspider	or did I misunderstand the above?
23:03:18	evan	must be a misunderstanding
23:03:25	evan	I don't think I said that
23:03:28	evan	I don't care too much
23:03:43	evan	i like having them be explicit
23:04:06	evan	but not super concerned about this
23:04:23	evan	the only one I can think of thats more concerning is Regexp
23:04:25	zenspider	interesting. ok.
23:04:28	evan	as I told ya last week
23:04:52	evan	the rest don't require deconstruction
23:04:54	evan	Regexp does
23:04:56	zenspider	yeah... that is the only one that might get folded back into UR...
23:05:02	zenspider	range does
23:05:05	evan	ah yes
23:05:06	evan	range
23:05:18	evan	the Range one is also for consistence
23:05:23	evan	since there is already a :dot2
23:06:09	drbrain	how do bytes get into an IO::Buffer?
23:06:17	zenspider	but dot2 is supposed to be for non-literal object ranges...
23:06:24	zenspider	drbrain: they walk
23:06:26	evan	drbrain: Scheduler.send_on_readable puts them in there
23:06:35	zenspider	running was deemed too hazardous
23:06:47	evan	zenspider: yeah, but i figured that having them in lit was an optimization
23:07:08	drbrain	I'm coming to the conclusion that it is not doing its job
23:07:22	evan	drbrain: it's possible.
23:07:37	evan	drbrain: b event.cpp:106
23:07:40	evan	in gdb
23:07:42	evan	then run it
23:07:43	drbrain	ok
23:07:47	evan	thats where an IO::Buffer is filled
23:12:35	zenspider	evan: wow you're inconsistent :P
23:12:42	evan	eh?
23:13:02	zenspider	optimize... deoptimize... make up your mind.
23:13:21	evan	i'm not an absolutist
23:13:24	evan	i get to change my mind
23:14:47	brixen	evan: in class Hash; def self.allocate; super(); end; end <-- what should msg.current_self be ?
23:14:53	brixen	at the point of super()
23:15:05	evan	Hash, the object.
23:15:31	brixen	not nil
23:15:34	brixen	IOW
23:15:37	evan	no
23:15:38	evan	not nil.
23:15:41	brixen	k
23:15:42	evan	if it's nil
23:15:45	brixen	it is
23:15:45	drbrain	evan: how does it get from TypedRoot<IOBuffer*> buffer to the IOBuffer*?
23:15:46	evan	then something isn't setting it.
23:15:48	brixen	k
23:15:53	evan	drbrain: .get()
23:15:56	drbrain	stepping through this, it seems to not go into IOBuffer->read_bytes
23:15:57	Defiler	I've never seen it not be nil, myself
23:16:05	Defiler	that I can recall
23:16:47	brixen	Defiler: http://pastie.org/262143
23:16:59	brixen	also, what is msg.module?
23:17:07	evan	brixen: it's filled in by the lookup mechanism
23:17:15	evan	brixen: it's the Module* where the method was found
23:17:22	brixen	hmm
23:17:48	evan	zenspider: http://rafb.net/p/WFVo5143.html
23:17:49	evan	that fine?
23:18:53	rue	msg.method_found_in # ? :)
23:21:27	Defiler	brixen: It is set in executor
23:21:32	Defiler	VMMethod::executor
23:21:48	Defiler	(amongst others)
23:22:02	evan	Defiler: current_self isn't
23:22:07	zenspider	evan: booteeful
23:22:11	evan	zenspider: ok.
23:22:38	Defiler	evan: I was answering him about module
23:22:42	Defiler	not current_self
23:22:56	evan	msg.module isn't set there either
23:22:57	evan	it's read there
23:23:27	evan	zenspider: do you have a better name you refer to the op_asgn2 logic by?
23:23:38	evan	just curious
23:23:47	brixen	evan: what's the idea of TaskProbe and would a probe at instructions.cpp make more sense for now than if 0/1 ?
23:24:11	evan	brixen: yeah, good call
23:24:16	evan	a probe call there would be better
23:24:19	brixen	k, I'll add it
23:24:37	evan	the idea behind the probe is to centralize all that noisy logic
23:24:42	brixen	cool
23:25:03	brixen	should we distinguish them or just all on or all off?
23:25:07	rue	Method overloading might help the cases where multiple different locations manipulate some data structure slightly differently
23:25:15	rue	Assuming those cannot be combined anyway
23:25:24	evan	brixen: dunno, thats why they're centralized, so we can easily change them all :D
23:25:35	brixen	right, but I don't want to turn them all on
23:25:38	brixen	is what I mean
23:25:42	evan	add a flag
23:25:49	evan	to turn just the one you want on
23:25:52	brixen	PROBE=opcode
23:25:54	brixen	something like that?
23:25:55	evan	sure
23:25:56	brixen	k
23:26:07	evan	every method in probe could have a bool flag also
23:26:10	evan	that indicates if it should fire
23:26:20	evan	then just wire up some easy logic to set the flags
23:27:07	brixen	I'm thinking just a set of bit flags
23:27:24	brixen	set with stuff like PROBE=1 all, PROBE=opcode just opcode
23:27:54	brixen	PROBE=1 // turn them all on
23:27:57	brixen	to be more clear
23:28:50	evan	sure, up to you.
23:29:35	evan	brixen: pushed a potential fix for that Hash problem
23:29:37	evan	give it a shot
23:29:43	evan	be sure to kernel:clean before though
23:30:04	brixen	k, thanks
23:32:54	drbrain	should break rubinius::IOBuffer::create stop somewhere?
23:33:00	drbrain	it doesn't seem to be called
23:33:19	drbrain	I'm trying to figure out who's setting used to 0
23:33:32	evan	no
23:33:42	evan	I don't think that we're creating IOBuffers with that
23:33:44	drbrain	so I figured break in IOBuffer::create, continue until it's created the buffer I want, then set a watchpoint
23:33:52	evan	the ones used are created entirely in ruby
23:34:13	evan	drbrain: calling ::create is not automatic
23:34:18	evan	it has to be wired up to prim
23:34:24	evan	which IO::Buffer doesn't have
23:34:32	drbrain	yeah
23:35:01	drbrain	http://rafb.net/p/8PFOCJ51.html
23:35:05	brixen	evan: before it was expected 1 given 0, now it's expected 1 given 3 :)
23:35:23	brixen	if we could convert those to $'s instead
23:35:24	evan	brixen: could you paste me the code?
23:35:27	evan	thats doing the call
23:35:27	drbrain	I added a begin/end p in IO::Buffer#shift_front
23:35:29	evan	the actual code
23:35:35	brixen	sure
23:35:46	drbrain	and it seems that something is wiping out my buffer
23:35:55	evan	drbrain: yeah! eeks.
23:35:59	evan	where did that buffer goe
23:36:29	evan	drbrain: there are only a few things that set @used = 0
23:36:32	evan	should be easy to find
23:36:35	brixen	evan: http://pastie.org/262156
23:36:40	brixen	the call is on 198
23:36:46	brixen	self.size += 1
23:37:06	evan	and could you paste the output again?
23:37:08	brixen	same code as in here: http://pastie.org/262107
23:37:09	evan	so I don't get confused
23:37:22	brixen	I'll have to create it, sec
23:37:28	evan	brixen: yeah, but that code isn't the actual code you're running
23:37:36	evan	i need to see if there is something else thats bleeding through
23:37:38	evan	and screwing it up
23:37:42	brixen	k
23:38:41	brixen	evan: then you want this too: http://pastie.org/262159
23:38:58	brixen	evan: the first is bootstrap/hash, the one just pastied is common/hash
23:39:28	evan	ok
23:39:33	evan	and the output when running it?
23:39:37	brixen	how much of the opcode do you want
23:39:41	brixen	I'm waiting for it
23:39:41	evan	the last 50
23:39:43	brixen	k
23:40:23	brixen	http://pastie.org/262161
23:40:36	brixen	there's a few extra heh
23:41:36	evan	um
23:41:39	seydar	drbrain: what was your problem from last night?
23:41:40	evan	it's not self.size += 1
23:41:41	evan	at all
23:41:58	evan	it's the @bins[bin] = Bucket.new key, nil, key_hash
23:42:03	brixen	yeah, this is different
23:42:09	evan	ok, so we fixed one
23:42:18	brixen	yeah, last was 198
23:42:28	evan	alrighty
23:43:55	evan	brixen: i have to ask
23:44:00	evan	why is Bucket a Tuple subclass?
23:44:18	brixen	at put
23:44:18	zenspider	the walrus said it was
23:44:24	brixen	would be faster than ivar lookup
23:44:32	zenspider	lolrus is a bucket expert, after all
23:44:34	evan	brixen: it's not though
23:44:43	brixen	well, easy to change
23:44:46	evan	ok
23:44:47	evan	just curious
23:44:47	brixen	why isn't it faster?
23:44:58	brixen	the code path for at is alot less it seems to me
23:45:01	evan	back in the day, I measured this
23:45:05	evan	well
23:45:11	evan	you're adding an extra method call in there
23:45:16	evan	that negates any savings
23:45:16	brixen	sure
23:45:22	evan	if there was any
23:45:29	zenspider	we really shouldn't be making design decisions (at this level) based on speed
23:45:53	brixen	I doubt this is the problem though
23:45:56	evan	it's not
23:45:59	evan	i was just curious
23:46:03	evan	alright, back on the trail.
23:46:19	zenspider	wanders off
23:46:44	evan	brixen: oh ha!
23:46:58	evan	that IS the problem.
23:47:07	evan	since it's Tuple subclass
23:47:12	evan	it's calling Tuple.new
23:47:26	evan	which is hooked directly up to tuple_allocate
23:47:46	brixen	ok, easy to fix
23:47:47	brixen	one sec
23:47:52	evan	k
23:48:14	evan	we should have a comment in Tuple that says if you subclass it
23:48:18	evan	you can't have an initialize
23:48:20	evan	because thats true
23:48:39	seydar	evan: could you call super instead?
23:48:44	evan	nope
23:48:50	evan	never gets that far
23:49:04	evan	Tuple.new goes directly to a primitive
23:49:06	evan	no initialize is called
23:49:25	evan	Tuple isn't really made to be subclassed
23:49:30	evan	it's too simple
23:49:33	evan	it's made to be delgated to
23:50:33	brixen	I have specs for this part that I can actually run without kill MRI or shotgun
23:50:59	evan	you're running them under vm?
23:51:26	brixen	yeah, with bin/mspec -tr
23:51:38	seydar	Would making a wrapper class around Tuple kill the speed increase?
23:51:38	evan	er?
23:51:40	evan	thats under MRI
23:51:43	brixen	and loading the bootstrap/hash and lib/core_bridge/hash
23:51:46	evan	i'm confused
23:51:58	evan	"run without kill MRI"
23:52:01	brixen	evan: sorry, in the cpp dir
23:52:01	evan	?
23:52:07	brixen	using MRI
23:52:15	brixen	if I load common/hash, it kills MRI
23:52:19	evan	i'm still confused.
23:52:25	brixen	nvm then
23:52:27	evan	k
23:53:53	evan	brixen: it runs under MRI because core_bridge/tuple.rb is totally wrong
23:54:07	evan	well, not totally
23:54:14	brixen	you don't have mine
23:54:15	evan	but how a Tuple is created doesn't match how Rubinius does it
23:54:17	evan	thus the disconnect
23:54:39	brixen	right, Tuple in MRI is Array and can be subclassed
23:54:50	brixen	anyway, it's a very simple change
23:54:54	brixen	and all the specs pass
23:55:01	brixen	so, I'll try running it under the vm
23:55:03	evan	k
23:55:54	brixen	I think I'll delete core_bridge anyway, it's never really worked out
23:56:03	evan	k
23:56:09	brixen	and it's impossible to really replace Hash in MRI
23:56:30	brixen	but I can test stuff like Hash::Bucket fine