carlosgaldino has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
carlosgaldino has joined #rubinius
dimday has joined #rubinius
elia has joined #rubinius
elia has quit [Ping timeout: 256 seconds]
Bwild has quit [Quit: leaving]
<cremes>
chrisseaton: I’m a huge fan of what’s going on at the jruby project. For my current perf needs (as well as Windows needs), it is really the only game in town.
<cremes>
I’m glad to hear you have these particular problems dialed in and a potential solution in the wings. Looking forward to JRuby 9k + Indy + Truffle + whatever.
<chrisseaton>
cremes: thanks - do you run Windows at the moment? 9000 doesn't build on Windows at the moment, and they need some help getting it working again if you have the time.
<cremes>
I do mostly run jruby on windows. I have some time the first 2 days of this week to lend a hand if you need it.
<cremes>
Time for bed. Leave me messages here/jruby if you need some help.
lopex has quit [Quit: Connection closed for inactivity]
meh` has quit [Ping timeout: 252 seconds]
dimday has quit [Ping timeout: 240 seconds]
diegoviola has quit [Quit: WeeChat 1.0.1]
pwh has joined #rubinius
havenn has joined #rubinius
havenwood has quit [Ping timeout: 250 seconds]
dimday has joined #rubinius
diegoviola has joined #rubinius
chrisseaton has quit [Ping timeout: 272 seconds]
chrisseaton_ has joined #rubinius
diegoviola has quit [Quit: WeeChat 1.0.1]
havenn has quit [Remote host closed the connection]
jeregrine has quit [Ping timeout: 272 seconds]
jeregrine has joined #rubinius
craigp has joined #rubinius
craigp has quit [Client Quit]
jeregrine has quit [Ping timeout: 272 seconds]
dlackty_ has quit [Ping timeout: 272 seconds]
DanielVartanov__ has joined #rubinius
DanielVartanov has joined #rubinius
DanielVartanov__ has quit [Remote host closed the connection]
DanielVartanov has quit [Remote host closed the connection]
craigp has joined #rubinius
noop has joined #rubinius
dlackty_ has joined #rubinius
jbr^ has quit [Ping timeout: 272 seconds]
jeregrine has joined #rubinius
jbr^ has joined #rubinius
craigp has quit [Remote host closed the connection]
craigp has joined #rubinius
<heftig>
brixen: are you just talking about changing require's implementation or about introducing new semantics?
|jemc| has quit [Ping timeout: 258 seconds]
havenwood has joined #rubinius
craigp has quit [Remote host closed the connection]
dzhulk has joined #rubinius
dimday has quit [Remote host closed the connection]
craigp has joined #rubinius
DanielVartanov_ has joined #rubinius
DanielVartanov_ has quit [Remote host closed the connection]
josh-k has joined #rubinius
lopex has joined #rubinius
benlovell has joined #rubinius
<yorickpeterse>
morning
<yorickpeterse>
cremes: the problem in my case is that for Oga I'm not going to write implementation specific code
<yorickpeterse>
cremes: having to do so for MRI, Rbx and JRuby would be far too time consuming
kagaro has quit [Ping timeout: 244 seconds]
kagaro has joined #rubinius
benlovell has quit [Ping timeout: 264 seconds]
benlovell has joined #rubinius
elia has joined #rubinius
goyox86 has joined #rubinius
JohnBat26 has joined #rubinius
elia has quit [Quit: Computer has gone to sleep.]
josh-k has quit [Remote host closed the connection]
elia has joined #rubinius
<goyox86>
morinig!
<goyox86>
morning!
elia has quit [Quit: Computer has gone to sleep.]
elia has joined #rubinius
havenwood has quit [Remote host closed the connection]
yxhuvud has quit [Ping timeout: 256 seconds]
elia has quit [Quit: Computer has gone to sleep.]
benlovell has quit [Ping timeout: 265 seconds]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 255 seconds]
diegoviola has joined #rubinius
craigp has quit [Remote host closed the connection]
craigp has joined #rubinius
benlovell has joined #rubinius
yxhuvud has joined #rubinius
craigp has quit [Remote host closed the connection]
craigp has joined #rubinius
johnmuhl has joined #rubinius
craigp has quit [Ping timeout: 272 seconds]
craigp has joined #rubinius
benlovell has quit [Ping timeout: 264 seconds]
enebo has joined #rubinius
benlovell has joined #rubinius
noop has quit [Ping timeout: 260 seconds]
noop has joined #rubinius
enriclluelles has joined #rubinius
noop has quit [Ping timeout: 258 seconds]
tenderlove has joined #rubinius
craigp has quit [Remote host closed the connection]
benlovell has quit [Ping timeout: 240 seconds]
benlovell has joined #rubinius
havenwood has joined #rubinius
benlovell has quit [Ping timeout: 244 seconds]
bennyklotz has quit [Quit: leaving]
benlovell has joined #rubinius
elia has joined #rubinius
noop has joined #rubinius
GitHub114 has joined #rubinius
<GitHub114>
[rubinius] brixen pushed 3 new commits to master: http://git.io/n2MPLw
<GitHub114>
rubinius/master b9f9c50 Brian Shirai: Removed whitespace in #require specs.
<GitHub114>
rubinius/master 6b79eaa Brian Shirai: Removed ruby_bug from #require specs.
GitHub114 has left #rubinius [#rubinius]
<GitHub114>
rubinius/master 27d4f07 Brian Shirai: Removed with_feature :require_19 from #require specs.
<goyox86>
Mmm let me check I was just running rubinius/benchmark But I had altered the GC settings to see the graph change
<brixen>
ah ok
<brixen>
goyox86: did you write up how you put all this together?
havenwood has quit [Remote host closed the connection]
<goyox86>
Nope in fact I was about to ask you if the blog post offer was still there. I'm gonna write about getting this up and running with StatsD + InfluxDB-StatsD plugin
<goyox86>
But I'll write it I was just checking if I could go to influxdb with statSD
havenwood has joined #rubinius
<goyox86>
brixen How do you plan to implement other runtime information such as for example the thread list which their status, classes loaded, and so on?
<brixen>
goyox86: yes, definitely would love for you to write up a post
dzhulk has joined #rubinius
<brixen>
goyox86: check out draftin.com if you don't have a preferred collaborative editing tool
<brixen>
it's pretty sweet
* goyox86
checking
<brixen>
goyox86: as for other runtime info, it depends
<brixen>
I'm trying to keep a clear separation between logging, monitoring, and analysis
<goyox86>
brixen Roger. Testing the draftin.com is really clean the UI
<brixen>
it's pretty nice
<brixen>
gives you a diff of collaborator edits and has a nice commenting feature
benlovell has quit [Ping timeout: 258 seconds]
diegoviola has quit [Remote host closed the connection]
<yorickpeterse>
I started the evening continueing my studies into LL parsing
<headius>
hooray
<yorickpeterse>
I'm now reading on mania, schizophrenia and string theory
meh` has joined #rubinius
<headius>
so parsing has driven you to crazy
<yxhuvud>
wouldn't be the first.
<yorickpeterse>
headius: the other way around
<yorickpeterse>
actually this originated from me theorizing about the afterlife and parallel universes while having dinner
<headius>
friends don't let friends LALR
<yorickpeterse>
I recalled some theory covered parts of this, hence I started googlging
<yorickpeterse>
* googling
<yorickpeterse>
I tend to get lost in these kind of thoughts :P
<headius>
yeah, I call it birdwalking after a behavior of the same name my professors often exhibited...wandering aimlessly from topic to topic
<yorickpeterse>
Oh, heh, we call that "ijsberen"
<yorickpeterse>
though that involves physical motion of walking in circles
<headius>
nice
<yorickpeterse>
it basically translates to "polarbearing"
<enebo>
yorickpeterse: if you do LALR(1) impl please share that code. I will convert to Java…but comment ths shit out of it too :)
<yorickpeterse>
heh, most likely I'll do LL
<yorickpeterse>
since LALR is a whole other bucket of wtf
<yorickpeterse>
I'm slowly getting the hang of LL
<yorickpeterse>
(that only took over a week)
<enebo>
hah well if the generated table is small enough then I might not care
<yorickpeterse>
In unrelated news, fascinating that hallucinations can be witness by using PET/MRI scans
<yorickpeterse>
and that apparently the brain actually takes different shapes/forms (more or less)
<yorickpeterse>
* witnessed
<yorickpeterse>
"The premier cause of auditory hallucinations in the case of psychotic patients is schizophrenia. In those cases, patients show a consistent increase in activity of the thalamic and strietal subcortical nuclei, hypothalamus, and paralimbic regions; confirmed via PET scan and fMRI.[11][12] Other research shows an enlargement of temporal white matter, frontal gray matter, and temporal gray matter volumes
<yorickpeterse>
(those areas crucial to both inner and outer speech) when compared to control patients."
* yorickpeterse
finds this really interesting, even though it's totally not related to Ruby/Rubinius :P
<yorickpeterse>
"High caffeine consumption has been linked to an increase in the likelihood of experiencing auditory hallucinations. A study conducted by the La Trobe University School of Psychological Sciences revealed that as few as five cups of coffee a day could trigger the phenomenon."
<headius>
huh, interesting
<yorickpeterse>
That certainly explains some of the weird shit in computing
<headius>
what auditory hallucinations are they talking about?
<yorickpeterse>
That's fascinating, it could mean that not only can you wipe memory but potentially _write_ memory to one's brain
<yorickpeterse>
by just plugging a power cord up their butt
<yorickpeterse>
(I'm simplifying things here obviously)
diegoviola has quit [Ping timeout: 264 seconds]
<yorickpeterse>
I eagerly await the day I can run Ruby on my brain
<|jemc|>
yorickpeterse: did you ever get to using any kind of "lazy" string buffers in oga?
<|jemc|>
as in, an object to wrap up a start and end index number as well as a reference to the original string?
<yorickpeterse>
|jemc|: No, not yet
<yorickpeterse>
and I doubt it would be of any benefits in RUby
<yorickpeterse>
* Ruby
<yorickpeterse>
Since the allocation of a custom class with some index is not going to be very different from allocating a string
<yorickpeterse>
I certainly don't see it being any faster/less intensive
<|jemc|>
yorickpeterse: yeah, I was about to go down that road and that's what I was thinking - my class isn't going to exceed the performance of a C++-based String
<yorickpeterse>
_if_ Ruby had a concept of "code point strings" this might be a different story
<brixen>
it could be more space efficient, of course
<yorickpeterse>
(that is, Unicode code points)
<yorickpeterse>
Rbx has byte strings, but they suck for multi-bytes
<yorickpeterse>
As in, it's hard to construct a string based on just bytes as you have to guarantee all bytes are there
<brixen>
they are just bytes :P
<yorickpeterse>
which makes less sense to explain when writing it down
<brixen>
String is what gives an interpretation of the bytes in a ByteArray
<yorickpeterse>
on second though, byte strings might work too
<yorickpeterse>
* thought
<brixen>
given a single large document string, you should be able to create a tree indexing it quite easily
<|jemc|>
yeah, I'm in the midst of trying to see if there's anywhere I can optimize the bytecode I "stole" from pegarus
<brixen>
this is, eg, what a suffix tree would do
<yorickpeterse>
brixen: this is problematic when the input string is streamed, instead of read as a whole
<brixen>
why?
<brixen>
how does streaming impact this?
<yorickpeterse>
well, it's problematic too for any form where you keep the input around
<yorickpeterse>
That is, in order to apply the index you need the raw input buffer to slice on
<yorickpeterse>
That means keeping it around in memory somewhere
<yorickpeterse>
Which in turn makes streaming useless since the input is all in memory
<yorickpeterse>
This _would_ work if the input was read as bytes, not a string
<|jemc|>
yeah, you'd have to keep it around, but you could at least "stream" into it
<|jemc|>
as in, having an "incomplete" string when you started - with no idea where the tail is until it's done
<yorickpeterse>
In my case that would mean having to keep `const char*` pointers around somewhere until they are used. I'd rather eat a chainsaw than deal with that
<yorickpeterse>
errrr `const char` pointers
<yorickpeterse>
not const char pointer pointers
<yorickpeterse>
so tl;dr for webscale parsing you'd need 1) IO#readbyte 2) ByteString
<brixen>
yorickpeterse: streaming does not dictate a single buffer
<brixen>
yorickpeterse: and the tree can retain the reference to the chunk
<yorickpeterse>
This is possible in Rbx, thus you _could_ probably make a pure Ruby Oga for Rbx that doesn't suck memory wise
<brixen>
@chunk, @start, @size or something like that
<yorickpeterse>
brixen: correct, but that means you have to keep the chunk around next to the index
<yorickpeterse>
It would certainly remove some of the String allocation overhead (until you need it), but it's not going to help that much memory wise
<yorickpeterse>
Most of my problems come from using C/Java, funny how that goes :P
<yorickpeterse>
Now that I remember, the byte slicing part is what I did prior to porting the lexer to C/Java
<yorickpeterse>
Although I think I still read the input as a String, then bytesliced on it
<yorickpeterse>
s/on it/it
havenwood has quit [Remote host closed the connection]
<yorickpeterse>
darn it, I should stop reading Wikipedia, it really impares my writing ability
dzhulk has quit [Quit: Leaving.]
<yorickpeterse>
Actually, I remember I did solve (mostly) the memory problem in pure Ruby, it was just the time it took to parse things that sucked
<yorickpeterse>
clearly we need asm.rb
<yorickpeterse>
Funny how Rbx already has that, more or less :D
josh-k has joined #rubinius
diegoviola has joined #rubinius
slaught has quit [Quit: slaught]
<|jemc|>
ah, and this is a great opportunity to use the new goto_if_equal instruction :)