benlovell has joined #rubinius
benlovell has quit [Ping timeout: 272 seconds]
<bennyklotz> @yorickpeterse: do you have a gh repo for your parser, so I can watch :D
<brixen> wow, so apparently autoload resolution in the C-API has been busted since 1c079a8f
<brixen> afaics
<brixen> which is jul 1, 2012
* brixen writes more C-API specs
<brixen> bundle --jobs=4 works
<jc00ke> w00t!
<jc00ke> w0000000000000000000000000000000t!
* jc00ke builds HEAD, gonna test
<jc00ke> Well, I suppose I have to wait until you push ;)
<brixen> uh
<brixen> lol
havenn has quit [Remote host closed the connection]
<brixen> I was about to ping you and figured I'd wait till the specs finish and I pushed
<brixen> the hilarity of the error that bundler actually output
<brixen> this is the real backtrace https://gist.github.com/brixen/a5971cc930abd2ace163
<brixen> wrong argument type Autoload (expected Module) being the actual error
pietr0 has quit [Quit: pietr0]
<jc00ke> no shit?! Damn.
<brixen> apparently, the prevalence of duck typing in Ruby means any output I show you for an exception is valid
<brixen> you don't need that backtrace for anything useful
<diegoviola> might this be the cause of Ezra passing away? https://twitter.com/ezmobius/status/362228490176692224
<diegoviola> this makes me very sad :'(
<brixen> alternatively, Rails has ruined everything
<brixen> diegoviola: yeah, he was doing lots better last time I saw him
<diegoviola> "I'm also not jumping back in too quickly as the stress had agravated my recently diagnosed type 2 adult onset diabetes which is what landed me in the hospital." -- Ezra Zygmuntowicz
<diegoviola> how can type 2 diabetes kill someone after only 1 year of being diagnosed?
<diegoviola> this is insane
<brixen> but with the Stuffstr work, he may have gotten stressed and down a bad road again
<diegoviola> this is scaring me, seriously
* diegoviola freaks out
<brixen> our diet is killing us
josh-k has joined #rubinius
<brixen> no question about that
<brixen> interesting docu if you want to watch something: http://www.fatsickandnearlydead.com/
<brixen> perhaps more worrisome, the so-called "metabolic syndrome" isn't associated with obesity
<evan_> one good thing about living in LA, it's easy to get real food.
evan_ is now known as evan
<diegoviola> is this the reason?
<diegoviola> please tell me it isn't
<evan> we don't know.
<diegoviola> ok
<brixen> metabolic syndrome is like cirrhosis of the liver without the drinking
<brixen> yeah, we don't know about ez
GitHub60 has joined #rubinius
<GitHub60> [rubinius] brixen pushed 3 new commits to master: http://git.io/euXkxA
<GitHub60> rubinius/master 5be264f Brian Shirai: Added C-API spec for autoload resolution from rb_class_to_path, rb_class2path.
<GitHub60> rubinius/master 36a5900 Brian Shirai: Fixed horrible abuse of Ruby syntax.
<GitHub60> rubinius/master 7772d05 Brian Shirai: Handle autoload in rb_path2class, rb_path_to_class.
GitHub60 has left #rubinius [#rubinius]
<brixen> jc00ke: please test your repro repo with master
<brixen> jc00ke: works for me :)
<jc00ke> Will test now
<diegoviola> our human bodies are too fragile
<brixen> jc00ke: gonna be on the train, will check in later
<diegoviola> anything can kill us :(
<diegoviola> *sigh*
<evan> we are beasts of a former age
<brixen> smiling and laughing generally hasn't been know to be lethal
<brixen> so we can do more of that
<evan> when cold winters killed most people
<brixen> s/know/known/
<evan> and fat stores were a survival trait
<evan> I know I live in an LA food bubble, but it feels like things are getting better on the food quality front.
guilleiguaran has quit [Ping timeout: 272 seconds]
andrewstewart has quit [Ping timeout: 272 seconds]
cpuguy83 has quit [Ping timeout: 272 seconds]
machty has quit [Read error: Connection reset by peer]
cezarsa has quit [Ping timeout: 272 seconds]
machty has joined #rubinius
andrewstewart has joined #rubinius
guilleiguaran has joined #rubinius
<jc00ke> @brixen hmm, still getting the Bundler::GemspecErrors, but lemme clear my cache and try again
cpuguy83 has joined #rubinius
cezarsa has joined #rubinius
<jc00ke> Hmm, yeah, maybe I'm not cleaning everything up correctly?
<jc00ke> `rm ~/.gem/rbx/2.1.0/specifications/* ~/.gem/rbx/2.1.0/cache/* ~/.gem/rbx/2.1.0/gems/*`
<jc00ke> `bundle --jobs=4`
dimday has joined #rubinius
machty has quit [Ping timeout: 272 seconds]
machty has joined #rubinius
alexsuraci_ has quit [Ping timeout: 258 seconds]
alexsuraci_ has joined #rubinius
johnmuhl has quit [Read error: Connection reset by peer]
johnmuhl has joined #rubinius
josh-k_ has joined #rubinius
josh-k has quit [Ping timeout: 264 seconds]
houhoulis has joined #rubinius
GitHub188 has joined #rubinius
<GitHub188> [rubinius] brixen pushed 2 new commits to master: http://git.io/xjakwg
<GitHub188> rubinius/master 6a86a57 Brian Shirai: Added C-API specs for more autoload resolution.
<GitHub188> rubinius/master 738976f Brian Shirai: Fixed autoload resolution in C-API.
GitHub188 has left #rubinius [#rubinius]
<brixen> jc00ke: there's surely some race there, otherwise that autoload issue wouldn't get hit with --jobs=4
<brixen> interesting I don't see it after fixing that
<jc00ke> hmm
<jc00ke> How are you clearing your local cache, and how are you running bundler?
<brixen> what do you mean by cache?
<brixen> ~/.gem/rbx/2.1.0/cache?
<jc00ke> Yeah
<brixen> by rm'ing that
<brixen> and bundle install --jobs=4
<jc00ke> And any other files in gems/ specifications/ etc?
<jc00ke> OK, just checking. I'm clearing out all 3 of those dirs
<brixen> I rm'd the whole parent dir
<cpuguy83> So is there something wrong with 2.3.0, what’s with the version bump?
<brixen> cpuguy83: what do you mean by "wrong"?
<brixen> 2.4.0 is a superset of 2.3.0
<brixen> so the minor version bumps
<cpuguy83> broken/bugged in some way
<cpuguy83> ie, should people _not_ use it.
<brixen> did you read Changelog and News?
<brixen> you should always use the latest rbx release
<cpuguy83> I did not check the changelog
<brixen> unless you have a breaking change that forces you to use a previous one
<brixen> rbx isn't MRI ;)
<brixen> if you are on a month old rbx, you are missing out
<brixen> we expect people to update to every new release
<brixen> anyway, dinner
benlovell has joined #rubinius
<jc00ke> @brixen I'll give the bundler thing another go in a bit
benlovell has quit [Ping timeout: 264 seconds]
travis-ci has joined #rubinius
<travis-ci> rubinius/rubinius/master (7772d05 - Brian Shirai): http://travis-ci.org/rubinius/rubinius/builds/42690324: The build passed.
travis-ci has left #rubinius [#rubinius]
<postmodern> thinking out loud here
<postmodern> but it seems like rubygems is the wrong tool to manage rubysl
<postmodern> since rubysl is patterned off the standard-library, rubysl libs don't need a version
<postmodern> the only version you care about is the ruby engine version, or rbx version
<brixen> not true at all
<postmodern> i almost think it would be a worth while project for rubinius to create a lower-level package manager
<brixen> what version is psych, minitest, bigdecimal, etc
<brixen> Rubinius X will
<brixen> Rubinius doesn't need to
<postmodern> brixen, when you care about the version, you use rubygems
<postmodern> hear me out
<postmodern> what if there was a lower-level package manager that essentially managed $LOAD_PATH
<brixen> like bundle --standalone?
<postmodern> basically
<postmodern> or like the `gem` method
<brixen> that exists
<postmodern> or python's import
<brixen> the gem method is pure evil
<postmodern> where you could enable/disable rubysl parts
<brixen> if gem didn't exist, rubygems could probably be not runtime right now
<brixen> postmodern: what problem are you trying to solve?
<postmodern> brixen, using stdlib without using rubygems
<brixen> why?
<postmodern> why not?
<brixen> that's not an answer
<postmodern> also i think it would be useful to have a simpler version of rubygems
<brixen> why are you trying to avoid rubygems?
<postmodern> one that you could build off of
<postmodern> as eventually we will need to rewrite rubygems at some point
<brixen> I think it would be excellent to have a completely different implementation of rubygems, for sure
<brixen> that's still not an answer
<postmodern> just thinking out loud
<brixen> ok, but I'm still trying to understand your problem
<brixen> pointless to talk about solutions without understanding the problem
<brixen> what about rubygems do you want to avoid and why?
<postmodern> im not talking about a specific problem
<postmodern> just an impudence mismatch i noticed with rubygems and stdlib
<brixen> then what are you solving?
<brixen> stdlib is a terrible idea
<brixen> stdlib should never have existed
<brixen> it's the worst Ruby code in existence
<brixen> precisely because it was tied to MRI releases
<brixen> if there had been anything close to rubygems, and core libs had been able to connect to a socket, stdlib would never have existed
<postmodern> well that and it was mostly written in 1997
<postmodern> but the idea of stdlib is commonly accepted by majority of programming languages
<brixen> who cares?
<brixen> commonly accepted should never be an argument for anything
<postmodern> I'm just stating a fact, not asking you to agree with it
<brixen> positing such an argument just makes me blindly rage
<postmodern> ok if you're going to be hostile, i'll drop the subject
<brixen> you're not stating a fact in a void, you're making an argument
<brixen> an utterly false one, but one made as if it had merit
<brixen> I'm not being hostile to you
<brixen> but you're reliance on "commonly accepted" is bogus
<brixen> s/you're/your/
<brixen> it is true that rubygems is a mess
<brixen> I have been beating that drum for well over a year
postmodern has left #rubinius ["Leaving"]
<brixen> I know exactly what needs to be fixed
<brixen> I don't have time to write it, promote it, deal with all the stupid bullshit bundler dealt with, etc
<brixen> at the same time, I won't hesitate to challenge any assertion that Ruby needs a stdlib
travis-ci has joined #rubinius
<travis-ci> rubinius/rubinius/master (738976f - Brian Shirai): http://travis-ci.org/rubinius/rubinius/builds/42694194: The build has errored.
travis-ci has left #rubinius [#rubinius]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 258 seconds]
sferik_ has joined #rubinius
sferik has quit [Read error: Connection reset by peer]
havenwood has joined #rubinius
josh-k_ has quit [Remote host closed the connection]
josh-k has joined #rubinius
johnmuhl has quit [Quit: Connection closed for inactivity]
josh-k has quit [Ping timeout: 250 seconds]
meh` has quit [Ping timeout: 258 seconds]
|jemc| has joined #rubinius
diegoviola has quit [Quit: WeeChat 1.0.1]
digitalextremist has joined #rubinius
digitalextremist has quit [Client Quit]
jnh has quit [Quit: Leaving...]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 244 seconds]
|jemc| has joined #rubinius
dzhulk has joined #rubinius
houhoulis has quit [Remote host closed the connection]
jfredett-w1 has joined #rubinius
havenwood has quit [Remote host closed the connection]
jfredett-w has quit [Ping timeout: 240 seconds]
|jemc| has quit [Ping timeout: 250 seconds]
slaught has quit [Quit: slaught]
JohnBat26 has joined #rubinius
tenderlove has quit [Remote host closed the connection]
tenderlove has joined #rubinius
tenderlove has quit [Remote host closed the connection]
tenderlove has joined #rubinius
tenderlove has quit [Remote host closed the connection]
tenderlo_ has joined #rubinius
|jemc| has joined #rubinius
tenderlo_ has quit [Remote host closed the connection]
benlovell has joined #rubinius
noop has joined #rubinius
dzhulk has quit [Quit: Leaving.]
benlovell has quit [Ping timeout: 264 seconds]
|jemc| has quit [Ping timeout: 256 seconds]
dzhulk has joined #rubinius
tenderlove has joined #rubinius
Ori_P has joined #rubinius
dimday has quit [Quit: Leaving.]
tenderlove has quit [Ping timeout: 245 seconds]
sferik_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dzhulk has quit [Ping timeout: 250 seconds]
dzhulk has joined #rubinius
Ori_P has quit [Quit: Computer has gone to sleep.]
Ori_P has joined #rubinius
goyox86 has joined #rubinius
Ori_P has quit [Ping timeout: 272 seconds]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 250 seconds]
goyox86 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<yorickpeterse> morning
heftig_ is now known as heftig
jfredett-w has joined #rubinius
goyox86 has joined #rubinius
jfredett-w1 has quit [Ping timeout: 252 seconds]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 240 seconds]
gtemple has quit [Read error: Connection reset by peer]
gtemple has joined #rubinius
elia has joined #rubinius
josh-k has joined #rubinius
<goyox86> morning!
<goyox86> Hey yorickpeterse which book did you decided to read?
<goyox86> about parsers and the likes
Ori_P has joined #rubinius
Ori_P has quit [Client Quit]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 258 seconds]
<yorickpeterse> goyox86: none so far
<goyox86> yorickpeterse Ah I remember seeing you discussing here something about the dragon book
<yorickpeterse> Yeah, I bought http://computationbook.com/ but that's about it
<chrisseaton> yorickpeterse: I have that on my to-read list - it looks really intereseting
<goyox86> I bought that one, (never read it thougu)
<goyox86> though
<chrisseaton> yeah, I have a lot of books like that
josh-k_ has joined #rubinius
josh-k has quit [Ping timeout: 250 seconds]
<goyox86> Man, I cant get rbx sending metrcis to statsd xD
<goyox86> I have statsd up. And I'm firing up my rails app like: RBXOPT=-Xsystem.metrics.target="statsd"-Xsystem.metrics.statsd.server="127.0.0.1:8125" bundle exec rails s
benlovell has joined #rubinius
goyox86 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
benlovell has quit [Ping timeout: 252 seconds]
pwh has quit [Ping timeout: 255 seconds]
pwh has joined #rubinius
tenderlove has joined #rubinius
sferik has joined #rubinius
tenderlove has quit [Ping timeout: 250 seconds]
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
josh-k has joined #rubinius
josh-k_ has quit [Read error: Connection reset by peer]
flavio has joined #rubinius
benlovell has joined #rubinius
sferik has joined #rubinius
benlovell has quit [Ping timeout: 245 seconds]
sferik has quit [Client Quit]
havenwood has joined #rubinius
sferik has joined #rubinius
sferik has quit [Client Quit]
sferik has joined #rubinius
goyox86 has joined #rubinius
johnmuhl has joined #rubinius
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sferik has joined #rubinius
sferik has quit [Client Quit]
sferik has joined #rubinius
goyox86 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
noop has quit [Ping timeout: 250 seconds]
sbryant has quit [Ping timeout: 258 seconds]
s3m has joined #rubinius
sbryant has joined #rubinius
tenderlove has joined #rubinius
<headius> dbussink, yorickpeterse: would either of you be interested in talking about Rubinius at FOSDEM? At the moment there's no rbx representation, and we don't like that
<yorickpeterse> headius: if I knew of a topic, yes
<headius> "Rubinius"
<yorickpeterse> That's like asking somebody to talk about the meaning of life :P
<headius> wow, you really are a fan
<yorickpeterse> eh?
<headius> bad joke...rubinius == meaning of life
<Rotonen> does rubinius have 42 of anything you could use as a topic? :P
<yorickpeterse> oh no, I meant it as in the ambiguity
<Rotonen> "42 ways to JIT"? :-)
<yorickpeterse> headius: wasn't the room something about language development specific?
<headius> anyway, we wanted to keep this from becoming a "look at my great new web framework" and put a focus on implementations etc
<yorickpeterse> Rotonen: from now one we won't show errors, only "42"
<yorickpeterse> So the annoying thing here is that brixen is doing the juicy stuff/toying with them, I'm just the janitor
<yorickpeterse> lets see when FOSDEM was again
s3m has quit [Ping timeout: 258 seconds]
<yorickpeterse> Hm 31st jan
<headius> brixen already said he wouldn't be able to make it, so I'm looking for EU-local folks
<headius> besides, you're basically eurobrixen now
<headius> for some reason I always hear trance whenever someone prefixes a word with euro-
<yorickpeterse> haha
<headius> nnn ts nnn ts nnn ts
yorickpeterse is now known as eurobrixen
<eurobrixen> There
<headius> perfect
<headius> whether you think you're qualified or not, it would be a shame for there to be no rbx representation at Ruby's trumphant return to FOSDEM
<eurobrixen> Yeah I'm actually quite interested in doing it
<eurobrixen> Is there any sort of experience level?
<eurobrixen> That is, is it expected for only hardcore people to show up, or should the topic be more accessible to everybody?
<headius> we figured making it a bit more hardcore would make it interesting to more folks outside Ruby
<headius> "50 new Ruby ORMs you haven't heard of" doesn't really draw them in
<eurobrixen> That's true
<eurobrixen> oh darn, deadline was yesterday already?
<headius> oh, it was, but we extended it to 12th
<eurobrixen> Ah right
<headius> mostly to get the last few folks we want in
<eurobrixen> heh
<mpapis> headius, not enough to fill in time?
<headius> I'm not sure how many we have right now...did you submit anything?
<eurobrixen> Hm, I think I might have some topics
<eurobrixen> headius: I'll do some talking with usbrixen and I'll see what I can cook up with
<mpapis> headius, I did RVM2 ... no description jet tho
<headius> eurobrixen: sure, feel free to bounce any ideas off me...I'm on the selection committee
<headius> mpapis: ok good
sferik has quit [Ping timeout: 250 seconds]
<eurobrixen> headius: will do
benlovell has joined #rubinius
sferik has joined #rubinius
benlovell has quit [Ping timeout: 250 seconds]
goyox86 has joined #rubinius
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
flavio has quit [Quit: WeeChat 1.0]
flavio has joined #rubinius
flavio has quit [Changing host]
flavio has joined #rubinius
tenderlove has quit [Remote host closed the connection]
tenderlove has joined #rubinius
sferik has joined #rubinius
digitalextremist has joined #rubinius
digitalextremist has quit [Client Quit]
tenderlove has quit [Remote host closed the connection]
<cpuguy83> Here’s the Dockerfile I’ve created for rbx, was going to go ahead and submit this as an “official” docker image if everyone is good with that. or if someone else wants to maintain that’d be cool tool. https://github.com/cpuguy83/docker-rbx/blob/master/Dockerfile
<brixen> cpuguy83: sweet
<brixen> cpuguy83: what's needed for maintenance?
<goyox86> cpuguy83 Nice!
<cpuguy83> Just need to make sure it’s up to date.
<brixen> we could certainly maintain it in the rbx repo, but we'd like to help push those responsibilities to other projects
<brixen> what would work better at this point: 1. a PR when we release, or 2. a ping to a maintainer that a release is out?
<cpuguy83> ie, when a new version comes through, update the RBX_VERSION var in the dockerfile, submit a PR to the official image manifest to update the commit hash used to build from
<brixen> ok
<eurobrixen> goyox86: did you look into using a Dockerfile for statsd?
<brixen> I'm going to post about helping existing rbx pkg maintainers and soliciting new ones, so I can include this
<eurobrixen> I somehow got it working on my laptop, but I don't really remember how
eurobrixen is now known as yorickpeterse
<brixen> goyox86: re statsd, can you telnet to 127.0.0.0:8126 and see the stats?
<cpuguy83> brixen: Would be totally awesome to have it within the main rbx repo. Then new tag == bump version in Dockerfile and submit PR to docker-library/official-images.
<brixen> goyox86: in other words, did you verify that your statsd instance is running by eg inserting a metric with curl and then checking it?
<brixen> cpuguy83: would we need to bump that in the release commit?
<brixen> cpuguy83: bbiab...
<goyox86> I've googled smoe days ago but I have statsd working on my machine sending data to influxdb with this -> https://github.com/bernd/statsd-influxdb-backend
<goyox86> Brixen havent verified with curl will do that :p
<cpuguy83> brixen: So docker-library/official-images has a manifest of all official images, which points to a git repo + commit hash to build from. So yeah, once the Dockerfile version is bumped, that would be the commit has to use.
<cpuguy83> *hash
<cpuguy83> brixen: So I’ll hold off submitting, then?
tenderlove has joined #rubinius
tenderlove has quit [Remote host closed the connection]
Ori_P has joined #rubinius
tenderlove has joined #rubinius
tenderlove has quit [Remote host closed the connection]
meh` has joined #rubinius
tenderlove has joined #rubinius
tenderlove has quit [Remote host closed the connection]
flavio has quit [Quit: WeeChat 1.0]
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sferik has joined #rubinius
diegoviola has joined #rubinius
tenderlove has joined #rubinius
Ori_P has quit [Quit: Computer has gone to sleep.]
<brixen> cpuguy83: to clarify, does the docker version bump need to exist *in* the tag
<brixen> eg, would I have had to bump docker version to 2.4.0 in a commit on or before the v2.4.0 tag?
<cpuguy83> brixen: No
<cpuguy83> Nope.
sferik has quit [Read error: Connection reset by peer]
<brixen> ok
<brixen> so there just needs to be a commit with the docker version bump, and that commit is used?
<brixen> used for the build, I mean
<brixen> the reason this is important for me to understand is that eg debian expects the files to be updated in/before the tag
<brixen> so I want to make sure I understand
<cpuguy83> brixen: No, our builder just reads the Dockerfile. So right now I’m pulling from releases.rubini.us… if it were in the main repo, I’d probably change that to use the git repo and checkout the tag version. In which case it wouldn’t work to update the Dockerfile before the tag is created (or rather, before the commit used for the tag)
sferik has joined #rubinius
sferik has quit [Ping timeout: 250 seconds]
pietr0 has joined #rubinius
<brixen> cpuguy83: sorry, you lost me on the last part there
<brixen> which is better for docker, building from the git repo or building from a tarball?
<brixen> I'd expect tarball because I bundle all the right gem versions, etc
<cpuguy83> Either way, really. With git, someone could easily take the Dockerfile and replace “v.2.4.0” with “master” and get everything built out for them.
<goyox86> brixen confirmed that statsd is working on my box with: echo "foo:1|c" | nc -w 1 -u 0.0.0.0 8125
<goyox86> I can see the data in influxdb
<brixen> cpuguy83: yeah, some people object to the build process doing network traffic, which makes the tarball nice
<brixen> cpuguy83: how about you just pick a process and we can revise it if needed
<brixen> I'd advocate for building from the tarball
<cpuguy83> works for me.
<brixen> goyox86: ok, good
<brixen> goyox86: so, check that you're setting the config values correctly
<brixen> goyox86: for rbx
noop has joined #rubinius
<brixen> goyox86: run it directly with rbx -Xsystem.metrics.target=statsd -Xsystem.metrics.statsd.server=127.0.0.0:8125 -G -S rails s
<brixen> goyox86: probably something like that
<brixen> -G makes bundle exec unnecessary
<brixen> -S rails will find the rails gem bin wrapper directly
<brixen> cpuguy83: ok, so what do you need from me at this point?
benlovell has joined #rubinius
sferik has joined #rubinius
<goyox86> brixen how do I see the logs coming from that RBX thread sending the metrics?
benlovell has quit [Ping timeout: 250 seconds]
<goyox86> OFF-TOPIC I'm using RBX to build my RBX and sometimes the process gets hung
<cpuguy83> brixen: Also a set of docs which will be used to populate the DockerHub page for rubinius: https://github.com/cpuguy83/docs/tree/add_rubinius/rubinius
<cpuguy83> This should be fairly static.
<cpuguy83> And lives within the docker-library/docs repo.
<cpuguy83> Here’s what that looks like for jruby: https://registry.hub.docker.com/_/jruby/
sferik has quit [Read error: Connection reset by peer]
noop has quit [Quit: Leaving]
pietr0 has quit [Quit: pietr0]
dzhulk has quit [Quit: Leaving.]
pietr0 has joined #rubinius
<brixen> cpuguy83: ok
<brixen> goyox86: there's no logs about sending the metrics, the thread just sends them
<goyox86> brixen Roger that I was talking about the some error messages I remember seeing in tha code
<brixen> goyox86: dunno what you mean by that
larstobi has quit [Ping timeout: 272 seconds]
<cpuguy83> brixen: And what I have in that repo is enough to give the same output as the jruby stuff.
<cpuguy83> But with rubinius info
goyox86 has quit [Ping timeout: 240 seconds]
elia has quit [Quit: Computer has gone to sleep.]
larstobi has joined #rubinius
slaught has joined #rubinius
havenwood has quit []
slaught has quit [Quit: slaught]
tenderlove has quit [Remote host closed the connection]
tenderlove has joined #rubinius
elia has joined #rubinius
dzhulk has joined #rubinius
elia has quit [Quit: Computer has gone to sleep.]
txus has joined #rubinius
txus has quit [Client Quit]
benlovell has joined #rubinius
benlovell has quit [Ping timeout: 272 seconds]
<cpuguy83> I noticed I have to set my locale to C.UTF-8 for rbx’s vendored bundler to work: invalid gem: package is corrupt, exception while verifying: invalid byte sequence in US-ASCII (ArgumentError) in /usr/src/rbx/rubinius-2.4.0/vendor/cache/bundler-1.7.7.gem (Gem::InstallError)
<headius> cpuguy83: what's your local normally?
<headius> locale
<cpuguy83> Just the default “POSIX"
<cpuguy83> Works installing mri and “gem install bundler"
<headius> johnmuhl: you can open an issue at any time, btw
<headius> oops, wrong channel
<johnmuhl> headius: I will
<cpuguy83> Well, LANG, LANGUAGE, and LC_ALL are unset
<headius> cpuguy83: LC=POSIX?
<headius> oh, ok
<headius> that's weird, isn't it?
<cpuguy83> It’s just default for a container. I only bring it up because bundler installs fine not using the vendored version.
<headius> interesting
slaught has joined #rubinius
pietr0 has quit [Quit: pietr0]
pietr0 has joined #rubinius
josh-k has quit [Remote host closed the connection]
josh-k has joined #rubinius
josh-k has quit [Ping timeout: 264 seconds]
<yorickpeterse> euh what the hell FOSDEM
<yorickpeterse> Activated my account but it doesn't let me log in
<yorickpeterse> perhaps my password is too secure
* yorickpeterse already went from 128 to 64 characters
* yorickpeterse is now using 24 :<
<yorickpeterse> oh FFS
benlovell has joined #rubinius
<yorickpeterse> riiiiight, apparently 24 is also too much
elia has joined #rubinius
<headius> hah
<headius> nice
<yorickpeterse> ....but I can't set it to a randomly generated string of 16 characters
<yorickpeterse> what the fuck is this
<yorickpeterse> at least I can infinitely reset my password
<yorickpeterse> right, apparently it takes it when you _do_ include a number
<yorickpeterse> RIIIIIIGHT
<yorickpeterse> It's apparently not length limited, it just requires a number
<yorickpeterse> what's a validation message?
benlovell has quit [Ping timeout: 244 seconds]
<yorickpeterse> well at least I'm in
<headius> woohoo
<headius> yorickpeterse: thank you
<brixen> oh the day I never ever ever ever have to deal with psych or nokogiri again
<yorickpeterse> That day might come soon, at least partially
<yorickpeterse> First I need to either fix Racc or write an alternative, then I need to make Rbx less slow
<yorickpeterse> Then we can start purging nonogiri
<brixen> yorickpeterse: working > fast
<yorickpeterse> well it does work already
<yorickpeterse> give or take some issues with HTML
<brixen> next would be a migration guide
<yorickpeterse> already done
<brixen> even better, a tool that rewrites your code
<brixen> :)
<brixen> yorickpeterse: very nice!
<brixen> next, marketing
<yorickpeterse> well I got featured in Ruby Weekly, finally :P
<brixen> I think this calls for a blog post
elia has quit [Quit: Computer has gone to sleep.]
<yorickpeterse> once I have some stuff running in production I'll be taking care of that as well
<yorickpeterse> which should be in ~2 weeks
<brixen> yorickpeterse: you want to target segments where people actually do things :p
<yorickpeterse> people starring the repo on Github doesn't count? :P
<brixen> yeah, it counts as much as page views
<yorickpeterse> hehe
<brixen> a high value, non-scalable activity would be to pick the top 10 gems using nokogiri and send PRs converting them to Oga
<yorickpeterse> gotta get that SEO going
<brixen> it should be easy to find out what those gems are
<yorickpeterse> I have some folks already adding support for their multi-xml/multi-sax-whatever Gems
<yorickpeterse> the AWS SDK is on my list of things to change myself
<yorickpeterse> but they do some crazy XML stuff there
<brixen> undoubtedly
elia has joined #rubinius
<yorickpeterse> headius: do I have to also submit slides/papers right away?
<yorickpeterse> brixen: but yeah, basically my plan for this week is to look into either fixing or replacing Racc
<yorickpeterse> Then starting next week I'd apply the results, spend some time making Rbx faster
<yorickpeterse> Once that's all done I can start making sure HTML support is solid
<headius> yorickpeterse: I don't think so
<yorickpeterse> headius: Ah ok
<headius> room-by-room and I don't think we care
<slaught> yorickpeterse: (re: racc) you might want to look at the antlr parser generator. I am just finishing up the ruby target & runtime for it.
<yorickpeterse> slaught: didn't antlr4 drop anything but Java output?
<headius> slaught: oh nice
<headius> hopefully without any C ext?
<yorickpeterse> Also, the requirement I have is that the output is 1) pure Ruby 2) faster than the current Racc setup
<slaught> in the official release. but antlr4 lets you write a new target easily.
<yorickpeterse> for 1) I can do with C/Java as long as my callback code is in Ruby
<slaught> the output and runtime is pure ruby.
<yorickpeterse> Looking at how god awful slow Racc is I'm quite certain this can be done faster without C
<yorickpeterse> * aweful
<slaught> the parser generator tool is in java. but that is a one time run on the grammar.
<yorickpeterse> Yeah I'm familiar with antlr
<slaught> i should be done with the attribution and writeup tonight. I will send you a link if you are interested.
<yorickpeterse> That would be nice, yes
<yorickpeterse> oh wait, antlr can still be used with a standalone lexer right?
elia has quit [Quit: (IRC Client: textualapp.com)]
havenwood has joined #rubinius
pietr0 has quit [Quit: pietr0]
<slaught> yes. you might have to change the lexer class interface a bit.
<yorickpeterse> ok I think I managed to submit something
<yorickpeterse> unless pentabarf barfed on it
<slaught> yorickpeterse: an antlr generated parser just needs the lexer to implement its tokenstream api. its like a dozen methods.
<yorickpeterse> Jesus
<yorickpeterse> It's literally 1 method in Racc
<yorickpeterse> oh ffs, the last resource I had of which I hoped it clearly explained how to construct LL(1) parsing tables also sucked
<yorickpeterse> none of these explain how to handle repetition for example
<yorickpeterse> and not a single one has a decent code snippet, it's either super abstract pseudo code or just formulas
pietr0 has joined #rubinius
pietr0 has quit [Client Quit]
pietr0 has joined #rubinius
dzhulk has quit [Quit: Leaving.]
<slaught> antlr will generate a lexer and parser from the single grammar definition.
<yorickpeterse> If the output of that were to be Ruby I can't use it, it would take far too much time and memory
<yorickpeterse> (been there, measured it, sadly couldn't stick with it)
<yorickpeterse> Not with antlr though, but with Ragel
<yorickpeterse> but even Ragel's Ruby output, while ugly, was pretty well optimized
<slaught> like I said you can use a custom lexer too.
pietr0 has quit [Ping timeout: 250 seconds]
pietr0 has joined #rubinius
yipstar has quit [Ping timeout: 255 seconds]
<johnmuhl> brixen et al.: opinions whether https://github.com/jruby/jruby/issues/2264 should be added to rubyspec or be just a regression test against jruby?
<brixen> johnmuhl: looks like a legit case to spec
<brixen> I didn't think ! was allowed in paths
<headius> I wasn't sure if it would be appropriate since it's the combination of ! and % that falls into special logic in JRuby for URLs
<brixen> I guess quoted it would work
<johnmuhl> brixen: OS X doesn't seem to mind it being part of path
<headius> it's a weird edge
<yorickpeterse> Linux also allows ! in filenames
<brixen> johnmuhl: if it's expected that Dir.glob handles that path correctly, it should be spec'd
<headius> brixen: thanks for clarifying...didn't want to clutter up rubyspec with odd one-off cases if they're really just jruby things
JohnBat26 has quit [Quit: KVIrc 4.3.1 Aria http://www.kvirc.net/]
<johnmuhl> brixen: ok I'll try to add a rubyspec soon
<brixen> johnmuhl: thanks
<headius> brixen: for future reference, at what point would it be considered too jruby-specific to add?
<brixen> a behavior that MRI doesn't have
<johnmuhl> added an issue to rubyspec so I won't forget
<johnmuhl> is there a better place to discuss rubyspec or is this channel ok?
<brixen> johnmuhl: here is fine
<brixen> there's also #rubyspec
<headius> brixen: thanks...I guess this remains a little grey, since there's nothing special about those paths on any impl but JRuby
<johnmuhl> fwiw I ran into this building an application to manage my music library and real artist/album/song names regularly include ! and %
<brixen> johnmuhl: yeah, they are unusual characters for paths considering their alternate meanings in shells
<brixen> eg cmd.exe
<brixen> slaught: oh, the wonderful world of double requires, remind me to tell you what we found
<slaught> brixen: i wait with dread and loathing
jnh has joined #rubinius
benlovell has joined #rubinius
lbianc_ has joined #rubinius
lbianc has quit [Ping timeout: 252 seconds]
lbianc_ is now known as lbianc
benlovell has quit [Ping timeout: 258 seconds]
yipstar has joined #rubinius
elia has joined #rubinius
yipstar has quit [Ping timeout: 264 seconds]
GitHub182 has joined #rubinius
<GitHub182> [rubinius] brixen pushed 2 new commits to master: http://git.io/Hmw3qA
<GitHub182> rubinius/master 98a1726 Brian Shirai: Unlock objects after fork() without locking.
<GitHub182> rubinius/master 8827765 Brian Shirai: Added MRI-specific GC-specific C-API defines for psych.
GitHub182 has left #rubinius [#rubinius]
<yorickpeterse> it actually works, though I probably really butchered it
<brixen> jc00ke: hm, I'm able to bundle you repro repo with --jobs=4
<brixen> jc00ke: gonna try on precise64
<jc00ke> @brixen OK, that's closer to my env (utopic64)
<brixen> well, I could do trusty64 too
<yorickpeterse> hm interesting, my LL hack is currently about as fast as Racc for the same grammar
<yorickpeterse> and that's shitty, pure Ruby vs shitty C
<yorickpeterse> and on Rbx my LL hack is 1,74 times faster than Racc
<yorickpeterse> yay JIT
<jc00ke> @yorickpeterse nice!
<yorickpeterse> and on jruby my LL hack is 3,74 times faster than Racc
tenderlove has quit [Quit: Leaving...]
tenderlove has joined #rubinius
elia has quit [Quit: Computer has gone to sleep.]
Ori_P has joined #rubinius
<yorickpeterse> so on Rbx my stuff is faster, but in general there it's still slower than the rest