<brixen>
jc00ke: if you have :foo, it's GC'd and then you create another :foo, I'm pretty sure it's going to get a different object_id
<brixen>
unless there were an absolutely perfect hash function on a pretty huge set of strings
<brixen>
and the object id is computed purely from :foo and nothing else
<brixen>
jc00ke: but yes, rbx 3 at least will GC symbols
<brixen>
it's probably a few hours work at most
lopex has joined #rubinius
elia has quit [Quit: Computer has gone to sleep.]
|jemc| has quit [Ping timeout: 255 seconds]
pd_ has joined #rubinius
ssedov has joined #rubinius
imajes_ has joined #rubinius
btcctf_ has joined #rubinius
Gibheer_ has joined #rubinius
heroux_ has joined #rubinius
pd has quit [Ping timeout: 264 seconds]
pd_ is now known as pd
pd has quit [Changing host]
pd has joined #rubinius
coffeeju1 has joined #rubinius
heroux has quit [*.net *.split]
btcctf has quit [*.net *.split]
Gibheer has quit [*.net *.split]
stass has quit [*.net *.split]
imajes has quit [*.net *.split]
coffeejunk has quit [*.net *.split]
imajes_ is now known as imajes
heroux_ is now known as heroux
coffeeju1 is now known as coffeejunk
havenwood has quit [Remote host closed the connection]
meh` has quit [Ping timeout: 245 seconds]
havenwood has joined #rubinius
centrx has quit [Quit: Science is organized knowledge, wisdom is organized life.]
Bwild has joined #rubinius
JohnBat26 has joined #rubinius
mqt_ has joined #rubinius
mqt_ is now known as mqt
Gibheer_ is now known as Gibheer
Gibheer has quit [Quit: leaving]
Gibheer has joined #rubinius
josh-k has joined #rubinius
josh-k has quit [Ping timeout: 265 seconds]
elia has joined #rubinius
elia has quit [Quit: Computer has gone to sleep.]
elia has joined #rubinius
josh-k has joined #rubinius
josh-k_ has joined #rubinius
josh-k has quit [Ping timeout: 256 seconds]
elia has quit [Quit: Computer has gone to sleep.]
<yopp>
brixen, it do get new object_id
<yopp>
brixen, also, symbols are frozen starting from 2.1
<yopp>
(or maybe even from 2.0)
robin850 has joined #rubinius
<yorickpeterse>
yopp: Symbols have been frozen since 2.0
<yopp>
so there no harm in GC'd symbols, if you are not using their object_ids.
havenwood has quit []
josh-k_ has quit [Remote host closed the connection]
mbj has joined #rubinius
Bwild has quit [Quit: leaving]
elia has joined #rubinius
josh-k has joined #rubinius
diegoviola has joined #rubinius
<diegoviola>
I'm pretty confused right now
<diegoviola>
rbx talks a lot about respect and having a great community
<diegoviola>
i.e. "We know this will be an important aspect of creating an environment of respect and support as we continue to explore how to improve the relationship between Rubinius as a product and project and those who use Rubinius."
* yopp
sets "Days since last Ruby community drama" counter to 0
<yopp>
diegoviola, I agree that ruby-core should include rubyspec in their development process in some way.
elia has quit [Quit: Computer has gone to sleep.]
<DireFog>
long time not checked what's going on here ^^
diegoviola has quit [Client Quit]
diegoviola has joined #rubinius
<chrisseaton>
yopp: I don't think anyone in that argument disagreed with Brixen that it would be nice if ruby-core would whole-heartedly adopted RubySpec
<chrisseaton>
personally I would be in favour of a formal semantics maintained by ruby-core, rather than a big bag of specs
<yopp>
rubydramas had counter and drama list :B
elia has joined #rubinius
<chrisseaton>
but I don't have the technical ability to write one
<brixen>
diegoviola: in typical fashion, the people promoting the drama and making personal attacks are misrepresenting what I said
<diegoviola>
brixen: I see, sorry about that
<brixen>
I said, based an a reasonable metric like percentage of 2.0, 2.1 and 2.2 features implemented versus rubyspecs written, neither MRI nor JRuby contribute to rubyspec
<brixen>
anyone can go verify that
<brixen>
no one has
<brixen>
least of all the people making personal attacks
<brixen>
I did not say I don't value contributions
|jemc| has joined #rubinius
<brixen>
I did clearly say that a few PRs do nat constitute contribution by the project as a whole in a meaningful fashion
<brixen>
I'm very happy to have this discussion publicly without personal attacks
<brixen>
which is why rubinius *is* adopting a code of conduct
<brixen>
and why I will not hesitate to ask people to leave if they continue making personal attacks
* DireFog
should probably /part then since three years of Rails have not made him a gentler person
<brixen>
DireFog: it's simple, just don't make personal attacks
<brixen>
I'm obviously frustrated by MRI's behavior
<brixen>
it's hurting Ruby
<brixen>
more people leave Ruby every day
<brixen>
I've spent a ton of time on rubyspec (and rubinius) to make Ruby better
<yopp>
maybe it's time to meet with jruby/mri teams and discuss why rubyspecs are only maintained by rbx team?
<brixen>
the one lesson I've learned more than any is not to put up with bullshit from people
<diegoviola>
yopp: I was going to suggest that as well
<brixen>
some people won't agree with that, but they are not contributing anyway
<brixen>
the rest of people have to deal with this sort of thing, and it makes your life miserable sometimes
<yopp>
It might be that they have some technical reasons, not political? Or they don't see value in contributing that way, and they have better idea?
elia has quit [Quit: Computer has gone to sleep.]
<|jemc|>
I'd say that ideally the biggest part of the burden would be on MRI and not as much on jruby (and rbx) because MRI is driving most of the language changes
<|jemc|>
it's like asking that a library you use maintains complete spec coverage
elia has joined #rubinius
<|jemc|>
others can step in and add specs if they are missing, but ideally all new features should come with specs
<yopp>
|jemc|, at least they can open issues and describe what changed
<yopp>
and how it's changed
<|jemc|>
asking them to integrate the process of running new versions of ruby against rubyspec to prevent regressions is also reasonable, in my opinion
<chrisseaton>
I don't want to start an argument again - but I will give one concrete example of why I didn't dive into contributing more RubySpecs when I started on Truffle: I opened my first tiny RubySpec PR about 18 months ago - it sat unmerged and without any comment until a couple of days ago
<|jemc|>
with my last comment being in reference to the recent regression with how kwargs behave (something about no longer calling to_hash?)
<|jemc|>
I wouldn't be surprised if that was an unconsious side-effect of another change rather than a design decision (but admittedly I haven't looked for a ticket on the subject to see)
<brixen>
chrisseaton: at least 3 people on the jruby team had commit access and could have helped you with that
<brixen>
chrisseaton: and you also could have bugged me about it more
<brixen>
rubyspec is a *huge* effort
<brixen>
so, an actually useful discussion if anyone has ideas
<brixen>
what should we use instead of IRC?
<brixen>
or should we continue to use IRC?
<brixen>
the reason I ask is because a lot of newcomers are not familiar with IRC
<brixen>
every time this has been told to me I brush it off because IRC works great for me :)
<chrisseaton>
I really like the markup that Gitter allows - I think it makes discussing code a lot easier - and it has an IRC interface so you can keep using that if you want
<yopp>
brixen, you should start by gathering jruby/mri core members on one place, to talk about what's going on
<yopp>
imho
elia has quit [Quit: Computer has gone to sleep.]
<|jemc|>
brixen: you're talking about where #rubinius users should meet in general?
<brixen>
yopp: I have
<brixen>
yopp: I've been working on this project for 8 years
<brixen>
I can point to multiple people with whom I've raised this issue who are unrelated to this project as well
<brixen>
mostly because JRuby and Rubinius teams used to work at EY
<brixen>
multiple managers can attest that I've raised this issue because it was such an issue
<|jemc|>
brixen: I haven't used slack or gitter but IRC is most convenient for me
<brixen>
and because charles has repeatedly made personal attacks about it
<|jemc|>
if gitter has a way for me to keep using IRC, that would be nice
<|jemc|>
If users have trouble getting to IRC we could do more for them in that regard - provide a web client on rubini.us for example
<brixen>
yopp: additionally, I've sat down with matz and MRI devs multiple times
<yopp>
and what they are saying?
<brixen>
yopp: only yugui, as 1.9.2 release manager, required MRI to make rubyspec pass
<brixen>
she simply said, I'm not releasing 1.9.2 until rubyspec passes
<brixen>
again, anyone can see this in the commit history
<brixen>
look at 2.0, 2.1, 2.2 features and look for PRs or commits from MRI or JRuby to rubyspec
<yopp>
and later? Why they stop doing that way?
<brixen>
and these people had commit bits
<yopp>
•stoped
<brixen>
I'm not stopping them
<brixen>
yopp: you'll have to ask them
<brixen>
I also tried to make rubyspec a part of a decent Ruby design process
<brixen>
you can watch the video
<brixen>
you can see the blog posts at brixen.io that I wrote clarifying the misunderstandings
<brixen>
you can see the redmine ticket about it
<|jemc|>
brixen: although it occurs to me that asking about whether we should keep using IRC to a group of users on IRC the day after christmas may give you some sampling bias :P
<brixen>
also, this point is perhaps subtle, but I don't have to agree with how MRI decides people can interact with their project
<yopp>
brixen, it's limited only by mri members?
<brixen>
yopp: well, let me clarify, there is no such thing
<brixen>
def facto, there is matz and the devs
<brixen>
my attempt to make it a formal thing with useful rules was dismissed
<brixen>
along with a nice campaign of FUD against it
<yopp>
with what reasoning?
<brixen>
including assertions that my rules required MRI devs "having lunch to not use Japanese" and other nonsense
<brixen>
yopp: you can read the ticket and infer what you want
<brixen>
also, these are really not trivial issues
<brixen>
the current MRI feature process is a total mess
<brixen>
Rails 5 stated it "requires Ruby 2.2" is laughable
elia has quit [Quit: Computer has gone to sleep.]
<brixen>
but because there is no process, *implementation details* like whether a Symbol is frozen are now Ruby version features
<brixen>
all the hacks around literal frozen Strings, too
<brixen>
or Symbol GC
<yopp>
when the feature is accepted and merged into the "official" Ruby,
<yopp>
the behavior should gradually be described by RubySpec. We may need
<yopp>
more members to maintain RubySpec. And we may need to communicate
<yopp>
more to distinguish "the spec" and the implementation dependent.
<DireFog>
wouldn't a feature ideally be specified *before* it's merged into official ruby?
<brixen>
it certainly should
<DireFog>
as in that whole "engineering" thing?
<brixen>
there were no tests in MRI for Thread::Backtrace::Location#path
<brixen>
Rails depends on a bug
<brixen>
we implement it by the docs
<brixen>
all our Rails 4.1 apps can run for a month without monkey patching
<DireFog>
Ruby by and large still feels too much like the entire *-lang.org environment despite having grown beyond the usual 3-people-including-all-users scope that's common for that realm :-|
<brixen>
it cost a significant number of tickets and dev time wasted
<brixen>
again, I don't have to accept that MRI says, just use Redmine
elia has joined #rubinius
<brixen>
I do try to search for things on that but it takes a lot of time
<brixen>
so I go read C code, which does not describe intent
<brixen>
or Changelog entries
<brixen>
etc
<brixen>
further, it's hilariously ironic that the people saying, "MRI says you have to do this, so you do" are bitching because I state how they have to work with my project
<brixen>
and the requirements I give are not arbitrary, they are necessary to not end up with MRI's test suite
<brixen>
if anyone wants to defend the quality of MRI's tests to say, the Haskell community, I'd love to attend that talk
<robin850>
brixen: I agree with what you said, especially regarding Rails and 2.2+. This is a really bad decision even though it can seem legit because a lot of work have been done on trying to provide support for JRuby and Rubinius (and I'm not saying that because I worked on that during the GSoC but because a lot of people are actually using Rails with other rubies)
elia has quit [Client Quit]
<robin850>
and obviously they aren't doing any effort regarding RubySpec!
<robin850>
s/they/the MRI team
elia has joined #rubinius
<yopp>
which is funny, because even matz said they should
|jemc| has quit [Quit: WeeChat 1.0.1]
<brixen>
robin850: thanks for all your help during GSoC, btw
<robin850>
brixen: My pleasure! RubySpec has been a real gold mine to understand better things, btw if you want some help with it, let me know :-)
<brixen>
robin850: heh, well, I've been thinking about RubySpec 2.0 :)
elia has quit [Client Quit]
<brixen>
one of the goals I've had for rubyspec is to combine the Ruby documentation with the specs
|jemc| has joined #rubinius
<brixen>
so, one thought for a better rubyspec would be a book in literate programming style where the tests are the examples
<brixen>
another thing I've been toying with is something like cucumber, but with better language parsing
<brixen>
ie, it's not just a front for a bunch of rspec manually written
<yopp>
brixen, that one "mystery ruby impl" was RubyMotion?
<brixen>
yopp: nope, Topaz
<yopp>
ah
<brixen>
also, Opal is doing some amazing stuff
<brixen>
someone I met recently is using it
<robin850>
brixen: Yep pretty good idea. This is what CommonMark is doing for its specs
<yopp>
so, seems like the main problem is simple: Ruby as language is relates on MRI, instead of relating to standarts
<yopp>
And everything else is just a side effect
<|jemc|>
heh, cucumber is pretty obnoxious but the intent is nice
<diegoviola>
yeah like the JVM is the only one to know how ot do things, sure
<brixen>
wow, I still have yarv-dev emails from < 2006
<brixen>
I tried to help with YARV before it was even in MRI
<yopp>
from my POV, seems like it became too personal
<diegoviola>
I really want to see RBX thrive as well as all the other implementations, Ruby is awesome as a language and it's annoying the world is moving to JS as well
<diegoviola>
when Ruby has been so amazing to work with all this time
<brixen>
diegoviola: I really want to see it thrive, too
<yopp>
huh
<brixen>
diegoviola: I have a big problem with the way matz designs features, though
<brixen>
I'd be willing to assert that matz is doing things harmful to the Ruby community
<diegoviola>
it does sounds like there's too much ego conflict here, why not put this aside and do what is best for Ruby
<brixen>
I'd like someone to suggest a metric and show evidence that is not true
<yopp>
^^ ++++
<brixen>
because everyone I talk to laughs at Ruby and advocates using other languages
<brixen>
Go, Clojure, JS, Haskell, Scala, plain old Java
<yopp>
brixen, as a enterprise customer I can tell why Ruby support is limited
<brixen>
and the same reasons come up over and over: performance, concurrency, and quality
<yopp>
1) No commercial support at all. No one doing this
<yopp>
2) Before 2.0, ruby had a messy release cycle
<yopp>
3) There no language specs, or whatever, so it's risky to invest, because you can't be sure, that tomorrow will be Ruby 3.0 that will be totaly backward incompatible, and you will be forced to invest unknown amount of money
<yopp>
as result you can't calculate TCO of ruby project,
<brixen>
yopp: yep
<DireFog>
yopp: not to mention that deploying anything is a total pain
<yopp>
DireFog, huh
<brixen>
deploying is certainly still harder than it needs to be
<yopp>
deploying anything on any language is almost always is a total pain
<brixen>
raise your hand if you've ever been bit by having a system ruby conflict with an app ruby
<DireFog>
deploying a ruby project either comes with a lot of tooling or just stuffing full devel toolchains on production machines
<brixen>
we built rbx from the beginning to run fine isolated in a single dir
<brixen>
DireFog: totally doesn't need to
<diegoviola>
deploying certainly got a lot better *for me* with things like chruby and ruby-install
<robin850>
brixen: O/
<yopp>
DireFog, that's because linux package management if pice of outdated crap
<brixen>
it's completely possible to deploy a single tarball
<yopp>
•piece
<brixen>
robin850: ?
<brixen>
robin850: oh, hand raised? heh
<robin850>
brixen: raise your hand :p
<diegoviola>
yopp: I disagree
<DireFog>
did you see the talk about how Flickr ended up deploying their Python tools?
<brixen>
robin850: heh, didn't recognize that
<yopp>
diegoviola, on what? :)
<diegoviola>
yopp: on saying that package management on linux is outdated crap
<yopp>
outdated in terms of content, not technology
<diegoviola>
yopp: on some distros packages are out of date, sure
<yopp>
tr/some/most/
<DireFog>
if you're talking enterprise, they'll be massively outdated
<yopp>
huh. Because you can't update just one service, without updating the half of the system
<yopp>
And you can't do that, in most cases
diegoviola has quit [Quit: WeeChat 1.0.1]
<yopp>
Because downtime can cost you a fortune
<brixen>
anyone play with asmjs much?
<yopp>
brixen, I remember, you said that you have customers for commercial support? :) My contract might not be renewed for next year, so I better start doing something else.
<yopp>
It will be tough time in Russia, so I can get good guys for good price :)
<brixen>
yopp: sucks about the situation in Russia :(
<brixen>
yopp: we should discuss this more
<brixen>
yopp: email me, brixen at gmail, sometime
<yorickpeterse>
yopp: you can always flee to Europe
<yopp>
yorickpeterse, am already on the move. Since June :)
<yorickpeterse>
Ah
<yorickpeterse>
come live in captalist pig Europe
<yopp>
But I don't have valid permission to stay for the long time, so I'm driving in circles around Europe :)
<yopp>
Right now I'm in Albania, in Jan I'll be moving to Spain for couple months
<yorickpeterse>
Yeah, you need a job to get a more long-term permit
<yorickpeterse>
or marry a European citizen
<yorickpeterse>
yopp: there are actually quite a bunch of Ruby shops here in Amsterdam that would probably hire you on the spot
<yorickpeterse>
yopp: the two companies to avoid are Booking.com and Tomtom
<yopp>
Too cold. And I'd like to do my thing, I'm a bit tired of doing someone else things
<yopp>
Booking is a Kingdom of Perl, half of Russia already moved there :B
<yorickpeterse>
The former is a slave sweat shop, the latter has a bunch of issues regarding LGBT/women if I recall
<yorickpeterse>
also the weather here isn't too bad
<yorickpeterse>
it actually gets pretty warm during summer
<yopp>
I pref south of Italy. +16 in the winter, and about 30 in the summer. Not so hot
<yorickpeterse>
That also happens to be the part of Europe that's doing less well atm :P
<yopp>
Food is amazing, sea is stunning. But immigration laws are bloody tough
<yopp>
Yeah, but it's way better that in Russia :)
<yopp>
And they don't have snow!
josh-k has quit [Remote host closed the connection]
<yopp>
I used to live in Siberia, with 9 months of snow. So I'm completed my snow quota for the rest of my life :)
<brixen>
yopp: they need Ruby programmers around Barcelona, I think
<brixen>
I would so love to live there
<yopp>
I never lived in Spain for a long time, so we're rented apartments for a couple month around Almeria
<yopp>
To get an idea of country. I've been in Barcelona and surrounding areas once, but it was more like "travel by"
<yorickpeterse>
yopp: either way, if you need anything in .nl just holler, I can probably hook you up
* yorickpeterse
sounds like a drug dealer
<yopp>
:B
<yopp>
Thanks, if things will go bad, I'll get to you to get /some/ ;D
<|jemc|>
If anyone hasn't seen Nim yet, they've got some interesting stuff going on
<bennyklotz>
rbx 3.0 should be compatible with ruby the language itself and rbx x is a superset of rbx 3, but what exactly should titanius be?
<bennyklotz>
also when introducing fun syntax in rbx 3.0 and at the same point saying "My objective for introducing these features into Rubinius 3.0 is to massively reduce the complexity of the current implementation of Ruby" is also confusing for me, is the new fun syntax only available for rbx internal ruby code?
<|jemc|>
bennyklotz: if I understand correctly, titanius would be the language platform rather than a language itself
headius has left #rubinius [#rubinius]
<bennyklotz>
|jemc|: thx, "we are extracting a more useful language platform out of Rubinius as the Titanius system" should read a little bit slower :D
<|jemc|>
bennyklotz: the paragraph above that helps to explain it
<|jemc|>
rubinius is already a great platform for building languages that are ruby-like (like my myco language)
<bennyklotz>
okay thx :)
<|jemc|>
titanius would expand on those internal features to make it a good platform for a wider variety of languages
<|jemc|>
which is something I'm excited about :)
<bennyklotz>
yep sounds really interesting :)
amsi has joined #rubinius
havenwood has joined #rubinius
<yorickpeterse>
bennyklotz: titanius will be the VM
<yorickpeterse>
the idea is to extract it and provide something similar to Parrot
<yorickpeterse>
rbx 3.0 is just rbx + better bytecode/jit/somet other things
goyox86 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]