brauliobo_ has quit [Remote host closed the connection]
brauliobo has joined #jruby
pawnbox has quit [Read error: Connection reset by peer]
pawnbox has joined #jruby
benlovell has quit [Ping timeout: 276 seconds]
donV has joined #jruby
brauliobo_ has joined #jruby
brauliobo has quit [Ping timeout: 276 seconds]
pawnbox has quit [Remote host closed the connection]
subbu_ss has quit [Ping timeout: 244 seconds]
brightball has joined #jruby
subbu_ss has joined #jruby
daicoden has joined #jruby
pawnbox has joined #jruby
daicoden has quit [Ping timeout: 244 seconds]
nirvdrum has joined #jruby
tomjoro has quit [Remote host closed the connection]
johnsonch_afk is now known as johnsonch
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #jruby
jeremyevans_ has quit [Quit: leaving]
benlovel1 has quit [Ping timeout: 276 seconds]
whitby has joined #jruby
benlovell has joined #jruby
benlovell has quit [Read error: Connection reset by peer]
Prasun has joined #jruby
thedarkone2 has joined #jruby
benlovell has joined #jruby
jeremyevans has joined #jruby
colinsurprenant has joined #jruby
bjfish2 has joined #jruby
lan5432 has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
blandflakes has joined #jruby
flujan has joined #jruby
<flujan>
Hello guys, I hava about 30 ruby process running on a box using ruby 2.3. I am trying jruby 9.0.5.0 but the memory is all consumed when I started the 30 processes. How can I run the same processes on the same java vm process?
tomjoro has joined #jruby
subbu is now known as subbu|afk
Prasun has quit [Quit: Leaving]
camlow325 has joined #jruby
<kares_>
flujan: depends on what kind of processes - you might want to refactor?
ITXpander has quit [Quit: Leaving.]
<kares_>
in general you could start the code they're running in Thread.new { ... } (assuming its thread-safe)
<flujan>
kares_: i am trying nailgun. but without success…
<kares_>
there's also a way for jruby to start in-process
<flujan>
I can start the process but they are not working.
vtunka has quit [Quit: Leaving]
<kares_>
you do not want nailgun on production ... at all!
<kares_>
flujan: its -Xlaunch.inproc=true
<kares_>
for system calls etc
<flujan>
kares_: its is not already default to true?
<kares_>
flujan: nope
<lan5432>
kares_: in reification, the ruby object that is going to become a java object, is it a Java object wrapped inside a class?
<kares_>
lan5432: not sure what you're asking for
<kares_>
flujan: also it will only start in-proc if it detects it can do so (which I am not sure how it does)
<kares_>
lan5432: seen reify explained yesterday - start up VisualVM and you should see - without reify you have a bunch of RubyClass instances for each Ruby class
<kares_>
with reify you have a unique Java class generated for each Ruby class
<kares_>
AFK
<flujan>
kares_: yeah, I will probably need to create threads instead of processes to run the scripts… :(
<lan5432>
I was checking for the issue with newinstance with parameters, and enebo told me the problem was related to reifying, so I thought Java was not finding the constructors or something
<kares_>
lan5432: recall I linked a related issue previously - in that issue you had a commit referenced that resolved the issue (but needed more work)
<kares_>
if that is the case I should have some WiP locally around that
<lan5432>
Yeah, but I'm still trying to fully understand reifying before I get to the issue itself
enebo has joined #jruby
Aethenelle has joined #jruby
jeremyevans has quit [Ping timeout: 248 seconds]
donV has quit [Ping timeout: 252 seconds]
Aethenelle has quit [Quit: Aethenelle]
Aethenelle has joined #jruby
skade has joined #jruby
subbu|afk is now known as subbu
benlovell has quit [Ping timeout: 244 seconds]
<GitHub161>
[jruby] enebo pushed 3 new commits to jruby-1_7: https://git.io/vVZOb
<GitHub161>
jruby/jruby-1_7 48f5908 Per Lundberg: Workaround for JRuby bug with UTF-8 encoding on non-English Windows...
<GitHub161>
jruby/jruby-1_7 afc3763 Per Lundberg: As suggested, let's try with a `FormatMessageW`-approach instead.
<GitHub161>
jruby/jruby-1_7 a53970c Thomas E Enebo: Merge pull request #3767 from ecraft/fix-win32-registry-error-handling-bug-with-utf8-windows-other-languages...
<GitHub4>
[jruby] enebo closed pull request #3767: Workaround for JRuby bug with UTF-8 encoding on non-English Windows (jruby-1_7...fix-win32-registry-error-handling-bug-with-utf8-windows-other-languages) https://git.io/vVmbf
shellac has quit [Quit: Computer has gone to sleep.]
rcvalle has joined #jruby
<enebo>
kares_: sorry I was away
_ko1 has quit [Ping timeout: 244 seconds]
<enebo>
kares_: yeah I do not think it would be bad to add…Playing Devil’s advocate though there are many Ruby Array/Hash methods we have not added to Java List/Map
<enebo>
kares_: So I guess I see two considerations: a) how common will someone want to dig b) will Java end up adding their own dig and we wil conflict
<enebo>
kares_: seems like for a) I don’t actually have any idea how popular dig will be but people obviously use it in Rails pre-2.3
<enebo>
kares_: for b) I suspect this is not very likely
<enebo>
kares_: I think asking headius to see what he thinks will probably see if he has any other insights but I am not against it
jeremyevans has joined #jruby
<enebo>
kares_: also I think this should be for Map and List if we do it
<eregon>
sounds like a nice way to improve Java interop, but there might not be many use-cases
<enebo>
eregon: yeah at one point too we thought about a concept called lightweights where data storage of Array/Hash was just a List/Map that would duck type
<enebo>
eregon: so the notion of having java types be more transparent is appealing
<eregon>
Scala has some auto conversions like that for collections from/to Java
<enebo>
eregon: I think the main issue with having a high level of transparency is all interaction with Java types force coercion costs. So in that sense if you plan on working on a simple datastructure from Java you are probably better off converting it to Ruby once
<enebo>
eregon: if we make it too transparent we will get endless this is slow reports
<eregon>
In the interop code I wrote I like to be very explicit, using Java conventions (like method names) but that's mostly for wrapping a library into a Ruby library
<enebo>
eregon: I used to do that but I ended up deciding that the object being in Java is an impl-detail
<enebo>
eregon: granted a potentially important impl detail
<eregon>
enebo: yeah, Ruby -> Java can be cheap, but not easily the other way around I guess?
<enebo>
eregon: gives me that hungarian notation feel now though
<eregon>
(RubyArray can just implement the List interface)
<jimbaker>
enebo, fwiw, we have found semantic equivalence/auto conversions in jython to work just fine. but users can always work with the underlying objects (with a little more work, and get more control)
<enebo>
eregon: RubyArray does actually implment List
<jimbaker>
at the very least it has been part of jython since the beginning :)
<enebo>
jimbaker: yeah we in general do not have semantic issues so much with our auto coercion it usually ends up being surprising to people there is coercion happening and they do not realize the cost
<enebo>
jimbaker: but most people don’t even notice it is happening
<enebo>
and frankly for most people the cost is not an issue
<jimbaker>
it really only impacts edge cases. like value vs identity comparisons
<jimbaker>
but again, we can offer more control as necessary
<enebo>
jimbaker: how is jython going?
<jimbaker>
i'm currently a bit busy: architecting a new openstack project; teaching a univ course on principles of programming languages; and oh yeah, jython :)
<jimbaker>
but 2.7.1 is close...
<enebo>
cool
<eregon>
jimbaker: do you happen to know about the object model of Jython? I'm actually doing research in object storage models these days.
<GitHub142>
jruby/master a9e3f89 Thomas E. Enebo: Comment out failing spec with reference to tracking issue #3771 (backslashes in URIs)
blandflakes has joined #jruby
<jimbaker>
eregon, i do in fact know the object model of jython very well. i'm tied up in meetings for a bit - switched over to my real job! - but happy to discuss on #jython later today or when it works for you
<eregon>
jimbaker: great! Let's discuss tomorrow if that works for you
<jimbaker>
yeah, friday is more open, that's for sure. i'm in boulder, colorado (america/denver); i will be around in the afternoon/no meetings
<kares_>
enebo: thanks - I stumbled across this as I handled all of the (Hash) methods for Map
<kares_>
since the Map wrapper acts like a Hash instance already
<kares_>
one use case for dig would be smt like a parsed (Java) json/xml structure containing List/Array/Maps
<kares_>
eregon: RubyArray already implements List
<kares_>
oh I see that was already noted above :)
<kares_>
also why I considered dig is that it leaks on all structures in Ruby that do []