<KDDLB>
This blog is essentially an ad for the Handshake ICO scam with a one-line "denial" of involvement mixed in there. It's obviously very unethical of Christel to not mention her own involvement in the scam which the blog post promotes.
<Guest75005>
This blog is essentially an ad for the Handshake ICO scam with a one-line "denial" of involvement mixed in there. It's obviously very unethical of Christel to not mention her own involvement in the scam which the blog post promotes.
<TReK21>
This blog is essentially an ad for the Handshake ICO scam with a one-line "denial" of involvement mixed in there. It's obviously very unethical of Christel to not mention her own involvement in the scam which the blog post promotes.
<klx``14>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<ChrisBr>
headius: hmm, difficult! So getting rid of the "resizing" issue would only be possible if we know the number of elements in advance (or use a very big array which would be a waste of memory)
<ChrisBr>
For kwargs, we should know the number of elements in advance, right?
<ChrisBr>
IIRC we do that with set. If you construct a new set with elements, we create the hash with the number of elements and no resizing is necessary (for constructing)
<ChrisBr>
However, this still has the disadvantage that knowing the number does not mean that we have a nice distribution in the hash (so we might still have collisions)
<ChrisBr>
but I would probably first go with the approach where we know how many elements we expect and construct the correct size so no resizing would be necessary
<ChrisBr>
this code should already be in place
<limbo0>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<ZLSA27>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
shellac has joined #jruby
fidothe_ has quit [Read error: Connection reset by peer]
drbobbeaty has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
drbobbeaty has joined #jruby
rdubya has joined #jruby
rdubya has quit [Ping timeout: 240 seconds]
rdubya has joined #jruby
<nikow11>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<kg23>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
shellac_ has joined #jruby
shellac_ has quit [Client Quit]
shellac_ has joined #jruby
shellac has quit [Read error: Connection reset by peer]
justinmcp has quit [Quit: No Ping reply in 180 seconds.]
<mikedlr17>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
bga57 has quit [Quit: Leaving.]
<basic`24>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
bga57 has joined #jruby
<Jguy>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
rdubya has quit [Ping timeout: 256 seconds]
rdubya has joined #jruby
shellac_ is now known as shellac
<GitHub113>
[jruby] mejackreed opened pull request #5276: Modifies underlying logic of slice_when / chunk_while to handle (master...fix-slice_when) https://git.io/fNDWV
<Cork18>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
xardion has quit [Remote host closed the connection]
<GitHub122>
[jruby] ChrisBr opened pull request #5278: Refactor and fix allocFirst with buckets for RubyHash (master...refactoring/hash) https://git.io/fNDPt
<iw00t9>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<iw00t9>
This blog is essentially an ad for the Handshake ICO scam with a one-line "denial" of involvement mixed in there. It's obviously very unethical of Christel to not mention her own involvement in the scam which the blog post promotes.
xardion has joined #jruby
<madduck10>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<headius>
ChrisBr: hi there
<headius>
so yeah all the cases I'm interested in are going to be known keys
<headius>
most kwargs calls are done with the literal syntax, and the ones passing around an opaque hash I don't care much about anyway because they're usually very generic code
<headius>
so given a literal call with foo:, bar:, baz: we would ideally create the lightest-possible hash for carrying them
<headius>
meanwhile I will also be working on just passing them on the JVM stack but we'll always have places that falls down and a Hash has to carry them
<headius>
I also experimented with having a dedicated per-thread kwargs hash that we clear and repopulate...that eliminates allocation entirely but it defeats some other optimizations because we're using a shared structure
<dysfigured10>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<MetaNova25>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<MetaNova25>
This blog is essentially an ad for the Handshake ICO scam with a one-line "denial" of involvement mixed in there. It's obviously very unethical of Christel to not mention her own involvement in the scam which the blog post promotes.
atm1sk has joined #jruby
<atm1sk>
hi all, i have been trying to get a clear answer on how to clean up JVM threads that nailgun server creates, is it even possible? is the only appropriate use of the nailgun server to fire up , run your jruby scripts and shut it down?
<mancha29>
Christel just posted this "denial" on the freenode blog https://freenode.net/news/spam-shake Why does this blog post mention "10.2 million" THREE times?
<headius>
@lopex if you can explain the $~ breakage I'll fix it
<headius>
atm1sk: unfortunately that is a limitation of using Nailgun
<headius>
I would suggest you spend your time looking at a Drip-style setup, where the background JVM is only used once
subbu is now known as subbu|away
<headius>
you can have it pre-boot JRuby at a minimum, and possibly parts of your app like spinning up Rails itself
rdubya has joined #jruby
headius is now known as JRubyGithub
JRubyGithub is now known as headius
<atm1sk>
headius: i am trying to pull some jmx stats out of cassandra without starting 2-4 JVMs for each script , nailgun just seems to keep growing and growing in threads no matter what type of params I pass it , be it java options or jruby options
<atm1sk>
its is extremely fast but the thread count is a concern
<headius>
atm1sk: are those threads that the JRuby app is starting or Cassandra?
<headius>
there may be a graceful way to shut those threads down
<atm1sk>
the jruby --ng-server
<atm1sk>
last night it launched ~1500 java threads
<atm1sk>
from having my monitoring system poll with jruby every minute
<headius>
you can get a dump of those threads by using jstack tool
<headius>
thread pooling doesn't really do anything in JRuby 9+
<atm1sk>
this is an el6 instance w/ jruby 1.7
<headius>
ok
<headius>
I'd say run a couple times and then get a jstack we can look at
<headius>
if they're Ruby threads then there's cleanup that's not happening in Ruby
<atm1sk>
do i have to end my script w/ some special exit?
<headius>
if they're Cassandra threads then perhaps there's some shutdown process for Cassandra you could call
<headius>
maybe?
<atm1sk>
i have been reading the nailgun source and i can't find any special code it looks for
<headius>
when the "main" script exits JRuby calls its own shutdown process, which at least closes streams and things (but does not forcibly kill threads)
<headius>
yeah nailgun is pretty simple
<headius>
but we provide a "NailMain" class that wraps the execution of JRuby...that could be made smarter to clean up threads, maybe
<atm1sk>
from what i am reading though , and looking at previous logs of this channel, others have had issues with threads overrunning their system w/ nailgun
<headius>
yes
<atm1sk>
drip does seem better
<lopex>
headius: reverted breaking commit for now, but it's was path with a good direction
<lopex>
headius: all because MatchData.regexp and MatchData.pattern bifurcation
<lopex>
headius: they were introduced so if the match doesnt escape, no need to create core objects
<lopex>
headius: there's many unnecessary checks now once so many optz were removed
<headius>
ok I understand
<headius>
the problem is that the use of $~ is so buried andmixed into the rest of regexp logic
<headius>
it's really hard to keep the sharing clean without having two paths for everything
<lopex>
headius: pushed more reverted plus a named_captures fix
<lopex>
headius: "l".scan("l"){$~.named_captures} was broken on 9.2
<lopex>
headius: we need to remove that dualism from there
<headius>
yeah overzealous removing $~ use from some paths
<lopex>
no, just RubyMatchData.pattern can be null
<lopex>
headius: also removed that useCount for now
<atm1sk>
headius: looks like i solved my problem , due to leaving my JMX::MBean.establish_connection object open the threads kept piling on , giving it .remove_connection solved the issue
<atm1sk>
i underestimated the persistence of the nail aspect
<ChrisBr>
headius: ok, just let me know what you have in mind or create issues and I can look into it
<ChrisBr>
btw: I saw now a few times that PHASE='-Prake -Dtask=spec:regression' times out, was this already the case before or could be introduced by the hash?