<kares[m]>
~ 160x improvement (concerns few internal raises as well as `raise type, msg, nil`)
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<lopex[m]>
kares: with default settings ?
<kares[m]>
lopex: what do you mean? jruby -X opts?
<lopex[m]>
yeah
<kares[m]>
yes, but its simply since the changes done do introduce the Ruby exception hierarchy into Java (or later) JRuby started paying the cost for still generating (un-used) trace in some cases
<kares[m]>
so its kind of a regression really
<kares[m]>
what's also interesting is that Java 11 is half the speed (slower) when it comes to creating+throwing exceptions compared to Java 8
<kares[m]>
haven't looked but recall headius possibly mentioning its due the introduced stack-walking?
subbu has quit [Ping timeout: 268 seconds]
subbu has joined #jruby
_whitelogger has joined #jruby
drbobbeaty has joined #jruby
_whitelogger has joined #jruby
anubhav8421[m] has joined #jruby
<lopex[m]>
maybe some changes wrt fillinstacktrace ?
lucasb has joined #jruby
nirvdrum has joined #jruby
<kares[m]>
likely so
xardion has quit [Remote host closed the connection]
xardion has joined #jruby
nirvdrum has quit [Ping timeout: 258 seconds]
<enebo[m]>
kares: how does this fit with our other work for '1 raise nil' optimization?
nirvdrum has joined #jruby
<headius[m]>
Yes it is the new stack walker
<headius[m]>
I made some changes to not use it if a full trace is requested
<headius[m]>
I believe it has been improved post-11 but obviously that isn't a huge help to us
ur5us has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]