mengu has quit [Remote host closed the connection]
<headius>
huh seems to have corrected itself when I rebased the PR
<headius>
enebo: seems some spec failures after your changes?
<rdubya>
headius, enebo: sorry, I was afk, that code is used on every query that gets converted to an ActiveRecord::Result object, which I believe is most of them
<rdubya>
I'm pretty sure the previous version of the adapter just returned arrays of data that had already been converted, but that wasn't playing well with how they refactored types in Rails 5, iirc
<enebo>
mengu: yeah I am not sure if he is just asking to understand better or if he is pushing back because it is extra support for them
<mengu>
enebo: we'll see
<enebo>
I really don't see them dumping those two values internally
<mengu>
i don't think it'd cause them no harm
<mengu>
*any harm
<enebo>
rdubya: ^ look at that issue and my comment to make sure we only need oid and mod. I do not see any other uses
xardion has quit [Remote host closed the connection]
<rdubya>
enebo: yeah those are the only values we need at this point, iirc I was able to get all the other info I needed publicly
<enebo>
rdubya: thanks for verifying...I searched our code but we can be tricky sometimes from ruby side
xardion has joined #jruby
<mengu>
how's ruby guys doing it
<mengu>
i did check it out but didn't get anything
<mengu>
too much magic
<enebo>
mengu: the native c ext works directly with oid and mod which is why ar is written passing those values around
<enebo>
mengu: I think those are significant non-internal things if you write to pg directly which is why JDBC adapter is storing them as fields
<enebo>
so I am pretty hopeful he will open those two values up
<enebo>
I am sympathetic to opening all of field if not all those fields/methods are more than their impl details
mengu has quit [Quit: Leaving...]
rdubya has quit [Quit: Leaving.]
<lopex>
headius: busy busy now, but next week will be better
<lopex>
headius, enebo: will unicode 11 be a blocker for our 2.6 compat release ?
shellac has quit [Ping timeout: 240 seconds]
rdubya has joined #jruby
rdubya has quit [Quit: Leaving.]
rdubya has joined #jruby
mengu has joined #jruby
<mengu>
night all
<mengu>
i don't know if its AR bug or ARJDBC bug, it does not take timeout parameter in the config into account
<mengu>
i'm investigating
<mengu>
looks like it is actually AR
<mengu>
yes
<mengu>
spec.config[:checkout_timeout]
<mengu>
this is nil, despite the fact that i set timeout: 60000 in database.yml
<mengu>
OH, wait, it's my mistake i think
<mengu>
goddammit
<mengu>
i use env var for db url
<mengu>
i should have put it there
<mengu>
meh
<mengu>
no problems :D
rdubya has quit [Quit: Leaving.]
rdubya has joined #jruby
<headius>
lopex: probably not? What typically changes in these updates?
<lopex>
headius: code ranges
<lopex>
mostly
<headius>
yeah but only on higher plane weird stuff right?
<lopex>
whole ranges for unicode
<headius>
I'm just trying to understand how likely it is a typical user would notice an update from unicode 10 to 11
<lopex>
of course all of them are esoteris
<lopex>
*esoteric
<lopex>
probably not
<lopex>
headius: that \X change is blocking the update
<lopex>
oh well,...
<lopex>
\X changed
<lopex>
so some emoji wont work :P
<headius>
I don't think it blocks then
<lopex>
yeah, just asking
<headius>
unfinished bullet for 2.6 support when we release if it's not in there
<lopex>
git df
<lopex>
er
<lopex>
yeah, only ranges
<lopex>
and no encoding changes
<mengu>
channel, may i ask what objectspace is? why is it disabled?
<lopex>
generation scripts stilll work ok though
<lopex>
mengu: weakref pressure
<lopex>
and contention
<lopex>
and a lot more allocation for weakrefs
<lopex>
so in turn gc pressure :P
<mengu>
gotta read on then
<mengu>
*that
<lopex>
headius: though, since there's not a whole lot working threads in jruby apps mostly, then tlabs are not a big issue right ?
<mengu>
looks like the newrelic gem has some problems then, opt/tomcat/webapps/ROOT/WEB-INF/gems/gems/newrelic_rpm-5.5.0.348/lib/new_relic/agent/transaction_time_aggregator.rb:101: warning: ObjectSpace is disabled; _id2ref only supports immediates, pass -X+O to enable
<lopex>
well, threads competing for allocation
<lopex>
mengu: but classes and singletons are in OS afair
<lopex>
mengu: mri used to us mark/sweep so it always had those objects in hand
<lopex>
*s/us/use/
<mengu>
huh, they are using ObjectSpace._id2ref
<mengu>
is there a jrubyish way of what they're doing?
<lopex>
enable it
<lopex>
that message is confusing
<lopex>
it should read "Since ObjectSpace is disabled.."
<lopex>
headius: ^^
<lopex>
so true/false/nil/1 are immediates in mris nomenclature
<headius>
mengu_: lovely
<headius>
they should not be using id2ref
<headius>
nobody should ever use id2ref
<lopex>
that's another story
<headius>
enabling objectspace will slow your app to a crawl on JRuby
<headius>
why does this project have no bug tracker
<mengu>
i am not going to enable objectspace, no worries
<mengu>
instead i'll get rid of the newrelic rpm gem
<mengu>
after switching to connection pooling and jndi
<mengu>
somewhere, my app is choking
<mengu>
no response from tomcat at all
<mengu>
just times out
<headius>
get a thread dump, we can see where it's stuck
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<mengu>
ahem, if i may, how do i do that when the app runs as war on tomcat?