drbobbeaty has quit [Read error: Connection reset by peer]
<headius>
enebo: so much of stdlib is in gems now I'm going to bump minor compat version to 3.1.4
<headius>
handful of gems getting updated, PR on the way
<enebo[m]>
ok
<headius>
enebo: the debian:8 image I had used for glibc thing seems to be really dead now... the server with updates for debian jessie no longer hosts those packages
<headius>
so I can't fix any other platforms easily right now and will just spin the rest of jnr releases with what I have
<enebo[m]>
yeah do what is possible
<enebo[m]>
really feel there is a business oppty for cross-platform build tool
<headius>
for sure
<headius>
so many half baked solution
<headius>
only jffi has been released based on this fix so I'm doing the full stack now
<headius>
gem updates merged to master
<enebo[m]>
ok
<headius>
enebo: did you just delete jruby staging repo?
<headius>
oh nevermind it's still there
<headius>
whew just worried I accidentally released it
<enebo[m]>
that is not current either
<enebo[m]>
but I do just full build just in case
<enebo[m]>
you can drop that if you are logged in
<headius>
I am and I will
<headius>
jnr stuff is out and propagating
<headius>
enebo: oh, joni
<headius>
did you get all fixes in you wanted? I can release that too
<enebo[m]>
no but I did realize that the reported problem was only represented by an unusual | symbol
<enebo[m]>
it likely is more common than that but I do not think we hold up for it. We do need it fixed obviously
<enebo[m]>
I do not recall anything else on joni other than config commit but if you wanted that in for debugging we could update it?
<headius>
yeah I think we should go ahead with update
<headius>
let me see what else is in there
<headius>
enebo: oh yeah the region stuff is in this
<headius>
so it will use more efficient packed object for single region matchers
<headius>
and hides all the internals of region
<enebo[m]>
oh yeah. did you ever observe is being faster?
<headius>
I did not try
<enebo[m]>
but encapsulation I think is a goal in itself
<headius>
maybe simple as benchmarking a single group match?
<enebo[m]>
but not very important for JRuby tomorrow
<headius>
so you want it or not
<headius>
it is more configs and region improvements
<enebo[m]>
A work-around would be to use FAST for case sensitire but SLOW for case-insensitive
<enebo[m]>
So proposed work-around is change Search.BM_IC_FORWARD to Search.SLOW_IC_SB_FORWARD in setOptimizeExactInfo.
<enebo[m]>
I will do it in the morning if you do not do it but this will stop the now two reported issues at the cost that their perf can be a lot slower
<enebo[m]>
This is not a solution but will allow us to work until we figure out problem with the case insensitive tables
drbobbeaty_ has quit [Remote host closed the connection]