<enebo>
etehtsea: I believe you it does but I am confused how StructLayout works :)
<enebo>
etehtsea: you mentioned we only test size in one platform. Could I trouble you to make a general test for size as well. It would some more confidence in alignment of stat on various platforms
<enebo>
etehtsea: thanks for updating the PR it looks ok to me and with some basic tests like you linked to I feel we will catch an issue if something is wrong
<enebo>
etehtsea: the fact you did the C sizeof thing also builds some confidence as well
<GitHub103>
[jruby] enebo closed issue #4600: Dir.tmpdir fails with error on AARCH64 https://git.io/v91SY
<enebo>
weirdly there are two atime failures on travis but those preceed your PR so who knows
subbu|lunch is now known as subbu
<etehtsea>
enebo the same specs are failing on my macos host
<enebo>
yeah I tracked this down to a PR merged a couple of months ago for nsec support
<enebo>
It may be a bad test or not :)
<enebo>
Not related to your stuff though
<enebo>
We will get this addressed for release though
etehtsea has quit [Ping timeout: 240 seconds]
donV has joined #jruby
<GitHub169>
[jruby] enebo pushed 2 new commits to jruby-1_7: https://git.io/v9yF1
<GitHub169>
jruby/jruby-1_7 8cdb01a Thomas E. Enebo: nuke .dev from pom.xml in finalName.
<GitHub169>
jruby/jruby-1_7 c831ac1 Thomas E. Enebo: Bump for 1.7.27.
<GitHub31>
jruby/master 7f6e378 Thomas E. Enebo: Fixes #4586. JRuby profiler raises exception when using the holidays gem
<GitHub76>
[jruby] enebo closed issue #4586: Jruby profiler raises exception when using the holidays gem https://git.io/v9EMS
<donV>
Hi all!
<nirvdrum>
Aethenelle: Did you find you needed to make that zipgroupfileset change then?
<Aethenelle>
that was to get the binaries into the main artifact?
<nirvdrum>
I think so.
<nirvdrum>
I just couldn't get the build file as written to work. You said you could, but I'm curious if it's just because you were on old version of the repo.
<nirvdrum>
I.e., did you make the change because I requested it or because you ultimately found it necessary?
<nirvdrum>
Aethenelle: The issue is I regenerated the constants on x86_64 Linux and get different values.
<nirvdrum>
The new values seem to match MRI.
<Aethenelle>
yeah... I typed that before I realized you were already there...
<nirvdrum>
I'm not sure if this a 32-bit VS 64-bit thing or values have just changed over time.
<Aethenelle>
there's a bunch of constants in places that aren't great
<nirvdrum>
And IIRC, I get a different set of values on SPARC. jnr-constants isn't segregated by platform currently, so that's going to be a bit messy to unwind I suspect.
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
hobodave has quit [Ping timeout: 240 seconds]
Aethenelle has quit [Quit: Aethenelle]
<headius>
nirvdrum: so the 64-bit values may be right then
<headius>
I know solaris had a few different sets of constants for 32 vs 64 on sparc and x86
<nirvdrum>
Unfortunately I don't have a 32-bit Linux machine readily available.
<nirvdrum>
I don't know if the constants were generated on 32-bit or not.
<headius>
yeah I'm not sure either, but it should be easy enough to get a VM up
donV has quit [Quit: donV]
<lopex>
headius: so monday is a release day ?
<lopex>
enebo: wrt that symbol patch, do java strings appear anywhere inbetween in the runtime ?