<kares[m]>
the id2ref came from a library hooking up a finalizer and I needed a way of undoing the finalizer before terminating the runtime
<kares[m]>
but since I changed how the library works so I have no strong feeling about shipping the id2ref patch - removed it from the 9.3.5 milestone
drbobbeaty_ has joined #jruby
drbobbeaty has quit [Ping timeout: 244 seconds]
<enebo[m]>
kares: ok cool. I think headius maybe said it best 'because I'd rather not expose more id2ref cases of we don't have to'
<enebo[m]>
Not that we shouldn't consider the PR but I also agree with that assessment
<kares[m]>
okay understood - let's see if the need pops up anytime soon, hopefully not!
<kares[m]>
also thought that it's slightly an edgy and weird use-case ...
<kares[m]>
fyi: I am into my last 2 weeks working at Elastic, than I will be off for some months taking a break.
<kares[m]>
just in case I forgot to mention that and I end up being non-responsive 😉
<enebo[m]>
kares: oh thanks for the update. hopefully this summer it will just be finalizing 9.4 finally
<enebo[m]>
I am also hoping 9.3.5.0 is enough for elastic.
<enebo[m]>
Our release stride has been a bit slower compared to 9.2 but it feels like the reports have not quite been as many (Although unfortunately things like the string dedup are pretty big problems)
<kares[m]>
enebo[m]: more than enough esp. given we (logstash) did not end up testing the 9.3 early on despite me trying to do so ...
<kares[m]>
9.3 looks good - couple JI regressions but I guess only heavy JI users, such as Logstash notices these ... atm the upgrade is postponed till 9.3.5 is out
<enebo[m]>
ah ok. cool.
<enebo[m]>
I think the new 9.4 kwargs branch will be done today
<enebo[m]>
kares: ractor is not planned for 9.4.0.0. Looking online I cannot find anyone who actually uses it. Mostly entries of people trying it but it not really working out well enough to be used...yet
<enebo[m]>
I think one of the larger problems is ractors do not interact well with existing library code too
<enebo[m]>
but also people expected to see some big perf advantage over a GIL and threads and seemingly there is not much gain
<enebo[m]>
The notion of safe isolation does not seem to be a big driving factor
<kares[m]>
good, honestly I did not get the point and it left me wondering if doing it in core straight away wout feedback was worth it ...
<kares[m]>
but what do I know, not a big fan of refinements either, which I've seen used on a project/library once, so far
<enebo[m]>
yeah refinements are being used a bit more now but it has been a very slow road
<enebo[m]>
I will reserve judgement on ractor but right now it is not at a place where it is getting adopted so we won't sweat it.
<byteit101[m]>
headius: should we try to fix the python issue, or just pre-compile it?
<headius>
byteit101: there's got to be a way to get it to recognize Python 3 in GHA, I will try a few things today too
<headius>
enebo: anything else to look at for 9.2 or 9.3?
<byteit101[m]>
oh god am I cursed both at work and personal projects now?!
<headius>
maybe something healed in GHA wrt python
<headius>
heh oops
<headius>
or it just wiped out the entire site
<headius>
well that's weird... gh-pages branch is empty?
<byteit101[m]>
I was messing with the deploy location, may want to undo that?
<byteit101[m]>
or maybe you have to move what was gh-pages to a new master, and it builds gh-pages?
<byteit101[m]>
also you can remove "list pythons"
<headius>
oh yes, so this is pushing a static build to gh-pages what was there before as a dynamic site
<headius>
but it is still empty
<headius>
I pushed the old gh-pages as master
<headius>
byteit101: I am going to add you as a committer so you can try to take it from here again
<headius>
my PR got borked because of the branch changes but all it did was copy your HEAD workflow over
<headius>
master is now where sources go and the build will deploy gh-pages as you were working on... try it again and maybe it will just work now
<byteit101[m]>
cool, I can't do anything rn, will look at it tomorrow evening if it's not fixed by then
<headius>
no problem
<headius>
perhaps it is just not including the generated site in the commit
<headius>
oh hah...now it failed this time
<headius>
jekyll 3.9.2 | Error: Broken pipe: env: can't execute 'python3': No such file or directory
<headius>
byteit101: so the python3 error is not from GHA but from the docker image that helaili/jekyll-action runs with
<byteit101[m]>
AAA!
<byteit101[m]>
[GitHub] Subscribed to 17 jruby repositories. You’ll receive notifications for all issues, pull requests, and comments that happen inside the repository.
<byteit101[m]>
I think I will disable most of those notifications
<byteit101[m]>
uh, so what does that mean? is that something to file an issue on the helaili/jekyll-actions repo? I've not used GHA or much docker before
<headius>
Welcome to the party pal
<headius>
I'm not sure. The docker image they use may just need to be updated to have Python 3 since there will be more and more plugins that want it
<byteit101[m]>
So the "list python versions" and the h/j-a run in separate contexts?
<headius>
Or we find a different way to build and publish
* byteit101[m]
looks at frequency of posts and think static builds seem reasonable