lucasb has quit [Quit: Connection closed for inactivity]
ur5us has quit [Ping timeout: 260 seconds]
nirvdrum has quit [Ping timeout: 260 seconds]
ur5us has joined #jruby
nirvdrum has joined #jruby
_whitelogger has joined #jruby
jacaetevha[m] has joined #jruby
<jacaetevha[m]>
Anyone here using JRuby in a Maven project? We have a few and builds have been failing because the Torquebox repo is down (apparently, permanently)
<jacaetevha[m]>
Need to find or create a replacement
jacaetevha[m] is now known as JasonRogers[m]
<headius[m]>
Ahh hello
<headius[m]>
How are you using that repo?
<headius[m]>
It has been EOL for a couple weeks now I think
ur5us has quit [Ping timeout: 260 seconds]
nirvdrum has quit [Ping timeout: 268 seconds]
<rwilliams[m]>
@headius:matrix.org: If theine is going to be the preferred method for speeding up start should it fall into the Jruby umbrella or projects? I gotta say it was a little frustrating (and fun) investigating all the ways to speed up startup time only to find that 1/3 worked and even the one that worked wasn't actively maintained and needed to be updated.
<rwilliams[m]>
I'm "maintaining" theine2 now but I imagine somebody can get ahold of Jan, I haven't tried emailing him, and get push access to Theine rubygems.
<rwilliams[m]>
This is for another discussion and I've brought this up before but Jan basically rewrote Theine into a fork of spring and was super close to getting it merged and it just died. I tried merging his fork into the Spring master but the tests failed spectacularly and its definitely way above my skill set.
<rwilliams[m]>
@jacaetevha:matrix.org: Didn't torque box die quite awhile ago(years). I remember being so excited for torquebox 4 and it just never materialized.
ur5us has joined #jruby
nirvdrum has joined #jruby
nirvdrum has quit [Ping timeout: 260 seconds]
ur5us has quit [Ping timeout: 260 seconds]
drbobbeaty has joined #jruby
nirvdrum has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
nirvdrum has quit [Ping timeout: 240 seconds]
nirvdrum has joined #jruby
shellac has joined #jruby
nirvdrum has quit [Ping timeout: 260 seconds]
drbobbeaty has joined #jruby
<rwilliams[m]>
So Jan got right back to me. I probably should have emailed him sooner before publishing Theine2
shellac has quit [Quit: Computer has gone to sleep.]
nirvdrum has joined #jruby
<headius[m]>
Yeah that happens 😃
<headius[m]>
It would be amazing if we could get this merged into spring proper, so everyone has it out of the box when they generate a rails app
shellac has joined #jruby
<rwilliams[m]>
Jan also offered some insight on improving theine https://github.com/rwilliams/theine2/issues/8 I guess he put in some PRs(since merged) to JRuby that would allow dropping screen as a requirement. I'm going to take a stab at it but it's definitely gonna take a good bit of learning on my part.
<rwilliams[m]>
I think his later work in the jruby spring fork may have implemented this
<nirvdrum>
headius[m]: You probably have it somewhere on your backlog, but it looks like there's been some activity in jnr-unixsocket that could use some attention.
<nirvdrum>
I don't use that particular JNR project for anything, but I'm happy to help out anyway that I can.
<headius[m]>
yeah I cycle through them periodically but they're on a long list of priorities
<headius[m]>
some days I have to spend debugging why github actions broke us
<headius[m]>
🙄
<nirvdrum>
I still have permissions on the project, but I didn't want to start merging things out of the blue. I'll try to triage the backlog.
<headius[m]>
if things look good I'd say go ahead
<headius[m]>
I can take a look too in a minute
<headius[m]>
I need to figure out a more reliable way to get GH notifications because email isn't it
<nirvdrum>
Okay. I'm not trying to push anything ahead. I just started reading through my growing backlog of GitHub notifications and saw that one.
<nirvdrum>
I'd need to freshen up on some of the particulars of domain sockets to really evaluate the test case. But, I'm happy the author wrote a test.
<headius[m]>
yeah seems like a great one to merge
xardion has quit [Remote host closed the connection]
xardion has joined #jruby
<JasonRogers[m]>
<rwilliams[m] "@jacaetevha:matrix.org: Didn't t"> Yeah, Torquebox died, but the Maven repo has been running for some time. My team came back from the holidays, started working on stories that depended on the repo, and boom ... all gone.
<JasonRogers[m]>
<headius[m] "It has been EOL for a couple wee"> Yeah, it hit us unexpectedly after folks starting working on stories for these projects after returning from holiday breaks.
<JasonRogers[m]>
There is rubygems.lasagna.io, but I'm not sure if that's meant to be public or not.
<headius[m]>
Jason Rogers: ugh sorry about that
<headius[m]>
they pinged us before they took down the server but none of our projects were using it anymore
<headius[m]>
I didn't figure there'd be downstream users
<headius[m]>
enebo: who was it that pinged about that server?
<headius[m]>
Jason Rogers: can you use the mavengems plugin that we use in the JRuby build?
<headius[m]>
nirvdrum: all PRs are addressed...if you don't have anything else you want in I can spin a release
<nirvdrum>
headius[m]: Cool. I actually didn't need anything. I was just responding to a GitHub notification.
<headius[m]>
email notification?
<JasonRogers[m]>
<headius[m] "Jason Rogers: can you use the ma"> I'll check it out. Thanks!
<nirvdrum>
Yeah, email notification.
joast has quit [Quit: Leaving.]
lucasb has joined #jruby
<enebo[m]>
headius: I don't understand what you are asking?
<headius[m]>
apparently downstream users like Jason Rogers were still using that torquebox rubygems server
<headius[m]>
so we're getting reports now about it being down
<headius[m]>
I don't remember who asked you whether they could shut it down
<enebo[m]>
ben pointed out it had filled up on disk and they wanted to take it down
<headius[m]>
enebo: so I'm not sure if we should ask for it to be restored or try to find a permanent home fo rit
<enebo[m]>
headius: I do not know very much about the server in the first place. I thought it was a proxy for finding maven artifacts as gems...
<enebo[m]>
and lasagna replaced that
joast has joined #jruby
<headius[m]>
For some reason I don't remember anything about lasagna
<headius[m]>
Seems like we need to figure out a long-term solution here
<headius[m]>
JasonRogers[m] it basically plugs into Maven and makes rubygems.org look like a maven repository
<enebo[m]>
What is actually being pushed? A gem? An artifact?
<headius[m]>
That's definitely the preference going forward, but I know non Maven builds like Gradle can't use it
<enebo[m]>
I assume this is just a way of gemifying existing artifacts
<enebo[m]>
but as I said I did not really know much about this in the first place
<headius[m]>
Well it basically just proxied ruby gems and provided a maven interface, right?
<enebo[m]>
oh it is the reverse?
<enebo[m]>
haha then I really didn't know
<headius[m]>
Clearly! 😃
<enebo[m]>
so it allows gems to appear as artifacts?
<headius[m]>
Yeah, I believe it's what we were using before the mavengem plugin
<rwilliams[m]>
Im such an ass, I thought they were still using torquebox 3 or something
<enebo[m]>
rwilliams: is using mavengem locally an option near term?
<enebo[m]>
I am trying to wrap my head around what was happening
<enebo[m]>
As someone maintaining some amount of infra for JRuby stuff I think we definitely do not want to be a datasource for published things (at least not long term)
shellac has quit [Quit: Computer has gone to sleep.]
<enebo[m]>
If I get the real value of mavengem and the notion of representing gems as artifacts it is not wanting to double source the same thing
<enebo[m]>
Since one can publish a rubygem to rubygems.org and then also publish it packaged as a maven artifact
<enebo[m]>
We actually do this in a sense for JRuby itself with jruby-jars gem + the jruby artifacts
<enebo[m]>
I also suspect publishing to maven is a pain in the ass compared to pushing a gem which is a second value add to that idea
<enebo[m]>
What torquebox was running was working but I am fairly certain it was a single machine and half suspect the backup/recovery was not a first class priority
<enebo[m]>
as such it maybe was an accident waiting to happen
<enebo[m]>
I feel bad about this and I admit I did not take much interest here (to the point of not fully understanding what it was)
<headius[m]>
It was JasonRogers[m] that mentioned his project was using that server
<JasonRogers[m]>
Using mavengem works for us
<headius[m]>
The value of the server is that it looks like every other Maven repository so you don't need a plug-in
<JasonRogers[m]>
And yes, it's meant to represent Gems as Maven artifacts.
<headius[m]>
JasonRogers[m] okay that's great
<headius[m]>
There may still be value in having a server that doesn't require a plug-in, but it should be somewhere more permanent anyway
<JasonRogers[m]>
But, having a server that doesn't require a plug-in seems like a real nicety
<headius[m]>
Yeah
<JasonRogers[m]>
(love the move away from Gitter, by the way)
<JasonRogers[m]>
Oh, umm.... just realized we have some Gradle-based projects that depend on gems too. Not sure how we're going to update them.
<headius[m]>
What do we know about that lasagna server?
<rwilliams[m]>
Speaking of gitter and riot are any well known tech communities using discord?
<rwilliams[m]>
Disregard
<enebo[m]>
Thinking outloud but could mavengem has a daemon mode which would install and then be runnable on each machine which will also do artifact resolution...possible updating .m2 settings for that
<enebo[m]>
Perhaps that is too hinky and error-prone
<headius[m]>
Well that would basically be the server we killed
<enebo[m]>
but it would run on everyones machine
<enebo[m]>
not just one normative one
<headius[m]>
If we could figure out the best way to run that server we could at least provide instructions
<headius[m]>
Ok yeah
<enebo[m]>
it would set localhost as a mvn repo
<headius[m]>
rwilliams we considered discord but really wanted to go with something 100% open
<headius[m]>
enebo I would think most organizations would prefer setting up an in-house machine rather than having builds depend on a local proxy
<enebo[m]>
I am happy riot seems to have addressed some stability/perf issues in last month
<headius[m]>
The issue here is that you have to put something into a pom file
<headius[m]>
Or the really difficult case, into your Gradle build which can't use the plugin
<enebo[m]>
headius: no doubt many would...just thinking it could be an alternative for people who don't want to maintain a server
<enebo[m]>
since it is just a proxy and contains no normative data
<headius[m]>
Yeah, but you would have to have localhost as a maven repository in your build
<enebo[m]>
ah yeah it would need to be in the pom wouldn't it...I was thinking along the lines of .m2/settings
<headius[m]>
I think you can set a bunch of default repositories, but localhost would seem very strange there too
<enebo[m]>
ok well it was just a thought and I will not continue on the trend of passing env vars :)
<headius[m]>
The ideal case would be getting rubygems.org to host a maven proxy
<JasonRogers[m]>
That would be really great!
<enebo[m]>
yeah it also occurred to me as well but I wonder what that sell would be like :)
<headius[m]>
Tough sell for sure
<JasonRogers[m]>
Jah
<headius[m]>
Maybe it's simple enough we could run it on some free tier cloud instance
<headius[m]>
Or get heroku to donate a server or something
<JasonRogers[m]>
I wonder if it could be run with an S3 file system. I'd be happy to donate an EC2 server, as a way of giving back to the community.
<enebo[m]>
is it simple enough to fit on a free dyno? :)
<headius[m]>
JasonRogers[m] thank you, we appreciate that. Hopefully we can get someone already hosting Ruby stuff to cut us off a slice
<JasonRogers[m]>
See https://matrix.to/#/!vyEDBdbmzqApWaugdd:matrix.org/$1578591834228055qfsyv:matrix.org?via=matrix.org&via=sibnsk.net&via=hackerspaces.be for how we changed our POM to make things work
<cshupp[m]>
Thanks!
<JasonRogers[m]>
my pleasure
<cshupp[m]>
I guess my local repo has hidden this from me
<headius[m]>
enebo: ^
<headius[m]>
I guess quite a few folks were still using the server ☹️ Sorry for the inconvenience, I think we need to get something back up
<headius[m]>
oh lasagna.io was Tyler's server right?
<headius[m]>
specifically to support jruby-gradle
<JasonRogers[m]>
It is Tyler's server, as far as I know
<headius[m]>
yeah
<headius[m]>
it's all coming back to me now
<headius[m]>
so at least he has a server and knows how to run it
<headius[m]>
I just pinged him about it
<headius[m]>
I'm going to lunch and errands and will be back in an hour or so
<cshupp[m]>
Wait you mean to tell me by chance you were just talking about this?
<JasonRogers[m]>
Yes ... it's all the rage. :) All the cool kids are ... nevermind
<cshupp[m]>
And I am speaking of the openssl error breaking the build
rtyler has joined #jruby
<enebo[m]>
news to me. Assuming it works on 8 then perhaps it is some module thing? Make an issue for it please.
<rtyler>
I have been summoned by the ghost of torquebox past
* rtyler
waves to headius[m]
<rwilliams[m]>
Lol
<cshupp[m]>
JRuby issue?
<rtyler>
headius[m]: do you all rely on any rubygems-servlets proxies? For JRuby/Gradle, we're preparing a 2.0 release which no longer relies upon proxies
nirvdrum has joined #jruby
<headius[m]>
rtyler: hi there
<headius[m]>
that's good news but we're debating whether it might be good to have a permanent proxy
<rtyler>
if you have somebody to pay for the hosting and bandwidth, it's a really simple stateless service to operate
<headius[m]>
yeah we were thinking of either trying to get rubygems.org to host it (unlikely) or someone like Heroku
<headius[m]>
I suppose I can blast out a tweet and see if anyone jumps on it
<headius[m]>
can you link me a write-up on how to?
<rtyler>
I already brought the topic up with the rubygems folks, they're not terribly keen on it