shellac has quit [Quit: Computer has gone to sleep.]
rtyler has quit [Ping timeout: 268 seconds]
rtyler has joined #jruby
drbobbeaty has quit [Ping timeout: 260 seconds]
shellac has joined #jruby
<rwilliams[m]>
headius: I've linked this before but Jan definitely did all the work to remove screen as a requirement for Theine via his work on the JRuby fork of Spring. I should be able to backport these changes into Theine. https://github.com/rails/spring/pull/449
shellac has quit [Quit: Computer has gone to sleep.]
shellac has joined #jruby
<kares[m]>
have been trying to use semmatic commit message prefixes
<kares[m]>
but tagged in e.g. "[refactor] ..." instead of "refactor: ..."
<kares[m]>
these are pretty standard but I did not put much thinking into when I want to use a specific tag e.g. "[ji] ..." do we rather have 2 tags like: "[fix] [ir] raise not clearing" ?
<kares[m]>
or "[fix] core: ..." vs. "fix: [ji] ..."
drbobbeaty has joined #jruby
rusk has quit [Read error: Connection reset by peer]
rusk has joined #jruby
lucasb has joined #jruby
nirvdrum has quit [Remote host closed the connection]
<headius[m]>
kares: yeah and first line of commit is already pretty short
<headius[m]>
thanks for the link
nirvdrum has joined #jruby
drbobbeaty has quit [Read error: Connection reset by peer]
drbobbeaty has joined #jruby
<enebo[m]>
I also use [] as a prefix I believe because someone added [noci] logic
<headius[m]>
travis has [skip ci] yeah
xardion has quit [Remote host closed the connection]
xardion has joined #jruby
rusk has quit [Remote host closed the connection]
nirvdrum has quit [Ping timeout: 268 seconds]
shellac has quit [Ping timeout: 268 seconds]
sagax has quit [Ping timeout: 240 seconds]
<headius[m]>
rtyler: your lasgana.io going to be up indefinitely?
<headius[m]>
I don't want to make you host a permanent option but at least we could direct people there until we have one
nirvdrum has joined #jruby
<rtyler>
it's literally not up
<rtyler>
it hasn't been for a year or two actually, I got tired of personally donating money to randos :)
<headius[m]>
Oh, okay. So at the moment, we have no rubygems proxy at all
<headius[m]>
enebo maybe we should just start an instance up under a jruby hostname and swallow the cost for the moment
<headius[m]>
Actually, I still have my IRC bouncer on a little arm instance, that would probably be sufficient
<JasonRogers[m]>
<rtyler "it hasn't been for a year or two"> Any idea of the estimated cost per month?
<headius[m]>
my arm box is a whopping €6/mo
<headius[m]>
hmm and apparently hung
<headius[m]>
enebo: does EY still run anything for us? I think the S3 is still out there?
<rtyler>
heh, it's not a lot of cost, I think it was 20-30 bucks a month
<rtyler>
the proxy doesn't proxy gem binaries, it does a redirect
<enebo[m]>
headius: s3 for windows executables
sagax has joined #jruby
<enebo[m]>
Jason Rogers: can you see if rubygem-proxy.torquebox thing happens to be up atm?
<enebo[m]>
I believe the server was brought back up with the logs nuked but it is tough to know if it is working
<JasonRogers[m]>
sure, one sec
<JasonRogers[m]>
enebo: I do get a response.
<enebo[m]>
Jason Rogers: that's encouraging
<rtyler>
it looks like it's back to normal, it still runs over HTTP however :)
<JasonRogers[m]>
right
<enebo[m]>
alright well that is a development
<JasonRogers[m]>
I think the `mavengem` approach is still better than that. Though, I think having a secure proxy would be super great.
<JasonRogers[m]>
I'm happy to help out where I can.
<enebo[m]>
well if we can leave this up for a while we can get some breathing room on how to handle any recommendation or migration
<JasonRogers[m]>
agreed
<enebo[m]>
http is perhaps the stick to encourage people to setup their own mavengem in places where that is reasonable
<enebo[m]>
for larger public uses perhaps we still need some deployment
<rtyler>
who still relies on this outside of the JRuby/Gradle community?
<JasonRogers[m]>
Jah ... And, it may help to further the concept of JRuby being a first-class citizen and not some orphan child project. (though, I'm not positive that is still a thing out in the wild these days)
<JasonRogers[m]>
rtyler: many POMs rely on it for gem deps. when embedding the JRuby engine inside another JVM language (for us, it's used in several legacy projects)
<rtyler>
ahh
<enebo[m]>
ok so I got word this server can stay up for a while without worrying about it too much but we need to maybe try and get ruby central to help us out here
* rtyler
wonders what this would cost if run on Lambda
<enebo[m]>
we were able to get ruby-lang to give us a mailing list which is not ruby central but we are basically representing published gems in a way so that they can get wider consumption
<enebo[m]>
maybe that is a good enough reason. I doubt it will be much load
<enebo[m]>
although I have no idea how many people use it obviously :P
<rtyler>
if ruby-central has a Kubernetes cluster or some thing that can run a container, then the overhead for them could be relatively low
* rtyler
shrugs
<enebo[m]>
ok well that would be a simple push ... it's just a container
<enebo[m]>
I was also wondering if this could potentially be put into rubygems server itself
<enebo[m]>
I no nothing of what is in that code but mostly it is just a template generating a pom isn't it?
<rtyler>
it definitely could, I haven't had the time to work on that
<JasonRogers[m]>
That would be ideal, I think. But, I think "they" aren't in favor of it.
<enebo[m]>
well times change but yeah I figure it will not be a simple sell
<enebo[m]>
perhaps the strongest argument is how much does maven change :)
<enebo[m]>
one small set/single route with a simple controller....
<enebo[m]>
I think a big push would just be a PR but that depends on how much work that would be
<enebo[m]>
If it ends up being a small benign PR with author(s) willing to support then it is an accept or tell someone why the work they did is too much of a burden
<JasonRogers[m]>
sounds reasonable
<enebo[m]>
We always have this struggle of being Ruby+. At some level we want to be zero overhead to the main Ruby community but that obviously is impossible
* rtyler
nods
<enebo[m]>
ok well I did not realize and I should have this is a servlet which calls a Java maven artifact for ruby sonatype tools
<enebo[m]>
so in order for what I said above to work we would need rubygems.org to switch to run with JRuby (they won't) or we would have to port all the logic from that artifact and it's dependent artifacts to ruby
<enebo[m]>
That is probably much too high of a task
<headius[m]>
not to mention the plugins probably reference other maven stuff all over the place to build up a proper model
<headius[m]>
doesn't seem feasible
<enebo[m]>
yeah it does
<enebo[m]>
ruby sona tools is a wrapper of sonatype tools
<headius[m]>
at the moment I don't see any option other than running a java server somewhere
<enebo[m]>
so it would mean converting the whole sonatype stack
<headius[m]>
and I highly doubt we'll get rg.org to run Java
<enebo[m]>
I have already conceded
<headius[m]>
I don't want to host it either but nobody has stepped up except Jason Rogers
<enebo[m]>
Unless they are heavily container-based then they may "host" us but we would still maintain it
<enebo[m]>
we should just find the cheapest minecraft server option and run it as a bukkit plugin
<headius[m]>
hah
<headius[m]>
well like I mentioned, my scaleway arm instance is €6 flat rate
<headius[m]>
if we agree that we're not going to be able to get this to run in rg.org then I think we should set up an instance under *.jruby.org for now and keep begging for a host
<headius[m]>
it will be the same service regardless of who hosts it eventually but we could at least get a permanent hostname
<JasonRogers[m]>
I'm in favor of that motion
<rtyler>
oh snap, are we about to get parlimentarian up in here
<headius[m]>
the floor has not recognized the delegate from California
<rwilliams[m]>
Lol
<enebo[m]>
What CNAME do we want?
<JasonRogers[m]>
hah
<JasonRogers[m]>
mavengems.jruby.org ??
<headius[m]>
that's better than I came up with
<enebo[m]>
sounds good
<headius[m]>
once it goes to https we'll have to deal with moving hostnames around but I think we can just stick any cname on it we want right now
<enebo[m]>
yeah should be fine for http
<headius[m]>
mavengems.jruby.org has my vote
<enebo[m]>
yeah looks just fine to me
<headius[m]>
just point that at the TB server until we decide next steps?
<rtyler>
that's not a bad start
<enebo[m]>
what is rubygems.tb addr?
<headius[m]>
my instants has 41GB but I don't know how much space this thing eventually consumes
<enebo[m]>
or hmm hahah I don't even remember if it is address or host for CNAME
<headius[m]>
rubygems-proxy
<rtyler>
there is not much disk space that's consumed, the proxy just stores XML caches locally
<headius[m]>
I'm fine running it but we know the tb server is solid, so as long as that's up let's point at it
<enebo[m]>
added ... not sure how long it will take to see it or whether I made a mistake or not
<rtyler>
I wonder how much bandwidth this thing takes up, I have gigabit at home and have already been running a few services ^_^
<enebo[m]>
mavengems.jruby.orgcanonical name = rubygems-proxy.torquebox.org.
<enebo[m]>
Name:rubygems-proxy.torquebox.org
<enebo[m]>
Address: 199.193.199.40
<headius[m]>
yeah it's not like we have thousands of people hitting this constantly, and maven caches stuff locally
<enebo[m]>
my machine seems to see something via nslookup
<enebo[m]>
snappy
<headius[m]>
Jason Rogers: can you give that new name a shot?
<JasonRogers[m]>
sweet
<JasonRogers[m]>
headius: I need to set up another test bed ... uno momento
<JasonRogers[m]>
Are you seeing traffic @headius?
<enebo[m]>
Jason Rogers: this is only a CNAME which is hitting the torquebox site
<enebo[m]>
but it has the advantage if you change to this we can repoint to a different server later and you won't go down
<JasonRogers[m]>
well, I've already gone through the process to switch over all of our projects to use `mavengem` ... so, I'm doing this in a test bed project
<JasonRogers[m]>
but, looks like it was successful
<enebo[m]>
ok well in your case you maybe should be but we have a plan here for eventual migration for public usage and you helped us!
<enebo[m]>
you even came up with the name
<JasonRogers[m]>
hmm... I think perhaps it pulled cached resources from our internal Artifactory
<JasonRogers[m]>
one sec
<headius[m]>
we'll figure out whether we're hosting it or someone else is, but it will always be that address
<JasonRogers[m]>
something's wonky ... probably my test setup
<headius[m]>
if cname doesn't work we could just alias to that IP
<headius[m]>
but for http it seems like it should be fine
<enebo[m]>
ah it does not work
<enebo[m]>
It ends up probably getting tripped up with virtual hosting
<enebo[m]>
so it does not match and another website on same web server is getting it?
<headius[m]>
ugh
<headius[m]>
yeah
<JasonRogers[m]>
OK. So, it's not my test project. :/
<enebo[m]>
I am going to ask to see if we can add it...should be a 1-line config change
<JasonRogers[m]>
Well, I have a release to finish putting together in order to get it out Monday morning. I need to focus on that for a while.
<enebo[m]>
but I don't know how huge a request that is
<headius[m]>
Jason Rogers: no worries, thanks
<enebo[m]>
Jason Rogers: yeah no problem. We will figure out what is up here
<headius[m]>
maybe I'll be running it after all
<enebo[m]>
final note today on mavengems cname...someone on monday will look into fixing the virtual host part
ARTSIOM has joined #jruby
ARTSIOM has quit [Client Quit]
nirvdrum has quit [Ping timeout: 268 seconds]
lucasb has quit [Quit: Connection closed for inactivity]