UtkarshRay has joined #rubygems
swills has joined #rubygems
swills has quit [Remote host closed the connection]
ur5us has quit [Remote host closed the connection]
_djbkd has quit [Quit: My people need me...]
UtkarshRay has quit [Remote host closed the connection]
ur5us has joined #rubygems
laith has joined #rubygems
UtkarshRay has joined #rubygems
laith has quit [Quit: Page closed]
ur5us has quit [Remote host closed the connection]
jottr has joined #rubygems
<jottr> Hi all. Could someone be so kind and explain to me what `gem pristine --all` does? I'm not sure I properly understand the explanation given at the offcial reference: "Restores installed gems to pristine condition from files located in the gem cache
<jottr> "
<jottr> What IS the pristine condition?
<mat_> jottr, I guess you hacked your installed files, and want them de-hacked
<jottr> mat_: hehe. I actually realized they go more into depth on the reference page. I should've scrolled down earlier. :)
<raggi> jottr: it just re-runs install for each gem specified
<raggi> jottr: and install rm's gem dirs if they pre-exist
<jottr> raggi: That's a good explanation. :)
<raggi> in some ways it's less useful than it used to be
<raggi> i used to use it all the time
<raggi> pre bundler, pre git, and when i used to deal wiht native library dependency issues on freebsd and windows
<jottr> Hm. So can I force the pristine command? I.e. `gem pristine json` returns "Skipped json-1.8.1, it is a default gem"
<raggi> heh
<jottr> There's no --force flag. Alternative would be to simply reinstall the gem right?
<raggi> no
<raggi> those are MRI packaged gems
<raggi> kind of a clusterfuck tbh, i never liked where we ended up
<jottr> ^^ What does that even mean? "MRI packaged gems"?
<raggi> ruby-core always had problems keeping their extra/contrib libraries up to date and managed independently of the main interpreter code
<raggi> many years after rubygems "won" the ruby packaging .. thing
<jottr> Hm. So json is a default gem. So it's impossible to uninstall it, right?
<raggi> ruby-core decided one way that might solve contrib library management challenges was to package them as gems
<raggi> but, what we ended up with is gems that are not gems, and it rears it's head in strange ways, such as this
<jottr> ...because of the native extensions?
<raggi> they're gems that you artificially can't uninstall because they're part of the interpreters public api
<jottr> ok
<raggi> and so removing them would mean you no longer have "the ruby standard library"
<raggi> still, they are gems enough that if you wanted to, you can manually pristine it with some tricks
<jottr> Ok. Maybe it'll make sense for me to actually describe my problem:
<raggi> yeah
<raggi> so
<raggi> gem pristine doesn't operate on bundlers vendor dirs
<raggi> `bundle exec gem prinstine` /might/ work, it once did, things may have changed
<raggi> but if it doesn't, just `rm -rf /Users/jottr/dev/mge/mge-rails/vendor/bundle/gems` then bundle again
<raggi> it looks like you switched ruby versions downward
<jottr> Yes.
<jottr> Can't recall why though...
<raggi> mixing vendoring and switching ruby versions is.. kinda evil
<raggi> porly documented errata
<raggi> anyway, that's fundamentally the source of the issue, these are gems stored in bundlers vendor path
<jottr> Hm. But that would mean I'm not allowed to switch ruby versions at all?
<raggi> you can, you just need to re-vendor
<jottr> Ah ok.
<raggi> bundler may have a new command for sorting this out
<raggi> did budnle exec gem pristine work?
<jottr> Nope. So I just did `rm -rf vendor/bundle/gems`
<jottr> `bundle install && bundle exec rake` results in:
<raggi> oh
<raggi> bleh, rm -rf vendor/bundle
<raggi> i forgot it had spec and cache dirs in there and stuff
<jottr> hehe
<raggi> technically you probably only needed ot do vendor/bundle/specifications
<raggi> but i can't remember the layout so well these days
<jottr> ...aaaand. Building therubyracer failed with building native extensions...
<raggi> heh
<raggi> oh god yeah, this whole v8 packed in a gem thing
<raggi> i remember this code
<jottr> I recall this https://github.com/cowboyd/therubyracer/issues/398#issuecomment-169063665 worked some time ago. But then it stopped working...
<jottr> Ok. So tried the same thing in above comment again:
<jottr> therubyracer built succefully, but:
* jottr headdesks
<jottr> My god. This issue is such a timesink...
<raggi> so i'm tired and really should be in bed
<raggi> so i'm not sure i'm game for debugging rubyracer right now
<jottr> raggi: nvm. thx for trying to help though. :)
<raggi> however, this looks like it's linked a v8 that is broken on your platform
<raggi> jottr: i'm goign to guess you're using homebrew
<jottr> raggi: yep
<raggi> brew update && brew install v8 then nuke you vendor again and bundle
<raggi> assuming there aren't actually some bugs in libv8/therubyracer this should fix stuff
<jottr> raggi: ok. will try. thx
<raggi> if all else fails, unless stuff changed i'm pretty sure sprockets or wahtever can use node instead of therubyracer
<raggi> so unless you use somehting that has a hard dependency on racer, just install node and remove racer, and be done wiht it
<raggi> remeber though: you may have it this way for deployments
<jottr> raggi: cool. I'll have to see which gems actually depend on rubyracer in that project.
<raggi> jottr: in most projects, it's just a runner for this: https://github.com/rails/execjs
<jottr> raggi: ah. good to know.
<raggi> people most often either cargo-cult in rubyracer or want/need it to avoid installing node/v8 in places where they do deployment builds
<raggi> execjs is a dependency of sprockets, which is part of the rails asset pipeline
<jottr> Ok. Well, the project will be deployed on heroku...
<raggi> if you look int eh execjs readme, it supports a bunch of runtimes
<raggi> so once upon a time you needed racer to build on heroku
<raggi> i think the cedar envs have node on them now, but you'll have ot try it and see
<raggi> (you can just use heroku shell or whatever it is to find out)
<jottr> So, I should just remove therubyracer from production: and move it into development: if at all...
<raggi> there ya go
<raggi> well
<raggi> in development
<raggi> just brew install node
<jottr> neat
<raggi> you'll find that breaks waaaay less often
<raggi> bedtime!
<jottr> raggi: thx for helping!
<raggi> np
<jottr> have a good one
* jottr nods
UtkarshRay has quit [Remote host closed the connection]
UtkarshRay has joined #rubygems
swills has joined #rubygems
relix has joined #rubygems
<relix> Hey, is it possible to add "--with-cppflags" to all gem install commands using .gemrc?
swills has quit [Ping timeout: 245 seconds]
jottr has quit [Ping timeout: 240 seconds]
jottr has joined #rubygems
imperator has joined #rubygems
swills has joined #rubygems
niska has quit [Ping timeout: 272 seconds]
<mat_> I have a question about the rubygems.org download CDN, FreeBSD currently uses http://production.s3.rubygems.org/gems/ and http://production.cf.rubygems.org/gems/ but it seems some newly released gems are not making it there, are those addresses legacy, or... ?
<mat_> poke drbrain, maybe ?
niska has joined #rubygems
c355e3b has quit [Ping timeout: 252 seconds]
cstrahan has quit [Ping timeout: 240 seconds]
davidjrice has quit [Ping timeout: 252 seconds]
ckrailo has quit [Ping timeout: 240 seconds]
ruurd has quit [Ping timeout: 250 seconds]
nz has quit [Ping timeout: 252 seconds]
fidothe has quit [Ping timeout: 252 seconds]
aquaranto has quit [Ping timeout: 252 seconds]
c355e3b has joined #rubygems
cstrahan has joined #rubygems
davidjrice has joined #rubygems
nz has joined #rubygems
aquaranto has joined #rubygems
fidothe has joined #rubygems
ckrailo has joined #rubygems
mjc__ has joined #rubygems
mjc__ is now known as mic_
mic_ is now known as mjc_
huoxito has joined #rubygems
relix has quit [Quit: Textual IRC Client: www.textualapp.com]
davidjrice has quit [Ping timeout: 240 seconds]
davidjrice has joined #rubygems
inev has joined #rubygems
ruurd has joined #rubygems
<inev> hi everyone. I'm having a few problems with the postgresql gem... i'm getting an incompatible library version, but i have no idea what could be wrong, i installed it using bundle, and its complaining about "activesupport-4.2.5"
<inev> (sirry if this is the wrong place to ask)
jottr has quit [Ping timeout: 252 seconds]
<imperator> inev, what OS, what version of rubygems, what version of postgres?
<drbrain> mat_: yes, they're legacy
<drbrain> you should use fastly
<drbrain> dwradcliffe: ↑
<mat_> drbrain, ok, I'll work on getting that changed
<drbrain> inev: try #ruby, this is for questions about rubygems itself
ur5us has joined #rubygems
<dwradcliffe> mat_: ping me if you have more questions. The cf domain is very old and Fastly is the only option now.
mjc_ has quit [Quit: Connection closed for inactivity]
huoxito has quit [Ping timeout: 240 seconds]
huoxito has joined #rubygems
huoxito has quit [Read error: Connection reset by peer]
huoxito has joined #rubygems
ernetas has joined #rubygems
<ernetas> Errno::ETIMEDOUT: Connection timed out - connect(2) for "api.rubygems.org" port 443 (https://api.rubygems.org/api/v1/dependencies?gems=cri)
<ernetas> Is this something that anyone is aware of or is it completely just me?
<ernetas> I tried multiple machines from multiple IPs. Also, clean setups. This happens when installing a bunch of gems - it usually works fine with installing single gems.
<ernetas> The IP it tries to connect to: 54.186.104.15
huoxito has quit [Remote host closed the connection]
swills has quit [Ping timeout: 250 seconds]
<ernetas> Nevermind, works now :)
<ernetas> But didn't change anything on my side.
ernetas has left #rubygems [#rubygems]
swills has joined #rubygems
jottr has joined #rubygems
jottr has quit [Ping timeout: 276 seconds]
huoxito has joined #rubygems
imperator has quit [Quit: Leaving]
jottr has joined #rubygems
ur5us has quit [Remote host closed the connection]