_djbkd has quit [Remote host closed the connection]
havenwood has quit [Ping timeout: 272 seconds]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kangguru is now known as kangguru_away
iamjarvo has joined #bundler
iamjarvo has quit [Ping timeout: 256 seconds]
jahkobi has quit [Quit: jahkobi]
Cidan is now known as zz_Cidan
robbyoconnor has quit [Ping timeout: 255 seconds]
_djbkd has joined #bundler
_djbkd has quit [Remote host closed the connection]
_djbkd has joined #bundler
x1337807x has joined #bundler
kfpratt has quit [Remote host closed the connection]
kfpratt has joined #bundler
kfpratt has quit [Remote host closed the connection]
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
_djbkd has quit [Remote host closed the connection]
livingstn has quit [Ping timeout: 264 seconds]
x1337807x has joined #bundler
houhoulis has joined #bundler
iamjarvo has joined #bundler
kgrz has joined #bundler
kgrz has quit [Ping timeout: 264 seconds]
kgrz has joined #bundler
kgrz has quit [Client Quit]
kgrz has joined #bundler
huoxito has quit [Remote host closed the connection]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
x1337807x has joined #bundler
havenn has quit [Remote host closed the connection]
jahkobi has joined #bundler
jahkobi has quit [Client Quit]
havenwood has joined #bundler
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
robbyoconnor has joined #bundler
huoxito has joined #bundler
huoxito has quit [Ping timeout: 265 seconds]
zz_Cidan is now known as Cidan
Cidan is now known as zz_Cidan
prime has quit [Ping timeout: 244 seconds]
houhoulis has quit []
houhoulis has joined #bundler
prime has joined #bundler
relix has joined #bundler
kangguru_away is now known as kangguru
<jacobat>
indirect: I know it doesn't. Bundler currently does not seem to solve my problem. That's why I'm here :)
<indirect>
jacobat: it sounds like you have a problem that’s completely unrelated to the one that bundler solves :)
<indirect>
(but maybe I’m wrong. feel free to describe your problem!)
<jacobat>
indirect: We're using Docker for development
<indirect>
that seems reasonable
<jacobat>
indirect: So we build an image with our application which stores gems in the default gem path /usr/local/whatnot
<jacobat>
Bundler stores git gems in there as well
<indirect>
that means you’re running `bundle install` without a --path option
<indirect>
probably you don’t want to do that
<jacobat>
No I do :)
<jacobat>
Hang on :)
<indirect>
hah, sure
<jacobat>
When we want to do development for that particular application we mount the hosts filesystem into the container overriding the application directory
<jacobat>
Typically we do that in /app
<jacobat>
Then if the user modifies their Gemfile we want to rebundle, but we can't rebundle to /usr/local because that path is not persisted between invocations of the container
<jacobat>
So instead we put gems installed afterwards in /bundle - a special directory that is shared between all containers
<indirect>
okay
<jacobat>
Now the issue is that we set GEM_HOME to point to /bundle so that any gems that have been changed from when the image was built to when the user runs bundle install will be installed in /bundle
<jacobat>
But this also means that bundler will no longer look to /usr/local for already installed git gems and will reinstall all git gems in /bundle
<indirect>
and that means that the gems installed in /usr/local are not available to your bundle
<indirect>
yes
<indirect>
I mean, does it look there for non-git gems?
<indirect>
if it does, that is news to me
<jacobat>
It does :)
<indirect>
I would expect it to only look in /bundle for all gems
<indirect>
huh
<indirect>
how about that
<jacobat>
Hehe
<indirect>
maybe Rubygems silently provides all the gems in GEM_PATH
<jacobat>
That would seem reasonable (without knowing any internals)
<indirect>
Bundler was (as I suspect you know) written specifically to put all dependencies in a single place, repeatably
<jacobat>
Well, it reuses regular gems that are in the gem path :)
<indirect>
so GEM_PATH is more or less incompatible with the underlying reason for Bundler’s existence
<indirect>
that’s a bug :)
<jacobat>
Oh crap! Please don't fix it!
<indirect>
I mean, it’s a bug that appears to be super convenient for you
<indirect>
but it’s definitely not by design
<indirect>
haha
<jacobat>
We have some rather heavy Gemfiles
<indirect>
it sounds like fixing it would take even more monkeypatching of Rubygems than Bundler already does
<indirect>
so it’s very unlikely to get fixed
<indirect>
if that makes you feel better
<jacobat>
:)
<indirect>
but in general, the point of bundler is to install all the gems in one place, so I’m very surprised to hear that you have had as much success as you have had with this plan that sounds quite hare-brained to me :)
<jacobat>
Yeah, I would prefer a simpler scheme as well ;)
<indirect>
are you unable to seed /bundle with the base image gems?
kgrz has quit [Ping timeout: 265 seconds]
<jacobat>
No, I could probably do that just by copying /usr/local/... there
<indirect>
so I personally would favor running bundle install --path .bundle inside each app directory before freezing the image
<indirect>
since that leaves the bundle files inside the app checkout as build/dev artifacts
<jacobat>
Remember that /app gets replaced with what's in the host file system ;)
<indirect>
oh
<indirect>
I see
<indirect>
you are doing even stranger and more complicated things than I initially thought :)
<jacobat>
Sure, it's possible - but that means every user has to spend however long a full install takes when most of the gems are already in place
<indirect>
why doesn’t the host FS have the bundle?
<jacobat>
Yeah, it's a bit of a wonky setup we have - the goal is to make the development experience for users as transparent as possible
<jacobat>
The host FS could have the bundle, but that would require each user to do a full bundle install
<indirect>
the host FS is seeded with the app source, why can’t it be seeded with the bundle as well?
<indirect>
(if the bundle is already present along with the source code, bundle install is a no-op)
<jacobat>
The workflow is that the user fetch the app repository from github and then starts a container
<indirect>
so when you say “most gems are already installed”, you either have most of rubygems.org in your image, or that’s probably not true
<jacobat>
I mean most gems in the Gemfile
<indirect>
so this is limited to a set of repos in a single org, or something?
<jacobat>
We build an image for each repos
<jacobat>
So when you work on app X you get the image for app X
<indirect>
and it’s impossible for you to provide a directory from the image that’s writable?
<indirect>
if that’s true, then yeah, I think copying GEM_HOME into /bundle and then setting GEM_HOME=/bundle is your best bet
<jacobat>
With Docker you can only shared FS paths that gets mounted from the outside.
<indirect>
I don’t mean shared, I just mean writable
<indirect>
you don’t need to persist changes to the bundle from the repo base
<indirect>
just allow them to be written
<indirect>
and then thrown away
<jacobat>
Sure you can write wherever, but it doesn't get shared between containers
<jacobat>
But if I have two Docker containers, one running my webserver and one running bundle install the gems will only be visible to the latter
<indirect>
and you want any gem any dev every installs in any commit to be available to every other dev ever?
<indirect>
oh
<indirect>
you have an entire container just for running bundle install? o_O
<jacobat>
That's why we have /bundle which is shared across containers
<indirect>
(I mean, yes, yes, lightweight containers, but still :P)
<indirect>
yep
<jacobat>
Well, we have a container where we start a shell and one where we run the webserver
<indirect>
how do you share code edits from the shell to the web server?
<jacobat>
The app directory is mounted from the host FS, so that's shared :)
<indirect>
I see
<indirect>
given those constraints, yep, sounds like it’s copying-to-seed-gems time :P
<jacobat>
:)
<jacobat>
That probably is the better solution yeah
<jacobat>
Thank you for taking your time to listen to me :)
<indirect>
np :)
debnathshoham has joined #bundler
debnathshoham has quit [Client Quit]
debnathshoham has joined #bundler
kangguru is now known as kangguru_away
kgrz has joined #bundler
houhoulis has quit [Remote host closed the connection]
kangguru_away is now known as kangguru
kangguru is now known as kangguru_away
havenwood has quit [Remote host closed the connection]
kangguru_away is now known as kangguru
kgrz has quit [Ping timeout: 244 seconds]
kgrz has joined #bundler
robbyoconnor has quit [Ping timeout: 245 seconds]
Muz has joined #bundler
<Muz>
Best way to specify that a Gemfile needs to lock the Ruby version to "1.9.3" OR "2.0.X"?
huoxito has joined #bundler
huoxito has quit [Ping timeout: 250 seconds]
kangguru is now known as kangguru_away
iamjarvo has joined #bundler
Pupeno has joined #bundler
Pupeno has joined #bundler
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
huoxito has joined #bundler
livingstn has joined #bundler
iamjarvo has joined #bundler
kangguru_away is now known as kangguru
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kfpratt has joined #bundler
havenwood has joined #bundler
iamjarvo has joined #bundler
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #bundler
kgrz has quit [Ping timeout: 244 seconds]
havenn has joined #bundler
havenwood has quit [Ping timeout: 264 seconds]
kgrz has joined #bundler
heroux has quit [Ping timeout: 272 seconds]
heroux has joined #bundler
arup_r has joined #bundler
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
iamjarvo has joined #bundler
iamjarvo has quit [Client Quit]
Pupeno_ has joined #bundler
mjc_ has quit [Ping timeout: 265 seconds]
igitoor has quit [Ping timeout: 265 seconds]
halorgium has quit [Ping timeout: 265 seconds]
shwoodard has quit [Ping timeout: 265 seconds]
callenb___ has quit [Ping timeout: 265 seconds]
JSharp has quit [Ping timeout: 265 seconds]
schisamo has quit [Ping timeout: 265 seconds]
Pupeno has quit [Ping timeout: 265 seconds]
whoojemaflip has quit [Ping timeout: 265 seconds]
halorgium has joined #bundler
igitoor has joined #bundler
schisamo has joined #bundler
callenb___ has joined #bundler
whoojemaflip has joined #bundler
shwoodard has joined #bundler
mjc_ has joined #bundler
kgrz has quit [Ping timeout: 265 seconds]
JSharp has joined #bundler
igitoor has quit [Changing host]
igitoor has joined #bundler
kgrz has joined #bundler
axl_ has joined #bundler
x1337807x has joined #bundler
iamjarvo has joined #bundler
kangguru is now known as kangguru_away
_djbkd has joined #bundler
smoitra has joined #bundler
<smoitra>
Hello I am willing to participate in gsoc 2015 I have visited the ideas page and I am interested in working in bundler
<varunwachaspati>
can anyone state a possible reason for this?
<varunwachaspati>
PS : I have checked with my internet connection everything works fine
<varunwachaspati>
I possibly suspect there is clash of dependencies I suppose as I had recently install some server deployment gem
zz_Cidan is now known as Cidan
x1337807_ has joined #bundler
x1337807x has quit [Ping timeout: 240 seconds]
havenn has quit [Remote host closed the connection]
smoitra has quit [Ping timeout: 252 seconds]
smoitra has joined #bundler
debnathshoham has quit [Quit: Connection closed for inactivity]
smoitra has quit [Quit: Leaving]
relix has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
kgrz has quit [Ping timeout: 265 seconds]
varunwachaspati has quit [Read error: Connection reset by peer]
relix has joined #bundler
x1337807_ has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
x1337807x has joined #bundler
kangguru is now known as kangguru_away
axl_ has quit [Quit: axl_]
axl_ has joined #bundler
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
khanh has joined #bundler
<khanh>
hey guys, i'm running into an issue with wrapping a group block within a source block. It looks like all gems get included by doing a Bundler.require(:default), which to me seems wrong, has anyone seen this happen? I'm on bundler 1.8.4
khanh has quit [Quit: Page closed]
_djbkd has quit [Remote host closed the connection]
huoxito has quit [Remote host closed the connection]
havenwood has joined #bundler
_djbkd has quit [Remote host closed the connection]
kangguru_away is now known as kangguru
_djbkd has joined #bundler
arup_r has quit [Quit: ChatZilla 0.9.91.1 [Firefox 36.0/2015022000]]
relix has joined #bundler
huoxito has joined #bundler
kgrz has joined #bundler
kgrz has quit [Ping timeout: 240 seconds]
livingstn has quit []
kangguru is now known as kangguru_away
maxyss_ has joined #bundler
maxyss_ has quit [Client Quit]
_djbkd has quit [Remote host closed the connection]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
_djbkd has joined #bundler
_djbkd has quit [Remote host closed the connection]
_djbkd has joined #bundler
relix has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
havenwood has quit [Remote host closed the connection]
triplecorn has joined #bundler
<triplecorn>
hi
<triplecorn>
glad there is a bundler channel
<triplecorn>
So can I upgrade the gem versions in a Gemfile - automatically?
<triplecorn>
bundle update tells me everything is updated although old versions are still in the Gemfile, it isn't changed.
havenwood has joined #bundler
huoxito has quit [Remote host closed the connection]
<triplecorn>
anyone?
<Radar>
triplecorn: bundle update just updates to the latest specified in the Gemfile.
<triplecorn>
Radar: any tool to update the versions in Gemfile to latest?
<Radar>
triplecorn: For instance, if you have gem 'rails', '~> 3.0' you won't get anything later than Rails 3.2.21
<Radar>
triplecorn: there's plenty: Vim, Emacs, Sublime Text 3, RubyMine...
<Radar>
Look at "bundle outdated" as well
<Radar>
triplecorn: There's no automatic process to do it because there's rarely a case where you want to upgrade ALL the gems in your Gemfile at the same time.
<triplecorn>
damn
<Radar>
Using Rails again as an example, I wouldn't want to go upgrading my app from 3.2.21 to Rails 4.2.