xalei has quit [Remote host closed the connection]
<apeiros_>
because it's not complex code and it's obvious that lines have been executed which are marked as not executed
<whitequark>
oh wow that is a lot of directories for a day old gem
<whitequark>
lemme check
<apeiros_>
I haven't spent more than a few minutes on it, though
<apeiros_>
maybe I missed something obvious
<apeiros_>
but I thought maybe there's some well known issue(s)
<whitequark>
no, never encountered any
ruby-lang754 has joined #ruby-lang
<apeiros_>
you can run the tests with simplecov using `COVERAGE=1 ruby test/runner.rb`
<apeiros_>
(you can set COVERAGE to any value really, it just tests for presence)
ruby-lang754 has left #ruby-lang [#ruby-lang]
<whitequark>
hm, the way you organize code is horrible
<whitequark>
but maybe it's just too WIP
<apeiros_>
hu?
srbaker has quit [Quit: Computer has gone to sleep.]
<apeiros_>
what's horrible about that?
<apeiros_>
referring to lib or test? or all?
<whitequark>
+x runner.rb without shebang, weird nonexisting stuff instead of a proper rakefile
<apeiros_>
ah, runner should not have +x
<whitequark>
and just this feeling that the repo is a lot of random files lying together
<apeiros_>
osx doesn't make that very visible
<whitequark>
btw I know what's wrong with simplecov
<apeiros_>
"weird nonexisting stuff" - how can stuff be weird, yet not exist?
itsmeduncan has quit [Quit: itsmeduncan]
itsmeduncan has joined #ruby-lang
<whitequark>
apeiros_: because there are six rake-related files or dirs, yet rake -T is totally empty
<whitequark>
apeiros_: re simplecov: you didn't read readme
<whitequark>
shall I explain the problem to you?
<apeiros_>
ah, that's infrastructure. I can drop in my rake stuff easily like that. I haven't because I don't need it. But maybe I should, and if it's just for other people using the project.
<apeiros_>
(it's a cross-project infrastructure, all other projects use the same)
<apeiros_>
yes, please @ explain
<whitequark>
apeiros_: readme instructs you to require simplecov before everything else
<whitequark>
and specifically your code
<whitequark>
I've fixed that and got 94% coverage
<apeiros_>
o0
<apeiros_>
the library code is being loaded after simplecov
<apeiros_>
unless…
<apeiros_>
no, the helper doesn't require library code either
<whitequark>
um
<whitequark>
I simply moved simplecov block to the top
<apeiros_>
yes, I see the difference too
itsmeduncan has quit [Quit: itsmeduncan]
<apeiros_>
I find it surprising… I interpreted "at the top" as "before you load the code which you test the coverage of"
<whitequark>
well that is exactly how you should interpret it
<apeiros_>
but that's not how it works, obviously
<whitequark>
lemme verify
<apeiros_>
unless I miss a require…
<apeiros_>
I used `p $".grep(/sorting/)`
<apeiros_>
and nothing from lib is loaded before simplecov is required
<apeiros_>
(from test/lib yes, but that's not code I want coverage-tested)
<whitequark>
hm, true
<whitequark>
no clue then. maybe that's how ruby's coverage works.
<apeiros_>
seems like
<apeiros_>
at least it's an easy fix
<apeiros_>
moved it and made a according comment
<apeiros_>
ah, hah, Smaller/Bigger#to_str - if only I remembered where that was necessary…
<apeiros_>
ugh, why does simplecov use blacklisting instead of whitelisting…
<whitequark>
argh
<whitequark>
I want to find the original developer of this stuff and beat his little moronic ass until he would explain me exactly which incoherent train of thought lead him to excise a perfectly good part of a haml template and replace it with a helper
<whitequark>
which is only ever called from a single place and which calls several partials and then concatenates the result with three strings
<whitequark>
I'm already there; what would I have to be concerned about?
havenwood has joined #ruby-lang
soypirate has joined #ruby-lang
Qfwfq has quit [Ping timeout: 255 seconds]
dzhulk has quit [Quit: Leaving.]
<apeiros_>
hm, maybe only that the padawan of a sith either dies before becoming the sith lord, or gets killed by his own padawan (if I remember that stuff right)
Qfwfq has joined #ruby-lang
benanne has joined #ruby-lang
<apeiros_>
whitequark: lol @ your tweet :D
<whitequark>
apeiros_: which one of them?
<apeiros_>
the force!
<whitequark>
ah, right
justinseiter has quit [Remote host closed the connection]
<whitequark>
it was surprisingly timely
davidbalbert is now known as davidbalber|away
ProfessorNim has joined #ruby-lang
<ProfessorNim>
hello all
<apeiros_>
indeed :)
<apeiros_>
funny, rewriting my rake tasks (haven't been using any for ages) - all gem tasks are now 2 liners: require 'rubygems/gem_runner' and `Gem::GemRunner.new.run …args…`
<whitequark>
apeiros_: I just use bundler gem tasks
<apeiros_>
what makes me a bit uneasy is that the whole env var dance happens in the gem executable
<whitequark>
though I could understand it if you don't
<apeiros_>
whitequark: I don't even want the rake tasks
mistym has joined #ruby-lang
<apeiros_>
I mean… rm *.gem && gem build *.gemspec && gem install *.gem # not that hard
<whitequark>
but why do you want to type it each time?
<apeiros_>
why would I?
<apeiros_>
there's bash_profile and alias
<apeiros_>
also ctrl-r
<whitequark>
and now you've only made it convenient for *yourself*, you egoistical prick :s
<apeiros_>
but I haven't forced my way of doing things on others, you control freak! :-p
<whitequark>
you surely did
<whitequark>
because I now have to do things your way.
<apeiros_>
o0
<dRbiG>
hello
<whitequark>
apeiros_: as every other gem has some way of doing this, with overwhelming majority just using bundler, I now have to either type them myself or recreate your aliae
<dRbiG>
i'm writing an app that uses beaneater, and the relevant question: is there some conventional way of writing rspec tests for stuff that involves deamons and networked stuff?
<dRbiG>
some sort of elaborate mocking?
<whitequark>
dRbiG: well kind of, you could always mock IO
<whitequark>
also don't test daemonization, just use one of the well-tested gems around
<whitequark>
you really don't want to write it yourself
<apeiros_>
whitequark: or you add whatever way you please and use that
<whitequark>
apeiros_: which you probably won't accept as a PR
<dRbiG>
whitequark: currently i'm trying to somehow write the tests on a 'leaving deamon'. i check if beanstalkd is running, then first descrive if the server class even initialised properly, now i'm thinking to write some sort of check for running while forking :)
<whitequark>
dRbiG: really, just don't
<apeiros_>
whitequark: you can maintain a fork :)
<whitequark>
apeiros_: I have zero wish to
<apeiros_>
whitequark: anyway, fear not, rake tasks are coming
<whitequark>
that is the problem
<whitequark>
dRbiG: gem 'daemons'
<apeiros_>
whitequark: the overwhelming majority only does `gem install foo` anyway
<dRbiG>
whitequark: yeah, i do have concerns about that approach too
<apeiros_>
and those are entirely unconcerned about those things
<whitequark>
apeiros_: who are more important to your project, long-term speaking: users or contributors? ;)
<whitequark>
yay 5am holy wars!
<apeiros_>
whitequark: I think I can live without devs who can't use neither the gem command, nor alias, nor implant whatever means they please
eklarskov has joined #ruby-lang
<apeiros_>
(nor ask what options they have)
<whitequark>
apeiros_: however if I would have to maintain a fork to work conveniently with your project, I probably just won't.
srbaker has joined #ruby-lang
<apeiros_>
whitequark: you'll have that issue no matter what. if the existing way doesn't please you, you either have to use it, maintain a fork, or walk away.
Stilo has joined #ruby-lang
<apeiros_>
you're just put off because my way isn't your way and argue in a circle…
<apeiros_>
i.e., so I do it with bundler - then comes foobar along who thinks bundle is shit and he doesn't want any of it, and because he doesn't want to maintain a fork to work conveniently with my project, he just doesn't.
marr has quit [Ping timeout: 248 seconds]
<apeiros_>
so - what's the point? :)
<whitequark>
apeiros_: I'm not really sure why would I want to push gems of your projects anywhere in the first place
<whitequark>
but as I don't really have anything else to do, let's go on
justinram has quit [Remote host closed the connection]
<whitequark>
*anything more interesting.
zmack has joined #ruby-lang
nertzy2 has left #ruby-lang [#ruby-lang]
eklarskov has quit [Ping timeout: 252 seconds]
earthquake has joined #ruby-lang
thufir_ has quit [Ping timeout: 246 seconds]
srbaker has quit [Quit: Computer has gone to sleep.]
dorei has joined #ruby-lang
slyphon has joined #ruby-lang
zmack has quit [Ping timeout: 260 seconds]
richardjortega_ has joined #ruby-lang
richardjortega has quit [Read error: Connection reset by peer]
richardjortega_ is now known as richardjortega
earthquake has quit [Quit: earthquake]
earthquake has joined #ruby-lang
netShadow has joined #ruby-lang
Qfwfq has quit [Ping timeout: 255 seconds]
Qfwfq has joined #ruby-lang
breakingthings has quit []
richardjortega has quit [Read error: Connection reset by peer]
dyln has joined #ruby-lang
richardjortega has joined #ruby-lang
sailias has joined #ruby-lang
slyphon has quit [Quit: slyphon]
havenwood has quit [Remote host closed the connection]
nXqd has quit [Read error: Connection reset by peer]
slyphon has joined #ruby-lang
drbrain has quit [Quit: Goodbye]
ProfessorNim has quit [Quit: Leaving]
drbrain has joined #ruby-lang
benanne has quit [Quit: kbai]
havenwood has joined #ruby-lang
thone has joined #ruby-lang
chendo__ has joined #ruby-lang
thone_ has quit [Ping timeout: 252 seconds]
KU0N has quit [Quit: Leaving...]
dyln has quit [Ping timeout: 264 seconds]
maxmanders has quit [Ping timeout: 256 seconds]
dyln has joined #ruby-lang
mistym has quit [Remote host closed the connection]
maxmanders has joined #ruby-lang
cirwin has quit [Quit: Leaving.]
klebervi_ has joined #ruby-lang
amerine has joined #ruby-lang
amerine has quit [Client Quit]
damienbrz has joined #ruby-lang
mistym has joined #ruby-lang
mistym has joined #ruby-lang
mistym has quit [Changing host]
thufir_ has joined #ruby-lang
Qfwfq has quit [Ping timeout: 255 seconds]
kurko__ has quit [Ping timeout: 252 seconds]
Qfwfq has joined #ruby-lang
kurko__ has joined #ruby-lang
wmoxam has joined #ruby-lang
mistym has quit [Remote host closed the connection]
wmoxam_ has joined #ruby-lang
wmoxam has quit [Quit: leaving]
wmoxam_ is now known as wmoxam
wmoxam has quit [Changing host]
wmoxam has joined #ruby-lang
postmodern has joined #ruby-lang
damienbrz has left #ruby-lang [#ruby-lang]
mercwithamouth has quit [Ping timeout: 260 seconds]
<firefux>
is it ok to buld Ruby 2.0 with 'make -j 4' ?
<lianj>
why not?
<firefux>
lianj: the rbenv build script juse -j 1 because of something openssl thing
<firefux>
s/juse/uses/
<lianj>
hm
pkondzior__ has joined #ruby-lang
<firefux>
lianj: oh it looks like is only on macs
* firefux
doesn't run macs
wallerdev has joined #ruby-lang
workmad3 has quit [Ping timeout: 245 seconds]
Qwak has joined #ruby-lang
<injekt>
morn
thone has quit [Ping timeout: 252 seconds]
Qwak has quit [Ping timeout: 256 seconds]
ryez has quit [Ping timeout: 245 seconds]
workmad3 has joined #ruby-lang
toretore has joined #ruby-lang
jxie has joined #ruby-lang
ruurd has quit [Quit: Leaving...]
<burgestrand>
apeiros_: Monitor is re-entrant and behaves like Mutex
Qfwfq has quit [Ping timeout: 255 seconds]
dorei has joined #ruby-lang
thone has joined #ruby-lang
ronaldo__ has joined #ruby-lang
ronaldo__ has quit [Client Quit]
ronaldo__ has joined #ruby-lang
savagecroc has joined #ruby-lang
<savagecroc>
hmmm
<savagecroc>
i've got to ssh onto a remote server, wait an hour to a database dump and then once it completes scp the file down.. how does this sound for a solution
<savagecroc>
get ruby to create a file on the remote server when it's done.. get ruby to pole every 10 secs to check if that file exists, once it does then scp it down
<savagecroc>
i have to nohup the database export just in case the ssh connection drops
sepp2k has joined #ruby-lang
<badeball>
savagecroc: instead of poling for the file, you could have it sent back. pg_dump -args && scp/rsync dump.sql yourcomputer.com:
<savagecroc>
badeball: ah yeah.. except the server won't have access to my computer
<savagecroc>
bit more annoying than that
<badeball>
is this a one time thing or are you setting up some kind of a backup solution?
<savagecroc>
i have to run the dump command as the postgres user otherwise it will ask for a password (and i don't want to take off the requirement in normal user accounts)
<savagecroc>
badeball: it's so developers can select dumps from various systems
<savagecroc>
and have them turn up on their computers
<savagecroc>
i.e. dev says i want xyz dump from this system, i get an email, i click yes or no.. script runs, they get the dump
ottbot has quit [Ping timeout: 252 seconds]
kuja has quit [Ping timeout: 252 seconds]
sush24 has quit [Quit: This computer has gone to sleep]
<badeball>
sounds like a time line where anything could happend and you'd need a script that can restart and resume from where it left. in that case, checking with small intervals might suit you.
kuja has joined #ruby-lang
<savagecroc>
yeah.. the file permission i don't know how to solve nicely though
<savagecroc>
best i can come up with is do the dump to /tmp/random_file_name
<bougyman>
savagecroc: you should create a different superuser account other than 'postgres'
<savagecroc>
then cp it to where it's supposed to go
<bougyman>
and you can alleviate having to type a password by creating a ~/.pgpass file
<badeball>
yeah, that'd be my suggestion as well. create a designated user with read permissions on all databases in question, capable of dumping them to wherever your users have permission to read them.
<savagecroc>
yes that's a nice idea
<savagecroc>
i like it :D
<savagecroc>
permissions never seem to be correct
<savagecroc>
unless you cp the file into the directory
<savagecroc>
scp and mv
<badeball>
you can change permissions on your own files with chmod.
<savagecroc>
end up with whatever they had
<savagecroc>
but i've got defaults on all the dirs
<savagecroc>
so for example on /srv/storage/dumps i have group:deploy:rwx,default:deploy:rwx
<savagecroc>
default:group even
<savagecroc>
i wish unix had someway of going.. ANYTHING that goes in this directory gets these permissions forevery
<savagecroc>
forever*
<bougyman>
try 04770
<bougyman>
if you want the group ownership/perms to be enforced
<savagecroc>
i try to avoid standard ACLs
<savagecroc>
and just have them all set to 770 oneuser:oneuser and just use setfacl for everything
jxie has quit [Quit: leaving]
<badeball>
I think you're going to have a hard time trying to avoid unix modes
sush24 has joined #ruby-lang
krohrbaugh has quit [Read error: Operation timed out]
<savagecroc>
what's the additional 0
kuja has quit [Ping timeout: 252 seconds]
kevinfagan has quit [Quit: Leaving...]
jxie has joined #ruby-lang
kuja has joined #ruby-lang
twintail has left #ruby-lang [#ruby-lang]
krohrbaugh has joined #ruby-lang
schaerli has joined #ruby-lang
<badeball>
the fourth bit, if present (defaults to 0 otherwise), specifies setuid, setgid and sticky bit.
<hdm>
Anyone else running into 'WARNING: No database support: TypeError superclass mismatch for class Cipher' trying to load openssl in ruby 2.0.0 via rvm (after a reinstall --force / pkg install openssl)
nachtwandler has joined #ruby-lang
kevinfagan has quit [Quit: Leaving...]
kevinfagan has joined #ruby-lang
poga has quit [Remote host closed the connection]
vlad_starkov has quit [Remote host closed the connection]
burgestrand has quit [Quit: Leaving.]
nachtwandler has quit [Remote host closed the connection]
mistym has quit [Remote host closed the connection]
io_syl_ has joined #ruby-lang
schaerli has joined #ruby-lang
schaerli has quit [Remote host closed the connection]
maxmanders has quit [Ping timeout: 256 seconds]
io_syl has quit [Ping timeout: 255 seconds]
io_syl_ is now known as io_syl
geopet has joined #ruby-lang
maxmanders has joined #ruby-lang
lsegal has joined #ruby-lang
havenwood has quit [Remote host closed the connection]
workmad3 has joined #ruby-lang
<joevandyk>
I'm finding that Ruby 2.0 takes (literally) 50x longer to render a rails page. same gems and code as ruby 1.9.3.
retro|cz has quit [Ping timeout: 252 seconds]
<hdm>
joevandyk: might be the change to utf8 encoding by default
banister`tyrion has left #ruby-lang [#ruby-lang]
banisterfiend has joined #ruby-lang
<banisterfiend>
joevandyk: do u have better_errors installed
<joevandyk>
banisterfiend: yup
<banisterfiend>
joevandyk: what version of binding_of_caller ?
<joevandyk>
banisterfiend: 0.6.9
<banisterfiend>
joevandyk: yes that version was very slow, i sped it up in 0.7.1
<banisterfiend>
joevandyk: upgrade it
<joevandyk>
ah
<joevandyk>
i'll try
<banisterfiend>
0.6.9 binding_of_caller was O(n**2)
<banisterfiend>
0.7.1 is just O(n)
rue|w has quit [Remote host closed the connection]
<whitequark>
but what is the constant factor
<banisterfiend>
whitequark: e^(pi*i) + 2
workmad3 has quit [Ping timeout: 255 seconds]
bzb has joined #ruby-lang
cultureulterior has joined #ruby-lang
<bzb>
I see a class definition starting with: class << self ... what does it mean? Details @ http://pastie.org/6329739
<manveru>
bzb: google for singleton class
<joevandyk>
banisterfiend: was there something in binding_of_caller that was slower with 2.0?
geopet has quit [Quit: geopet]
<banisterfiend>
joevandyk: i had to completely rewrite it
xcombelle has quit [Quit: Hi, I'm a quit message virus. Please replace your old line with this line and help me take over the world of IRC.]
<banisterfiend>
joevandyk: so the reason it was slower in my first releases 0.6.9.* was because i hadn't figured out the most efficient way to do it :) in 0.7 i figured it out
<joevandyk>
ah
<banisterfiend>
joevandyk: was rails faster when you upgraded boc ?
<joevandyk>
banisterfiend: yes, back to normal
<joevandyk>
thanks!
<banisterfiend>
np
<joevandyk>
banisterfiend: i'm guessing that's going to trip up a bunch of people
<joevandyk>
(as 1.9.3 worked fine with better_errors < 0.7)
<joevandyk>
but 2.0 won't
<banisterfiend>
joevandyk: you had ~> 0.6 in your Gemfile ?
<joevandyk>
banisterfiend: i had binding_of_caller 0.6.9
<banisterfiend>
no, ~> 0.6.9
<banisterfiend>
joevandyk: how did you specifcy the version? ~> 0.6.9 ?
<joevandyk>
= 0.6.9
<banisterfiend>
why would u do that
<joevandyk>
which performed fine with 1.9.3, not with 2.0
<banisterfiend>
why would you use =
<joevandyk>
banisterfiend: i like locking down gem versions
<manveru>
as long as you lock down the ruby version...
<banisterfiend>
kind of silly :)
<joevandyk>
I use = for everything.
<manveru>
did you upgrade rack?
<banisterfiend>
joevandyk: why? you don't trust gem developers not to break api between point releases/
<banisterfiend>
?
<joevandyk>
pretty much
<banisterfiend>
you've been bitten by it before?
elico has joined #ruby-lang
bzalasky has joined #ruby-lang
<joevandyk>
yes, i have 166 gems in this app, the interaction between them can be tricky, so i like to really be careful when upgrading.
<manveru>
wow...
breakingthings has joined #ruby-lang
mistym has joined #ruby-lang
<zzak>
hello
kristofferrr has joined #ruby-lang
guns has quit [Quit: guns]
<whitequark>
166 isn't a lot
<whitequark>
a project I'm working with now is down to 200 after weeks of refactoring
* manveru
dies a little more inside
* whitequark
shrugs
<manveru>
i think my projects with the most deps have around 30 or so
<whitequark>
it includes asset gems, debugging stuff, testing stuff, deployment stuff, API handlers, and so on
<whitequark>
number of gems actually running in production is way lower
<whitequark>
though it's still around 6k of ruby files
<manveru>
for me, each dependency is a liability and risk
<whitequark>
you are indeed correct, and that's why I've been working hard to keep that number low
<whitequark>
note though that there isn't a single gem I can really remove from production. for each one a business task exists which justifies its presence.
bcardarella has joined #ruby-lang
<whitequark>
(except for THREE form builders in one rails app. whoever added them all should be fired.)
Mizpah has joined #ruby-lang
mistym has quit [Remote host closed the connection]
<bcardarella>
Just upgraded to 2.0 and when I try to install a gem I'm getting "Permission denied - /Users/me/.gem/specs"
<bcardarella>
I chmod +rw on the directory and I own it but I'm still getting that
<manveru>
+x ?
kevinfagan has quit [Quit: Leaving...]
<bcardarella>
lemme try +x
<manveru>
whitequark: well, i try to split as much as possible to make each part fast to test
<manveru>
the app i work on atm has like 200k LoC including dependencies... even pulls in a whole fork of V8 to minify js :P
<bcardarella>
manveru: that seems to have worked
<bcardarella>
ty
<whitequark>
manveru: yeah, testing :/
* whitequark
hates rails tests against DB with a passion
<manveru>
indeed
<whitequark>
especially that on this particular project preloaders tend to fail
<whitequark>
well, zeus just doesn't generally work on linux (fucking go)
<whitequark>
and spring for some reason screws up all rails internals when launching rspec
<manveru>
the zeus load balancer?
<whitequark>
manveru: the zeus checkpointing application preloader
<whitequark>
github:burke/zeus
<manveru>
hm
<manveru>
and what's wrong with go and linux?
<manveru>
all issues i can find are closed
erichmenge has quit [Quit: Arrivederci!]
<whitequark>
manveru: it randomly explodes with a null pointer
<whitequark>
they might've fixed it already, I'm not using it for months
<manveru>
ok
<manveru>
dunno... "our app starts too slow due to too many dependencies; i know, let's solve it with yet another dependency" :P
<whitequark>
manveru: it's not like you can reduce the amount of dependencies anyway.
<manveru>
you can make smaller applications
<whitequark>
no
<whitequark>
well, in your particular case, maybe yes. generally this isn't true.
<manveru>
not sure how you think packing everything into a single huge process is a good solution
<manveru>
i mean, i get that that's what rails advocates mostly... but nobody seems to question that much
bcardarella has left #ruby-lang [#ruby-lang]
<whitequark>
haha, this actually reminds me of another debate
<manveru>
then again, they are also fine with the routing, so i guess nobody gives a shit about anything
<whitequark>
manveru: btw, what's wrong with routing?
willyw has quit [Remote host closed the connection]
workmad3 has joined #ruby-lang
<manveru>
it depends on http methods, is tied to the model, is implemented with regular expressions, is rather impossible to manage without an ide or `rake routes`...
erichmenge has joined #ruby-lang
<manveru>
i mean, it may be better than php routing, but what isn't :P
<whitequark>
manveru: do you have better options?
<whitequark>
(in particular, "implemented with regexps" is a rather bullshit argument)
<manveru>
yeah, because we never have any regex exploits
Qwak has joined #ruby-lang
<whitequark>
again: do you have better options?
<manveru>
sure, but that would be advertising the stuff i built, so you wouldn't care anyway
<whitequark>
this reminds me of DCI proponents. looks good in theory, but isntantly falls apart in practice
<whitequark>
manveru: please do advertise it
<manveru>
ramaze/innate
<whitequark>
oh, ramaze. did you write it?
<manveru>
yes
<whitequark>
I actually think it's pretty cool. Haven't had time to explore it in details yet
<hdm>
anyone else running into this exception with openssl on ruby 2.0? => cipher.rb:61:in `<class:Cipher>': superclass mismatch for class Cipher (TypeError)
<hdm>
irb / require doesn't trigger, but require() from within the app does
<manveru>
hdm: full backtrace?
<zzak>
manveru: a small rails project that i wasn't going to get to
<zzak>
you said "i have to work with rails atm for the first time"
<hdm>
ruby -r openssl ./app seems to hit the same issue too, /usr/local/rvm/rubies/ruby-2.0.0-p0/lib/ruby/2.0.0/openssl.rb
<hdm>
clean rvm / forced reinstall / rebuilt pkg openssl after hitting this during the first test
<manveru>
yeah, it works fine here with rvm and 2.0.0-0
kristofferrr has quit [Quit: ❤]
<manveru>
ruby -r openssl -e ''
<hdm>
i figured if something was monkeying with the openssl namespace, running ruby -r openssl to load it would have sovled it
<manveru>
fails too?
<hdm>
that works fine, same with irb
<hdm>
smells like some gem dependency did a 'module OpenSSL' monkeypatch, but forcing openssl to be required before the app starts still hits it
bzb has quit [Quit: I hate to quit but my bladder's full. :-(]
<manveru>
hm
<manveru>
maybe some patch for http stuff?
bzb has joined #ruby-lang
cultureulterior has quit [Quit: cultureulterior]
<hdm>
digging to see what could have caused it, moved gemfile out of the way, etc
<hdm>
must be some dep screwing it up
<manveru>
net-ssh maybe?
<manveru>
hm, no
benanne has quit [Quit: kbai]
bzb has quit [Read error: Connection reset by peer]
<hdm>
yeah, it doesn't seem to hack up any of the openssl classes, just calling them fully qualified
bzb_ has joined #ruby-lang
<manveru>
well
<hdm>
odd that the require is causing a warning on the constants
<manveru>
you could try to find the implementation by checking $LOADED_FEATURES
<manveru>
in particular in /usr/local/rvm/rubies/ruby-2.0.0-p0/lib/ruby/2.0.0/openssl/cipher.rb to cut it down a bit
kurko__ has joined #ruby-lang
reppard has joined #ruby-lang
<hdm>
just seems odd - why would ruby -r openssl ./code_which_requires_openssl trigger a warning on constant redefinition
<hdm>
unless $LOADED_FEATURES was being scrubbed of the original require (doesnt seem to be the case), prior to the second require
<manveru>
Kernel#load would do that
<manveru>
not that anybody uses that much... but who knows
<hdm>
true, but the callstack there is a require()
<manveru>
the second, yeah
<manveru>
you don't see the first require
<hdm>
the first is on the command line
<hdm>
i wonder if ActiveSupport is not compatible or horking it up somewhere
<manveru>
i don't see anything defining Cipher here...
<hdm>
just thinking its require() overloads could be breaking somewhere, ill print a callstack on the openssl.rb require and figure out how its getting loaded a weird way
<manveru>
yeah
<manveru>
might be autoload or threads
<manveru>
can't really say without a way to reproduce
ascoppa has joined #ruby-lang
schaerli has joined #ruby-lang
Wildhoney has quit [Ping timeout: 255 seconds]
schaerli has quit [Remote host closed the connection]
jbsan has quit [Quit: jbsan]
tylersmith has joined #ruby-lang
Stilo has joined #ruby-lang
reppard has quit [Ping timeout: 252 seconds]
<hdm>
manveru: that is bizzare, $LOADED_FEATURES already contains "openssl.rb" when its loaded the second time
<hdm>
yet the callstack is still 'require'
<hdm>
normally require would skip over in that case
jbsan has joined #ruby-lang
<manveru>
openssl.rb maybe, but openssl.so ?
<hdm>
well, openssl.rb is definitely being loaded twice when it shouldnt
<hdm>
and it is what loads openssl.so
<manveru>
aye
<manveru>
did you find a way to reproduce it simple yet?
<hdm>
not yet, its something way after the gem deps are loaded
<hdm>
#5 there is the caller for the second incorrect load, im guessing its something bured in rubygems or active_support or something
<hdm>
using rubygems 2.0.0.rc.2, looks like its definitely to blame
<hdm>
will try to find a stable repro
reppard has joined #ruby-lang
havenwood has joined #ruby-lang
postmodern has joined #ruby-lang
bzb_ has quit [Quit: I hate to quit but my bladder's full. :-(]
mjio has joined #ruby-lang
reppard has quit [Ping timeout: 256 seconds]
jwaldrip has joined #ruby-lang
<hdm>
manveru: even weirder, traced every call to the original require and the loaded_features leading into it - they are identical, but the ~9th require "openssl" causes the original require to reload it from disk again
<hdm>
points back to ruby 2.0.0's require() as the culprit
maxmanders has quit [Quit: Computer has gone to sleep.]
<postmodern>
whitequark, seems to also happen on 1.9.3 with rspec 2.13.0
solars has quit [Ping timeout: 260 seconds]
gix has quit [Ping timeout: 240 seconds]
cultureulterior has joined #ruby-lang
<whitequark>
postmodern: btw, did you encounter random build failures with travis?
<whitequark>
something is really wrong with rubygems. timeouts, SSL errors, ...
tylersmith has quit [Quit: tylersmith]
cultureulterior has quit [Client Quit]
willywos has joined #ruby-lang
gix has joined #ruby-lang
havenwood has joined #ruby-lang
geopet has joined #ruby-lang
willywos has quit [Remote host closed the connection]
nXqd has quit [Ping timeout: 252 seconds]
<whitequark>
postmodern: $ gem install rspec
<whitequark>
ERROR: While executing gem ... (Gem::DependencyError) Unable to resolve dependencies: rspec requires rspec-core (~> 2.13.0)
<whitequark>
I'm not exactly sure why does this happen or how do I fix it
<Mizpah>
Hi Folks - Is there any type of common newbie error that causes the code in a partial (with a table) to render once per row ? i.e. <table> header (row1) </table> <more code>, <table> header (row2) </table> <more code> etc ? I calling a partial via render @game_classes -- the partial used to work, but am unsure when I introduced the error - so its hard to step back! If nothing springs to mind will make a gist :)