<Guest2452>
how can i clear gem's downloaded gem files?
Guest2452 joined #ruby-lang
Guest2452 left #ruby-lang
zenspider joined #ruby-lang
mksm joined #ruby-lang
<mksm>
How can i clear gem's downloaded gems?
<havenn>
Guest2452: If what you are asking is how to uninstall all gems, this is what I do (sane or not): `gem list`.split("\n").each { |gem| system "gem uninstall -aI #{gem.split(" ").first}" }
<zenspider>
mksm: do what?
<mksm>
havenn, not really. I want to force gem to refetch the gem and install
Nisstyre joined #ruby-lang
<erikh>
$GEM_HOME/cache/ is probably where you want to look, or each item in $GEM_PATH's /cache/ dir
<erikh>
but that's a semi-educated guess.
<zenspider>
mksm: why?
<mksm>
zenspider, i want to reinstall a gem that has some modifications and it keeps installing the old one.
<mksm>
version numbers are the same (i know)
<andrewvos>
mksm rvm clean or something?
<andrewvos>
mksm: Or just delete the gems path
<andrewvos>
Oh wait. Will that delete cahce I wonder?
<zenspider>
rm the .gem file from the corresponding cache directory (use gem list -d gemname)
onio joined #ruby-lang
<zenspider>
then just reinstall
<zenspider>
and then chastise the author for publishing a gem with changes without modifying the version
<zenspider>
esp if that author is you :P
<mksm>
andrewvos not gem cache
<mksm>
zenspider, thanks. will install always overwrite?
nouser joined #ruby-lang
<zenspider>
yes
looopy joined #ruby-lang
looopy joined #ruby-lang
statim joined #ruby-lang
replore_ joined #ruby-lang
uniqanomaly__ joined #ruby-lang
io_syl joined #ruby-lang
io_syl joined #ruby-lang
babinho joined #ruby-lang
lele|w joined #ruby-lang
slyphon joined #ruby-lang
nouser left #ruby-lang
<andrewvos>
What does everyone use for irssi notifications?
<shevy>
xchat
<shevy>
,D
<zenspider>
emacs :P
<steveklabnik>
alt-tab
<andrewvos>
*disapproving look*
<shevy>
haha
<steveklabnik>
oh, yeah, sorry, i'm on a Mac. command-tab
<steveklabnik>
:p
<andrewvos>
HAH BUSTED
<andrewvos>
Just seems primitive that I have to ^n to find peoples messages that have been around for ages.
<steveklabnik>
i mean, it highlights when you get mentioned
<steveklabnik>
at least, mine does.
<andrewvos>
hmmm.
<andrewvos>
Mine doesn't
<steveklabnik>
odd
<steveklabnik>
irssi 0.8.15 (20100403 1617)
<steveklabnik>
regular old arch linux.
<andrewvos>
No idea what version I'm on
<steveklabnik>
irssi 0v
<steveklabnik>
-v
<andrewvos>
Running on an ec2 instance for months. Too scared to touch it cause i have no backups
<andrewvos>
Ah. Same version
<andrewvos>
Anyway, who cares. Night everyone
<j`ey>
andrewvos: /hilight andrewvos
Diaoer joined #ruby-lang
<andrewvos>
j`ey: Worth a try
<j`ey>
andrewvos: hey hey hey yellow text
looopy joined #ruby-lang
s3mi0 joined #ruby-lang
<shevy>
lol
zedUNDginger joined #ruby-lang
neoesque joined #ruby-lang
edwardsharp joined #ruby-lang
tenderlove joined #ruby-lang
Oloryn_lt1 joined #ruby-lang
<zedUNDginger>
can anyone think of a way a method can get a Method object of itself without using its name
<zedUNDginger>
Mon_Ouie: ping
Asher joined #ruby-lang
<vereteran>
where is the site or source for hash table used in ruby (that st.c file in sources)? it credits to Peter Moore, but googling doesn't help :( it seems like he doesn't have a site or something similar
<vereteran>
steveklabnik, no, i want original source, this one is edited for ruby. see the file: it includes either ruby/ruby.h or if its not ruby source then st.h
<steveklabnik>
...what?
<zenspider>
headius: oi
<vereteran>
steveklabnik, it was modified for ruby, its not original part of it. i want original source
<zedUNDginger>
steveklabnik: he means the internal hash implementation used by ruby, st.c
<headius>
hi there...I'm getting to that pull request :)
<zenspider>
headius: the more aaron and I talk about this, the more we want to push back on this and have platform predicates instead... what was your argument against that?
<zenspider>
eg: putting 'skip "bug 1234" if jruby?' in tests directly
<vereteran>
steveklabnik, thank you! seems to be it
<headius>
zenspider: the set of tests we will or will not pass over time is going to change...maybe on a daily basis
<zenspider>
that st.c is going to have drifted a long way out from ours
<headius>
having guards in the tests themselves is not going to be efficient
<headius>
plus I don't have commit access to MRI, so I can't remove or change the guards as we pass them
<steveklabnik>
zenspider: since 1985? that's such a short time!
<zenspider>
headius: I can fix the commit bit issue
<zenspider>
and I assume you'd be doing this all the time but merging the tests to mri only periodically?
<zenspider>
which I think you should be doing already... as you prolly have modifications/fixes/improvements to mri tests already, no?
<headius>
as with rubyspec, I'd like our test run to pull a specific revision of tests at the start
<headius>
no, we have tried to avoid modifying them at all
<headius>
we have removed a few tests that we can't pass because we have no other way to mask them
<headius>
the ideal situation would be that we can just grab a certain rev of MRI repo and run with a set of tags against it
<headius>
then I never have to commit or modify the tests in any way
<zenspider>
wouldn't it be preferable to take those removed tests and merge in a skip line?
<headius>
it would be preferable to what we do now, but having an external list of tests to skip would better isolate us
<headius>
I don't have a strong preference, but skip lines specific to JRuby living in MRI's tests seems really ugly to me
<headius>
and then what if other impls want skips too?
<zenspider>
ok. I'm gonna record this dialog off and talk to aaron about it. I still think you're wrong... but I'm open to the change. :P
<headius>
it's a mess
<zenspider>
I have a test up and running for it already
<zenspider>
that's NOT a mess
<zenspider>
that's GREAT
<headius>
show me what you mean
<headius>
maybe I'm not understanding how the tests would look
<zenspider>
it means at once glance you can see the state of the official tests across all the major impls
<zenspider>
def test_ary_blah_blah; skip "bug 1234" if jruby?; assert_blah; end
<headius>
hmm
<headius>
well if it were my codebase I'd hate it, but it does serve the same function
<zenspider>
by internalizing the data, third party libs, auditors, etc, can all see at a glance what the state is
<headius>
I'm also a little surprised that ruby-core would agree to it
<zenspider>
lemme argue your case with aaron over dinner and see where it goes
<zenspider>
I'm not really thinking of asking them :P
<zenspider>
ok. dinner. bbl
<headius>
ok
lsegal joined #ruby-lang
hahuang65 joined #ruby-lang
babinho joined #ruby-lang
looopy joined #ruby-lang
roberto_bob joined #ruby-lang
s3mi0_ joined #ruby-lang
neoesque joined #ruby-lang
wyhaines joined #ruby-lang
<jmontross>
what is the best way to compare two files to determine if they contain same text? One is an erb template and the other is a file that the erb template will replace if they are not the same
<jmontross>
yeah, i'm planning to md5 them…. thanks zed
<Spooner>
If they are large, it can be faster to compare both sizes first (if they aren't the same size, they aren't the same).
<jmontross>
thanks. that is good idea.. they are only like 8 lines of text for a /etc/environment
<porcelina>
uhm, byte by byte comparison is faster than md5
<Spooner>
Speed probably doesn't matter then (and anyway, if they are the same most of the time, you'd still need to compare them).
<porcelina>
unless you're storing it
<porcelina>
since you have to read both files either way
<porcelina>
and byte by byte comparison has no hashing and can shortcut as soon as you hit a dissimilar byte
<jmontross>
really? i heard md5 was faster…. that makes perfcet sense tho - you'd store the md5 if you had to keep comparing same fil
<porcelina>
if it's like 3 files, md5 is faster
<lianj>
"thanks zed" sounds scray in this channel
<steveklabnik>
lol
<Spooner>
He is just paying homage to my game. He should pay me to use the name though :D
<porcelina>
what game?
<Spooner>
Zed and Ginger :D
<lianj>
surfing cats, nice!
<Spooner>
Not enough people making Ruby games!
<porcelina>
what are you writing it in? ruby?
<Spooner>
Ruby with Mon_Ouie's Ray library, but I usually use Gosu.
<porcelina>
i thought about using ruby with a custom c backend... but it just seems like a lot of work
<lianj>
did you enjoy writing it in ray vs gosu?
<Spooner>
Ray does have a lot more magic features (like glsl) but I found gosu a bit more mature.
<porcelina>
are there any 3d libraries?
<Spooner>
You can use 3D in Ray, but in Gosu you need to use the OpenGL gem.
<Spooner>
I didn't do any 3D though. All just 2D sprites with some shaders.
<lianj>
modern gl is mostly pushing mesh data to the gpu and using shaders
<porcelina>
yeah
<porcelina>
if you load and manage all your meshes in c it seems like you should be able to do all the high level stuff in ruby with no real hit on performance
<Spooner>
I think you can happily load objects into Ray.
<Spooner>
Yeah, they both work on all 3 OSes and I know Ray works on Rubinius and jRuby (not on Windows, of course :D).
<lianj>
porcelina: you can attach a physicslib and opengl via ffi. prepare buffers and have ffi pointers to them. then just pass the pointers from opengl to physics and let both c libs handle the hard work
tphummel joined #ruby-lang
<Spooner>
Yes, the main problem is that Ruby is slow and all the C++ underneath is fast.
<Spooner>
In trying to make a game that runs at a decent FPS, of course.
<porcelina>
yeah
<porcelina>
that's why i was thinking custom c backend might be easier
<porcelina>
instead of loading opengl and phsics with ffi just load them in c and load your own c in ffi
<Spooner>
Depends what you want to do. The existing libraries do fine. I have made chipmunk+gosu games and they run well, but hooking them up via Ruby does burn CPU, I guess.
<lianj>
i did that at first and went back to pure ffi. not that its the best solution in any case, but works fine for my cases
<tphummel>
hi everyone. sorry in advance for being a noob. i have a question about doing the equivalent of "next" within map or reduce, where i want to return memo for the next iteration of the map/reduce. return breaks out of the containing function which is too much.
<steveklabnik>
tphummel: break?
<steveklabnik>
maybe?
<lianj>
for example, im using gstreamer and gl. the frame buffer never hits ruby code, you pass the pointer of it directly to gl
<tphummel>
can i pass my memo var to break?
<ryanf>
I think break might still be too heavy
<ryanf>
but doesn't next work?
<steveklabnik>
yeah
<porcelina>
uhm, doesn't map reduce need a result?
<porcelina>
if you're memoizing just do memo || calculation...
<tphummel>
yeah, that's what i thought. b/c you return your memo at the end of each run
<porcelina>
yeah
<porcelina>
and it should shortcut out if it finds the memo
<porcelina>
i'd do like { |i| memo[i] || memo[i] = (calc) }
<tphummel>
ok, i think i see
swarley joined #ruby-lang
<porcelina>
be nice if there was like .map.memoize or something that did it for you tho
Diaoer joined #ruby-lang
Wardrop_ joined #ruby-lang
Diaoer joined #ruby-lang
<tphummel>
i can accomplish what i'm trying for with an if statement. if satisfaction==false memo else do_my_logic(). but that doesn't seem as easy to follow as an early return/short circuit at the top of the reduce
towski joined #ruby-lang
<porcelina>
"require 'ray'; Ray.game 'hi' do end" crashes irb. :(
<Spooner>
Try the examples (in the gem folder)?
Xzyx987X_ joined #ruby-lang
<Spooner>
That does work for me though.
<porcelina>
it gives me a bad x opcode error. :\
<porcelina>
i think i didn't install it properly...
<porcelina>
yeah, the hello world example in the gem folder fails to create a window
<Spooner>
Actually, that snippet crashes for me in my Linux VM, but is fine in Windows with Ruby 1.9.3.
<porcelina>
what's it say on your linux?
<porcelina>
Major opcode of failed request: 1 (X_CreateWindow)
<porcelina>
this?
<Spooner>
Hmm, but Mon_Ouie says it works fine on his proper linux box.
<Spooner>
I just get /usr/share/ruby-rvm/gems/ruby-1.9.2-p180/gems/ray-0.2.0/lib/ray/game.rb:98: [BUG] Segmentation fault (but I just set up the VM and have not had Ray working on it yet).
<porcelina>
hmm
<Spooner>
Mon_Ouie is the person to talk to about it breaking. All I know is that it needs some sort of shader support (nothing advanced) and doesn't like Intel integrated cards.
<Spooner>
I did all my dev work on Win7 and I can't remember if I got it working in my old VM before I reinstalled the VM.
<Spooner>
This is one problem I found with Ray. Gosu has been around longer and has a larger user-base, so bugs on machines other than developer have had more opportunity to be fixed :)
<porcelina>
yeah
<porcelina>
idk, it's weird cuz if i can't open an xwin then... i shouldn't have a desktop environment. :\
<Spooner>
There is a #ray on this IRC server if you are interested in doing more with the library.
neoesque joined #ruby-lang
igotnolegs joined #ruby-lang
<Spooner>
You'd think so, wouldn't you?
thone_ joined #ruby-lang
<zenspider>
back
dv310p3r joined #ruby-lang
<zenspider>
what is RUBY_PLATFORM on jruby?
<headius>
"jruby"
<headius>
er
<headius>
platform? "java"
<headius>
RUBY_ENGINE is "jruby"
<headius>
trying to come up with the cleanest guard?
<headius>
maybe minitest should have some built-in guards? there's not that many impls
<headius>
platform == "java" is the classic way and works everywhere
<zenspider>
seems the safer bet
<zenspider>
kk
<headius>
unfortunately there's no RUBY_ENGINE on 1.8.7, which still annoys me
<zenspider>
right
brianpWins joined #ruby-lang
<zenspider>
is there ANY reason for an mri guard?
<headius>
if I write tests that MRI fails I might want to commit them until they're fixed
<shevy>
hmm perhaps yugui or whatever was her name could add RUBY_ENGINE
<headius>
MRI ain't perfect either, of course
<zenspider>
YES IT IS HOW DARE YOU
<headius>
shevy: matz said he didn't want to backport RUBY_ENGINE, unfortunately :(
amerine joined #ruby-lang
sulo joined #ruby-lang
<headius>
I sometimes wonder how he distinguishes changes that should be backported from changes that should not...we asked for this early in 1.8.7 cycle, and I doubt it would have hurt anyone to add it
<jammi>
mksm: thanks, but I think the Socket.ip_address_list will be fine. I'll just do a Socket.respond_to?(:ip_address_list) condition
<jammi>
a C extension as an extra dependency will just cause problems for people without compilers, like many windows users
kitallis joined #ruby-lang
<mksm>
yep
towski joined #ruby-lang
dr0id joined #ruby-lang
vauban joined #ruby-lang
twittard joined #ruby-lang
frangiz joined #ruby-lang
<postmodern>
jammi, that C-extension will probably not work on Windows, even when pre-compiled
<postmodern>
jammi, although you could copy it's logic and use FFI
<postmodern>
jammi, maybe even research the equivalent calls necessary on Windows
<postmodern>
if a C extension is less than 300 lines, generally it can be written in fewer lines with FFI
<jammi>
postmodern: yeah, that's usually the case for non-trivial pieces of C
vauban_ joined #ruby-lang
<postmodern>
jammi, and this code just introspects struct ifaddrs's
<jammi>
anyway, this is just a nice-to-have feature
<jammi>
nothing functionality depends on, just some nice status information to display
pabloh joined #ruby-lang
TheMoonMaster joined #ruby-lang
yxhuvud joined #ruby-lang
Banistergalaxy joined #ruby-lang
macmartine joined #ruby-lang
jredville joined #ruby-lang
Swimming_Bird joined #ruby-lang
s3mi0 joined #ruby-lang
s3mi0 joined #ruby-lang
kitallisii joined #ruby-lang
achamian joined #ruby-lang
ecin joined #ruby-lang
zmack joined #ruby-lang
edwardsharp joined #ruby-lang
tallship joined #ruby-lang
Asher joined #ruby-lang
amerine joined #ruby-lang
ilyam joined #ruby-lang
zerokarmaleft joined #ruby-lang
skipper joined #ruby-lang
burgestrand joined #ruby-lang
Tref joined #ruby-lang
michael_mbp joined #ruby-lang
dnjaramba joined #ruby-lang
JohnBat26 joined #ruby-lang
fridim_ joined #ruby-lang
heftig joined #ruby-lang
mytrile joined #ruby-lang
setmeaway joined #ruby-lang
[zaf]-Coral joined #ruby-lang
ddollar joined #ruby-lang
achamian joined #ruby-lang
TheMoonMaster_ joined #ruby-lang
robotmay joined #ruby-lang
sulo_ joined #ruby-lang
Tref_ joined #ruby-lang
burgestrand joined #ruby-lang
setmeaway joined #ruby-lang
michael_mbp joined #ruby-lang
dc5ala joined #ruby-lang
Manhose_ joined #ruby-lang
roadkith joined #ruby-lang
yorickpeterse joined #ruby-lang
<bnagy>
just had a weird regex question... 'asdf:01:aaa9' how to match \d+ which is NOT preceded by :, ie it should match 9 not 1
<erikh>
[^\d]\d+
<erikh>
you get the idea if you want to add a + after the c-class or w/e
<erikh>
ah, sorry: [^:]\d+
<zedUNDginger>
what about my girl /(?!:)(\d+)/
<Ruthenium>
that would still match 1
<bnagy>
? that totally fails to even try to do what I said :)
<erikh>
ok
<erikh>
have fun
<zedUNDginger>
bnagy: do u like my girl
<zedUNDginger>
this is my girl /(?!:)(\d+)/
<bnagy>
that will match 1
<zedUNDginger>
oh
twittard joined #ruby-lang
<bnagy>
erikh: when I said it totally fails it referred to the first answer, the second one is the first obvious thing to try, followed by the negative lookbehind
ryanf joined #ruby-lang
Mchl joined #ruby-lang
thrcka joined #ruby-lang
Wardrop joined #ruby-lang
<bnagy>
oh... maybe /[^:\d](\d+)/
<bnagy>
I bet that's recursive to hell :S
solars joined #ruby-lang
<shevy>
yeah
<shevy>
that shows how many here understand regexes really well
molgrew joined #ruby-lang
<bnagy>
not understanding regex is a defensive mechanism
<bnagy>
yeah it's not recursive at all
Manhose_ joined #ruby-lang
hebz0rl joined #ruby-lang
<erikh>
it backtracks
<erikh>
that's not recursion.
gnufied joined #ruby-lang
<erikh>
I never use zero-width and explicit negative stuff personally because it always breaks on me. backtracking is reliable and easy to read.
<erikh>
(and also nice and posix compliant)
jasox joined #ruby-lang
mksm joined #ruby-lang
srbaker joined #ruby-lang
stepnem joined #ruby-lang
<bnagy>
I don't think I have ever used negative assertions.. btu then from memory the ruby engine didn't support them once, right?
<bnagy>
I have used positive lookahead, rarely
<bnagy>
but in general I hate having to use regex :/ This was just a weird question because it wasn't immediately obvious how to do it
gix joined #ruby-lang
cyri_ joined #ruby-lang
<shevy>
hehe
<shevy>
I don't mind regexes as long as they stay simple!
<RomyEatsDrupal>
hi… rubynuby here with a a very brief ( less than 20 lines) testing script she wrote in an attempt to learn readers and writers… can someone help explain why my script ran into errors ?
<injekt>
sopheak: each animal class is a subclass of Animal
<MistyM>
sopheak: Dog.class == Dog, but Dog.superclass == Animal
<sopheak>
yep, that's in ruby, right?
tomzx joined #ruby-lang
<injekt>
that's in any language which implements inheritance
<MistyM>
(Or rather, Dog.class.superclass)
<sopheak>
ah i see, that's my confusion
<sopheak>
thank for correcting :)
<porcelina>
well, you'd be right in java tho
<porcelina>
if you say Animal woggus = new Dog();
<porcelina>
you couldn't call woggus.bark() because java would only know about the superclass
<porcelina>
but in ruby you could call bark
<porcelina>
and if it happened to be a cat it'd just throw an error
<sopheak>
procelina, that's correct
<sopheak>
unless the superclass has bark()
<porcelina>
yea
<sopheak>
ruby is better then :)
<porcelina>
yea ruby is significantly easier to deal with
<robbrit>
duck typing ftw
<sopheak>
yeah, but i'm not yet familiar with all the sign
<porcelina>
well, it's not quite duck typing, because there's no typesafety whatsoever
<sopheak>
just start from last two days
<porcelina>
unless you write x.methods.contains or whatever
<injekt>
porcelina: that's what respond_to? is for
<injekt>
perfect for duck typing
<porcelina>
oh right i forgot about that. ^_^
<injekt>
you dont care if your object is a Duck, you just care if it .quacks() :)
<porcelina>
i'd still rather be able to overload params tho
<porcelina>
that's the one thing i'm missing that true duck typing solves
<injekt>
true duck typing?
<porcelina>
like, actual typechecking
<porcelina>
and not manual
<injekt>
that's not duck typing then
<injekt>
if you're type checking
<porcelina>
you're checking that a type responds to something
<injekt>
right
<porcelina>
but the language does it for you
<injekt>
you dont care about the type
<porcelina>
ruby won't check for you
<porcelina>
except to throw an error if it's wrong
<injekt>
of course
<porcelina>
so the only way to overload a method is to throw a ton of if's in it
<injekt>
i dont understand what you'd want
<injekt>
as an alternative
<injekt>
if you want method overloading you already want a static type system
<porcelina>
something like def make_noise (animal with quack); animal.quack; end
<porcelina>
def make_noise (animal with bark); animal.bark; end
<injekt>
ahh rather than checking for if respond_to? everywhere
<porcelina>
yeah
<injekt>
if you have a dog and a duck which is of type animal
<porcelina>
i'm sure they'd be equivalent, just prettier syntax for it
<porcelina>
doesn't really matter what type they are
<injekt>
you should be using the same method to talk
kyrylo joined #ruby-lang
<injekt>
this is where interfaces in java are nice
<porcelina>
yea
<injekt>
defining a spec
<sopheak>
may i ask you what these statement mean?
<sopheak>
def make_noise (animal with quack); animal.quack; end
<injekt>
describing what porcelina would like
phlipper joined #ruby-lang
<porcelina>
uhm, they're how i want ruby to be. ^_^;
<sopheak>
ah :)
<porcelina>
they're not actual ruby
<injekt>
I like that there's only ever one method signature
solars joined #ruby-lang
<porcelina>
i think it's actually necessary
<porcelina>
since you can add methods at runtime
<porcelina>
it wouldn't know whether you wanted to override or overload
<injekt>
well, it would just always override
<porcelina>
yea i guess
<porcelina>
crap, i'm late for school. :\
<porcelina>
tty
<porcelina>
ttyl*
imajes joined #ruby-lang
<sopheak>
thanks guys for all of your help
<sopheak>
i'm 1AM here, ttyl*
<j`ey>
np glad to help
io_syl joined #ruby-lang
io_syl joined #ruby-lang
srbaker joined #ruby-lang
outoftime joined #ruby-lang
RomyEatsDrupal joined #ruby-lang
srbaker joined #ruby-lang
musl joined #ruby-lang
ivorybishop joined #ruby-lang
yorickpeterse joined #ruby-lang
srbaker joined #ruby-lang
necromancer joined #ruby-lang
<necromancer>
hey guys
<necromancer>
when shell-scripting with ruby, is it ok to just do "foo bar" when i want to print to screen, or do i have to use puts()?
RickHull joined #ruby-lang
<MistyM>
No, it won't be printed to screen unless you explicitly print it. (Except in irb, of course.)
publicvoid_ joined #ruby-lang
RickHull joined #ruby-lang
<necromancer>
MistyM: thanks
<MistyM>
No problem!
<necromancer>
MistyM: so one more...i'm basically writing an init.d script for this server called Red5. it is started by running a script called `red5.sh`, and outputs all of its logs to the terminal. so when i close my session, it kills the server.
butchanton joined #ruby-lang
<necromancer>
MistyM: i'm writing an init.d script to launch on startup and also keep it running as a bg process. what i'd like to do is test whether my `exec` call to red5.sh worked. is there a way to capture the exit code using ruby's backtick syntax?
<MistyM>
necromancer: Exit status is stored in the special global variable $? (until the next time a process is spawned, then it gets overwritten)
<MistyM>
Also, are you capturing the stdout using the backtick syntax? If not, #system is probably the one you're looking for - it returns true for a zero exit status, false for non-zero. (Exit status is still available by $?)
<necromancer>
ok
<necromancer>
oh wow thanks workmad3
<necromancer>
:)
<workmad3>
necromancer: what are you doing with red5 then? :)
RickHull joined #ruby-lang
<necromancer>
workmad3: you don't wanna know ;-)
<workmad3>
necromancer: as long as it's not 'attempting to compile it' then it can't be too bad :P
<necromancer>
workmad3: this client required me to build them a SWF that allows them to record voice to the Red5 server, then once the recording is done it adds to a queue which runs ffmpeg to convert the FLV to WAV, then sox to mix the WAV with another background music track, then ffmpeg again to transcode to MP3
<workmad3>
necromancer: interesting :)
<necromancer>
workmad3: all so people can eventually purchase these tracks in this music store-like environment, but it's for like spoken word over bg music for trauma therapy
<necromancer>
oh yeah i can't imagine compiling it
<necromancer>
i had to compile ffmpeg though that was a chore
<workmad3>
necromancer: tbh, the biggest problem with red5 is that they can (and do) break anything and everything between versions... if you have a version that works with your app *save it somewhere and never delete it*
<workmad3>
(hence why I forked that version specifically for the company repo :) )
saysjonathan joined #ruby-lang
<necromancer>
workmad3: interesting...i'll have to keep that in mind
saysjonathan joined #ruby-lang
<necromancer>
workmad3: how did you fork it from svn to git? i had to get my copy off svn i believe
<workmad3>
necromancer: I just committed the version provided by the contractors who did the red5 stuff to git
<necromancer>
workmad3: that looks just like my /usr/share/red5 dir actually...is it possible to just package that up in a git repo?
<necromancer>
my thought is i can just clone the repo to /usr/share/red5 on any machine and it will "just work"
<necromancer>
no sense building from ant if i don't have to right?
<workmad3>
yeah :)
<workmad3>
I build the red5 application with maven... I spent a couple of days with that at one point making sure all the dependencies built into the app didn't conflict with the stuff in the lib/ dir of red5
<workmad3>
but the server... no way I'm going to touch that if at all possible
outoftime joined #ruby-lang
<necromancer>
workmad3: yeah i was very confused by red5 and its lack of docs. having not touched java in almost 6 years now, it was very daunting. so we had to solve it in other ways. i would have rather just done all of this work using the Red5 app. i think it's possible to do everything i'm doing totally inside the Red5 app since it's java and it's got the stdlib
<necromancer>
but right now the queue is handled by a PHP script (boss's orders unfortunately) and a series of shell scripts executed by php exec()
<necromancer>
err shell commands*
achamian joined #ruby-lang
<href>
haha, necromancer: epic nickname.
kitallis joined #ruby-lang
ecin joined #ruby-lang
kain joined #ruby-lang
<workmad3>
necromancer: yeah, red5 is basically a java webapp (the default server runs embedded tomcat or jetty, not sure which). It can do all sorts of good long running service stuff
<necromancer>
workmad3: oh cool. i think it runs tomcat iirc, since ant is part of the tomcat project
<necromancer>
haha thanks href
towski joined #ruby-lang
[zaf]-Coral joined #ruby-lang
<workmad3>
necromancer: ant is done by apache, same as tomcat yeah... but jetty is an apache project nowadays too ;)
<workmad3>
oh wait, no, jetty is an eclipse project
<necromancer>
hmm
<necromancer>
how is eclipse doing these days? last i checked it was buggy and crashed a lot ;)
<workmad3>
but still, red5 is neither and can make use of whatever it feels like :)
<workmad3>
I don't use it... I use vim :P
edwardsharp joined #ruby-lang
benanne joined #ruby-lang
mrsolo joined #ruby-lang
sulo joined #ruby-lang
robbyoconnor joined #ruby-lang
brianpWins joined #ruby-lang
heftig joined #ruby-lang
amerine joined #ruby-lang
r0bby joined #ruby-lang
hahuang65 joined #ruby-lang
phlipper joined #ruby-lang
RickHull joined #ruby-lang
NemesisD joined #ruby-lang
<NemesisD>
hi all. im trying to figure out how to sanely monkey patch a Module such that my method gets inserted into the call chain but can "super" into the module's method
<NemesisD>
the only way i've found to do it is to instead target the class that the original module is included into, everything else outright replaces the original module's method
r0bby joined #ruby-lang
io_syl joined #ruby-lang
io_syl joined #ruby-lang
jmontross joined #ruby-lang
RegEchse joined #ruby-lang
Squarepy joined #ruby-lang
sepp2k1 joined #ruby-lang
imperator joined #ruby-lang
Squarepy joined #ruby-lang
perryh_ joined #ruby-lang
Squarepy joined #ruby-lang
Squarepy joined #ruby-lang
twittard joined #ruby-lang
<macmartine>
I downloaded the Ruby docs from http://ruby-doc.org/stdlib-1.9.3/ and there's no css so it looks like crap -- pretty unusable. Is there another source?
Indian joined #ruby-lang
<Mon_Ouie>
You can generate the documentation from the source code with rdoc or yard
RomyEatsDrupal joined #ruby-lang
<macmartine>
K, thanks. Weird the d/loads would be like this.
Swimming_Bird joined #ruby-lang
io_syl joined #ruby-lang
io_syl joined #ruby-lang
RickHull joined #ruby-lang
RickHull joined #ruby-lang
r0bby joined #ruby-lang
RegEchse joined #ruby-lang
kitallis joined #ruby-lang
jmontross1 joined #ruby-lang
josh9 joined #ruby-lang
jkprg_ joined #ruby-lang
<josh9>
can i use Ruby 1.9.3-p0 or will i face compatabilities isssues with gems?
<mihar>
1.9.3 works fine for me
<mihar>
but i could be just lucky
Carnage\ joined #ruby-lang
<theconartist>
read the release notes josh9 they aren't that long
<josh9>
theconartist: sure. thanks
<theconartist>
and it would be pretty hard to find any problems upgrading
<yxhuvud>
josh9: ask again in a year when more people have tried upgrading :)
<lianj>
nothing serious, but ruby -e 'p :a,:b,'; ruby -e 'p(:a,:b,)' in 1.9.3 is still strange. bug or intended?
robotmay joined #ruby-lang
<lianj>
josh9: most of what works in 1.9.2, is fine. only this syntax change broke some of my libs :Ð
<macmartine>
params[:user]['service_ids'] ||= [] fails with "<IndexError: string not matched>". I see why, but how do I set it to an empty array if the key doesn't exist?
<zenspider>
IndexError ?
<zenspider>
meh. #rubyonrails
<macmartine>
Yes, it's try to access the index 'service_ids' of a hash. But that key doesn't exist
<zenspider>
real hashes return nil when they don't match the key
<erikh>
dollars to donuts though zenspider is right
<slyphon>
imperator: :D
<zenspider>
damnit... now I want popcorn
<imperator>
erikh, ah yes that's the one
WillMarshall joined #ruby-lang
<imperator>
net-ssh for me, with its 3 bajillion prereqs
<erikh>
heh
<zenspider>
imperator: :((((
<imperator>
zenspider, the perl module
<zenspider>
then add some more ('s
<zenspider>
ruby's is worth 4
<zenspider>
perl is prolly 8? 9?
<macmartine>
zenspider: ah, of course. params is a hash, but params[:user] isn't. long day. thanks
<imperator>
i think the total prereqs for net-ssh-perl, with some of the optional stuff installed, came out to over 50 dependencies last time i checked
<macmartine>
erikh: ^^ thanks
<imperator>
one ( for each one, i say
<erikh>
macmartine: enjoy
<zenspider>
hah. that's a totally fair measure. someone should write a gem for that