Bwild has quit [Remote host closed the connection]
Bwild has joined #rubinius
Bwild has quit [Ping timeout: 252 seconds]
elia has quit [Ping timeout: 265 seconds]
_elia has joined #rubinius
nirvdrum has quit [Ping timeout: 252 seconds]
gtemple has joined #rubinius
pitr-ch has quit [Quit: Computer has gone to sleep.]
<yorickpeterse>
YAY more AWS errors
<yorickpeterse>
can't even fucking schedule some data because the aws-sdk shits itself
elia has joined #rubinius
_elia has quit [Ping timeout: 240 seconds]
elia has quit [Quit: (IRC Client: textualapp.com)]
elia has joined #rubinius
nirvdrum has joined #rubinius
<cremes>
@brixen: I know you have a lot of priorities, but is one of them getting travis-ci to use an rbx newer than 2.1.1? I have a feeling that’s waiting on your build system improvements...
<yorickpeterse>
cremes: Travis has 2.5
<yorickpeterse>
use rbx-2 in your config file
<cremes>
hmmm, maybe my travis.yml is out of date or something. what’s the blessed syntax these days to get the latest rubinius?
pitr-ch has quit [Quit: Computer has gone to sleep.]
meh` has joined #rubinius
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<amsi>
cremes: IO#popen runs a program on the system, but instead of the process's stdin and stdout going to the terminal, they're connected to #read and #write (or equivilent helpers) of the object IO#popen returns
<cremes>
amsi: yeah, i get that much. part of what confuses me is that i don’t understand how the spec *ever* works.
<cremes>
how does the child ever get the “if” statement to execute? it isn’t inside a block!
<amsi>
off the top of my head my guess would be it's using fork magic, where str ends up holding a reference in the parent but is nil in the child. so that if statement would be executed after the fork happens
<cremes>
amsi: hmmm, that might be the trick. i do see a fork happening here. thanks for the idea!!
<amsi>
np, good luck!
<amsi>
cremes: i just found in the docs where it says "A string equal to "|-" will create another Ruby instance as a subprocess.", so IO.read("|-") is doing that fork
<cremes>
amsi: yes, i know that much. i’m currently up to my eye-balls in an IO class refactor. I’m trying to get the read_specs to pass and the #popen code is a monster of inscrutability!
<yxhuvud>
that is some serious black magic
<amsi>
ah, i see. that's a big undertaking. i've been dealing a lot with socket IO lately; it seems like the IO class is a crazy world with unexpected edge cases
<cremes>
amsi: you ain’t kidding
gtemple has quit [Ping timeout: 240 seconds]
cypher23 has quit [Ping timeout: 264 seconds]
sferik has joined #rubinius
cypher23 has joined #rubinius
cypher23 has quit [Ping timeout: 246 seconds]
cypher23 has joined #rubinius
GitHub84 has joined #rubinius
<GitHub84>
[rubinius] chuckremes pushed 4 new commits to ffi-io: http://git.io/FKuQ
<GitHub84>
rubinius/ffi-io 2399e9e Chuck Remes: simplify by taking advantage of existing class method
<GitHub84>
rubinius/ffi-io 037cd57 Chuck Remes: add #pipe function
GitHub84 has left #rubinius [#rubinius]
<GitHub84>
rubinius/ffi-io dddd972 Chuck Remes: add a convenience method for testing call failure
<cremes>
what’s the recommended way to pull down new commits from master and merge them to a local branch?
<cremes>
i’m guessing it’s…
<cremes>
git checkout master
<cremes>
git pull
<cremes>
git checkout <branch>
<cremes>
git merge master
<cremes>
is that right?
<yorickpeterse>
You can just pull straight from a branch
<yorickpeterse>
and if you pull with --rebase you don't get merge commits
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sferik has joined #rubinius
<|jemc|_>
all my git repositories on github are actually hg repositories on my local machine that get translated by hg-git - although it'd be nice if github started supporting hg like bitbucket
<yorickpeterse>
like I said, the UI is shit
<yorickpeterse>
granted I'm too used to it by now, but I'd love for Git to sort that out
<yorickpeterse>
(one can only dream)
<yorickpeterse>
But the branching support is pretty darn baller
<|jemc|_>
git branches are pretty much equivalent to hg bookmarks
<yorickpeterse>
granted hg and git are quite similar in various ways
<|jemc|_>
in fact hg-git translates them as such and I have no problems interacting with github that way
<|jemc|_>
plus hg has anonymous branching so if you have two developers working on the same branch, they don't have to merge until they're ready
<|jemc|_>
and you don't have to update every time you pull from a remote - update is a separate command similar to checkout
* |jemc|_
stops evangelizing now
slaught has joined #rubinius
<yorickpeterse>
ooh, also almost 4 years ago since I started using Vim
<|jemc|_>
yorickpeterse: have you given your talk yet?
<yorickpeterse>
Caturday
<yorickpeterse>
I'll see if I can share the stream link once it's up
<yorickpeterse>
although I might not have wifi before/during the talk
<|jemc|_>
want to add a slide comparing valgrind output between rbx and MRI? ;)
<yorickpeterse>
eh, that's out of the scope, I'm not talking about anything performance related
<|jemc|_>
it's not performance, it's leaks and invalid memory management
<|jemc|_>
(valgrind --tool=memcheck)
<|jemc|_>
just thought I'd mention it as rbx only had a few problems compared to MRI where the problems were staggering and brixen and I have been working to clean up the rbx problems
<yorickpeterse>
I know, but it's out of the scope since I won't be talking about performance/memory/etc
<yorickpeterse>
It's primarily a talk about the process behind rbx and some examples (String#scrub and the venerable Kernel#caller_locations in this case)
<cremes>
git pull REMOTE master - doesn’t work
<cremes>
do i have to setup REMOTE to be something specific?
<yorickpeterse>
well, REMOTE has to point to whatever remote you want to pull from
<slaught>
cremes: REMOTE is just any remote uri you have setup. see ‘git remote -v’ for the current setup and git remote command lets you add new ones
<cremes>
geez, i never took the time to learn git except for the basics.
<cremes>
i have a clone of the rubinius repos. i have a commit bit for it, so i can commit directly. i created a branch and pushed it to the rubinius repos. how do i get the remote master to merge in with my local branch? if the answer is “set the remote” then i’ll start googling for help on that.
noop has quit [Ping timeout: 276 seconds]
<heftig>
brixen: melbourne doesn't parse this: "def a b:; c []; end" if you use linebreaks instead of semicolons
<heftig>
actually, it's the linebreak after : that's significant
<heftig>
eval "def a b:; c []; end" works but eval "def a b:\n c []; end" doesn't
<yorickpeterse>
or just `git pull --rebase origin master` (when on your own branch)
<cremes>
oh shit… i followed your first way. on the ‘git rebase master’ step it is going through all of my *old* commits and complaining about merge conflicts.
<cremes>
do i really have to go through every fucking commit from the beginning and resolve conflicts? argh… this sucks.
slaught has quit [Quit: slaught]
elia has joined #rubinius
slaught has joined #rubinius
slaught has quit [Client Quit]
<|jemc|_>
heftig: can you open an issue (or perhaps a PR with a spec)?
<cremes>
something is fucked up with my branch history. it loops on some of the commits. maybe someone can pull down the ffi-io branch and look at it? i think it’s fucked.
<cremes>
i need help unfucking it. this is way over my (git) head.
Bwild has joined #rubinius
<cremes>
okay, got my git woes worked out thanks to headius.
<cremes>
master is merged into my branch. no conflicts. yippee...
sferik has quit [Read error: Connection reset by peer]
sferik has joined #rubinius
tenderlove has quit [Quit: Leaving...]
tenderlove has joined #rubinius
diegoviola has joined #rubinius
diegoviola has quit [Client Quit]
|jemc|_ is now known as |jemc|
diegoviola has joined #rubinius
diegoviola has quit [Read error: Connection timed out]
diegoviola has joined #rubinius
slaught has joined #rubinius
slaught has quit [Client Quit]
diegoviola has quit [Quit: WeeChat 1.1.1]
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]