<yorickpeterse>
"vm/llvm/artifacts/jit_compiler.cpp.o:(.note.stapsdt+0xa8): undefined reference to `rubinius_jit__function__end_semaphore'" what the fuck
<IPGlider>
has anyone been able to configure and build on osx yosemite? I'm getting "Checking for 'llvm-config': not found" and it does not download a prebuilt llvm
<yorickpeterse>
IPGlider: prebuilt support has been removed
<yorickpeterse>
make sure that you brew install llvm and that it has been linked
<yorickpeterse>
brew link llvm # or something like that
<IPGlider>
oh, ok thank you!
|jemc| has joined #rubinius
noop has joined #rubinius
slaught has joined #rubinius
<yorickpeterse>
right, so using clang doesn't work either
<yorickpeterse>
argh
<|jemc|>
yorickpeterse: your undefined reference thing sounds like something is not getting linked
<yorickpeterse>
I suspect it's rbx trying to include extra shit related to dtrace
<yorickpeterse>
but since I don't truly have dtrace it craps out
<yorickpeterse>
since it only does this when I pass --with-dtrace
<yorickpeterse>
tonight's todo: see if I can hack in metrics for cache invalidations (besides inline cache), and get fuckign systemtap working
IPGlider has quit [Quit: WeeChat 1.0.1]
meh` has joined #rubinius
JimmySieben has joined #rubinius
<yorickpeterse>
Hm, I reckon I can add systemtap/dtrace support to Oga
<headius>
interesting choice...I can't say I've heard of an XML library that exposes dtrace probes
<headius>
of course, if you do that, it won't work anywhere that doesn't have dtrace
<yorickpeterse>
it would be conditionally of course
<yorickpeterse>
But come to think of it, the C side of things is tiny compared to the test
<yorickpeterse>
* rest
<yorickpeterse>
not sure what use the probes would be
<headius>
yeah, I couldn't think of anything :-)
<yorickpeterse>
heading home, bbl
jeremyevans has quit [Ping timeout: 255 seconds]
jeremyevans has joined #rubinius
amsi has joined #rubinius
tenderlove has quit [Quit: Leaving...]
tenderlove has joined #rubinius
diegoviola has joined #rubinius
JimmySieben has quit [Quit: JimmySieben]
Rotonen has quit [Ping timeout: 250 seconds]
JohnBat26 has joined #rubinius
abbe has joined #rubinius
byprdct has quit [Ping timeout: 265 seconds]
byprdct has joined #rubinius
havenwood has quit [Remote host closed the connection]
havenwood has joined #rubinius
<yorickpeterse>
hm, so the probes file is properly generated
<yorickpeterse>
which leads me to suspect it might not be linked properly
noop has quit [Ping timeout: 240 seconds]
amsi has quit [Ping timeout: 265 seconds]
postmodern has quit [Quit: Leaving]
jnh has joined #rubinius
<jnh>
good morning
JohnBat26 has quit [Remote host closed the connection]
josh-k has joined #rubinius
<yopp>
hm. how do you think what are the key points for ruby crash course?
<brixen>
yopp: do you mean a course on dealing with Ruby crashes?
<brixen>
or a rapid learning of Ruby?
<yopp>
;) rapid deep learning
<brixen>
rapid deep learning I'm not sure
<brixen>
but learn to program if you don't know Ruby at all
<brixen>
+ sandi metz's book
<brixen>
practical object-oriented design in Ruby
<brixen>
the first is elementary but covers the syntax well
<brixen>
the second is essential Ruby knowledge
<yxhuvud>
ime, deep OO skills can easily jump the language barriers, but weak OO skills will write soup whatever the language.
<yxhuvud>
sometimes I feel sandi should be clobbered by a yagni hammer, but there is a lot of stuff as well.
<yxhuvud>
*lot of good stuff
<jc00ke>
yxhuvud: maybe yagniy (yagni+yet) but it's good to know the concepts regardless.
JimmySieben has joined #rubinius
<yxhuvud>
indeed.
<yxhuvud>
the problem with the yet part is that it might as well be never, as demands shift over time.
<jc00ke>
yxhuvud: sure, but when there is demand... then you have the tools! She definitely doesn't encourage over-engineering, that's fo sho
<yxhuvud>
some of her talks have been quite painful in that regard. But yes, learn to recognize badly structured code and learn ways to make it less bad.
<yxhuvud>
don't forget that sometimes the best teacher is actually working with the problem. perhaps let them solve some simple things doing pair programming, and then discuss the experiences and solutions afterwards?
<brixen>
yopp: one of our staff made that
<brixen>
yopp: it focuses on some Rails stuff by tries to build good Ruby practices
<brixen>
s/by/but/
<yopp>
uh. my whole idea is to address "rails developers rarely give a shit about ruby"
byprdct has quit []
<brixen>
we very much expect our Rails devs to give a shit about Ruby
<brixen>
there are 18 of them sitting in a training course with Sandi at this very moment :)
<yxhuvud>
yes that is also a good point. what will they use ruby for? if rails, maybe look into how some rails magic is implemented, and similar for chef or puppet if you are using either of those
<yopp>
yxhuvud, yeah, something pretty much like: theory block and then solving practical problem with this new theory
<yopp>
yep, about rails I'd like to write a tiny ORM and tiny analogue of action_pack with routing to give an idea how this solved in rails
<yopp>
uh, no free shipping for old EU man
Rotonen has joined #rubinius
<yopp>
brixen, thanks for the book, samples looks very nice
<yorickpeterse>
yopp: didn't read all, what level is the target group?
<yorickpeterse>
Total beginners?
<yopp>
yorickpeterse, average programming skills, weak ruby skills
<yorickpeterse>
Ah
<yorickpeterse>
ok that greatly simplifies things
<yopp>
think of average-salary rails programmer ;)
<yorickpeterse>
https://rubymonk.com/ this used to be decent, not sure if it still is
<yorickpeterse>
hey I'm an average-salary ruby programmer
<yopp>
see, you don't need this courses!
<yorickpeterse>
I'd be worried if I did :P
<yopp>
"The `lambda` keyword "
<yopp>
oh my
<yorickpeterse>
yopp: honestly the best way to learn as an intermediate is to just do shit
<yorickpeterse>
That is, don't bother too much with papers/books/etc, get down and dirty with some code
<yopp>
yep, that's why I'd like to spent 60% of time writing code
<yopp>
•spend
<yopp>
and not just dummy foobar code, but solving some real world problems
<yopp>
And real problems are welcome :)
byprdct_ has joined #rubinius
<yorickpeterse>
*cough* rbx *cough*
<yopp>
yorickpeterse, mostly they are needs good knowledge of rbx. but if you have some "average level" issues in mind, it will be awesome
<yorickpeterse>
hm
<yorickpeterse>
Hm, seems Rbx doesn't link the dtrace probes when building the vm somehow
<yorickpeterse>
that is, I don't see any mention of it in the gigantic daedalus output