josh-k has quit [Remote host closed the connection]
benlovell has quit [Ping timeout: 272 seconds]
mbj has quit [Ping timeout: 256 seconds]
mbj has joined #rubinius
havenwood has quit [Ping timeout: 246 seconds]
havenwood has joined #rubinius
diegoviola has joined #rubinius
diegoviola has quit [Remote host closed the connection]
diegoviola has joined #rubinius
|jemc| has quit [Quit: WeeChat 0.4.3]
flavio has quit [Quit: WeeChat 0.4.1]
mbj has quit [Ping timeout: 244 seconds]
mbj has joined #rubinius
diegoviola has quit [Ping timeout: 246 seconds]
diegovio1 has joined #rubinius
diegovio1 is now known as diegoviola
meh` has joined #rubinius
havenwood has quit [Remote host closed the connection]
elia has quit [Quit: Computer has gone to sleep.]
elia has joined #rubinius
elia has quit [Client Quit]
|jemc| has joined #rubinius
elia has joined #rubinius
mbj has quit [Quit: leaving]
elia has quit [Quit: Computer has gone to sleep.]
amclain has joined #rubinius
slaught has joined #rubinius
slaught has quit [Client Quit]
tenderlove has quit [Quit: Leaving...]
havenwood has joined #rubinius
|jemc| has quit [Quit: WeeChat 1.0.1]
|jemc| has joined #rubinius
diegovio1 has joined #rubinius
diegoviola has quit [Ping timeout: 265 seconds]
byprdct has quit []
arrubin has joined #rubinius
DanielVartanov has joined #rubinius
diegovio1 is now known as diegoviola
havenwood has quit [Remote host closed the connection]
havenwood has joined #rubinius
DanielVartanov has quit [Remote host closed the connection]
<brixen>
well, finally openssl C-ext compiles again
<brixen>
I wonder if it will segv too
<brixen>
MRI sure doesn't make it easy to run tests
yxhuvud has quit [Remote host closed the connection]
postmodern has joined #rubinius
kagaro has quit [Ping timeout: 240 seconds]
kagaro has joined #rubinius
<yorickpeterse>
I fucking hate C
<yorickpeterse>
and null pointers
diegoviola has quit [Quit: WeeChat 1.0.1]
<yorickpeterse>
Oga's XML lexer shits itself when 1) you're reading in input as an IO, per line 2) it bumps into an attribute value with a newline
<yorickpeterse>
because then the pointer to the character is suddenly NULL
<yorickpeterse>
when it's supposed to be at _least_ "\n"
|jemc| has quit [Ping timeout: 250 seconds]
<Rotonen>
yorickpeterse: an IRL thing i have once run into the wild: a BOM in the middle of an xml attribute keyword
<Rotonen>
yorickpeterse: obviously from a bug, but paved over with a 'rules and assumptions' datasheet, due to having passed certification and was not to be certified again
<yorickpeterse>
heh
<yorickpeterse>
I have no idea why this even breaks when there's a newline, I suspect Ragel somehow fucks up its state
<yorickpeterse>
so ragel has two pointers: ts (token start) and te (token end), ts is NULL here, te is a string like it should be
<yorickpeterse>
because....logic
<Rotonen>
random amounts of BOM in xml in SOAP can really make your day :>
<yorickpeterse>
it then craps out because it's basically doing (te + 1) - (NULL + 1)
<yorickpeterse>
well, time to crank out gdb and see what the heck is going on
<yorickpeterse>
which....doesn't even work
<yorickpeterse>
since it's not a C error, but a Ruby error
<yorickpeterse>
great
elia has joined #rubinius
<brixen>
yorickpeterse: I updated openssl lib on os x and the openssl C-ext starts segfaulting because C
<yorickpeterse>
sounds reasonable
<yorickpeterse>
something-something if Ruby was webscale I wouldn't have to use C
<brixen>
yorickpeterse: I'm coming to the opinion that we must make something that is 10-100x faster to write than C and no more than 5x slower running
<brixen>
the only way to replace C is to replace the things written in C
<yorickpeterse>
brixen: you mean like Go? :P
<brixen>
but that's a lot of work so you must be able to work much, much faster
<brixen>
no, I don't mean like Go or Rust for that matter
<brixen>
you're not going to be 10x faster writing an app in Rust
<brixen>
even if it's a much nicer language to start with
<brixen>
which it is
<yorickpeterse>
it was a joke :P
<brixen>
jokster
<brixen>
go write some C
<brixen>
:)
<brixen>
but the point is, people focus too much on the perf
<brixen>
without appreciating the need to be much, much faster to write
<brixen>
IMO
<yorickpeterse>
well, there's a pretty big difference between a 3,5 second runtime, and 500ms (in my case)
<yorickpeterse>
with those 3,5 seconds being the bare/bear minimum
<brixen>
well, you'd be outside my 5x there
<brixen>
of course, that's not a hard number but yeah, you can't take 3.5s to do something the user won't wait 500ms for
<brixen>
pretty sure Java is still "the language of choice for high performance, low latency applications"
<brixen>
re "Node.js is the language of choice for high performance, low latency applications and has been powering everything from robots to API engines to cloud stacks to mobile web sites."
<brixen>
and since when is Node.js a language
<brixen>
this is from Joyent itself
<brixen>
apparently Node is still an elaborate practical joke
<havenwood>
i've always suspected
<yorickpeterse>
brixen: you're forgetting one thing here
<brixen>
the language in this post reminds me of the very early "what is Node" video with sound bites from people who were painfully clueless
<yorickpeterse>
brixen: Joyent is Joyent, they'd be stupid not to boast Node as something other than "THE BEST DARN THING SINCE SLICED BREAD"
<brixen>
yorickpeterse: everybody boasts that
<yorickpeterse>
Meanwhile I'll happily wait for Node to actually do parallelism properly
<Rotonen>
node is fairly good for avoiding iowait on a single core virtual machine ie. small amazon instances
<Rotonen>
but for that as a driving design metric i actually prefer python3/tornado
<Rotonen>
for real hardware with SMP or NUMA (or HMP for freaky tech of future), things get as hairy and complex as ever
<Rotonen>
and there might be a conclusion drawn why i hang around here and like this particular project quite a lot :-)
<brixen>
OTOH, I'm curious to see if they can pull off the board thing
<brixen>
Ruby can't do that
<brixen>
one more thing Node can do that Ruby can't