mishok13 has quit [Read error: 110 (Connection timed out)]
ikaros has quit ["Leave the magic to Houdini"]
tmaeda is now known as tmaedaZ
ulfdoz_ has joined #ocaml
munga_ has quit [Read error: 113 (No route to host)]
ulfdoz has quit [Connection timed out]
ulfdoz_ is now known as ulfdoz
B1DOrD has quit [Read error: 110 (Connection timed out)]
eldragon has left #ocaml []
ski__ has quit ["Lost terminal"]
_unK has quit [Remote closed the connection]
_unK has joined #ocaml
ski__ has joined #ocaml
seanmcl has quit []
rothwell has joined #ocaml
johnnowak has joined #ocaml
<johnnowak>
hello all. has anyone managed to successfully managed to call ocaml (compiled via ocamlopt) from C on a platform? It works fine for me on OS X but fails to link due to missing symbols on linux and freebsd.
<johnnowak>
*ahem* on any platform other than OS X rather.
seanmcl has joined #ocaml
seanmcl has quit []
jonafan has quit [SendQ exceeded]
<infoe_>
hm, im getting build errors for godi-lablgl on conflicting types for Togl_PostRedisplay, then several others
johnnowak has quit []
ygrek has joined #ocaml
Submarine has joined #ocaml
deavid has quit [SendQ exceeded]
rothwell has left #ocaml []
beauty has joined #ocaml
<beauty>
morning
deavid has joined #ocaml
beaut has joined #ocaml
beauty has quit [Read error: 110 (Connection timed out)]
beaut is now known as beauty
ttamttam has joined #ocaml
_zack has joined #ocaml
jcaose has joined #ocaml
Snark has joined #ocaml
ttamttam has quit ["Leaving."]
munga_ has joined #ocaml
_unK has quit [Remote closed the connection]
Associat0r has joined #ocaml
_zack has quit ["Leaving."]
_unK has joined #ocaml
verte has joined #ocaml
thelema has quit [Remote closed the connection]
thelema has joined #ocaml
sramsay has joined #ocaml
sramsay has quit [Client Quit]
<Camarade_Tux>
morning :)
beauty has quit ["YourCodeSucksException"]
julm has quit [Remote closed the connection]
julm has joined #ocaml
valross has joined #ocaml
ikaros has joined #ocaml
bind_return has quit [Read error: 110 (Connection timed out)]
bind_return has joined #ocaml
slash_ has joined #ocaml
jcaose_ has joined #ocaml
Yoric has joined #ocaml
mishok13 has joined #ocaml
jcaose has quit [Read error: 110 (Connection timed out)]
munga_ has quit [Read error: 113 (No route to host)]
robocop has joined #ocaml
slash_ has quit [Client Quit]
verte has quit ["~~~ Crash in JIT!"]
drunK_ has joined #ocaml
drunK_ has quit [Remote closed the connection]
tmaedaZ is now known as tmaeda
_unK has quit [Read error: 145 (Connection timed out)]
<yziquel>
exit
yziquel has quit ["Leaving"]
Associat0r has quit [Read error: 104 (Connection reset by peer)]
<Yoric>
hi
seanmcl has joined #ocaml
ski__ has quit ["Lost terminal"]
BiDOrD has joined #ocaml
mbac_ has joined #ocaml
mbac has quit [Read error: 54 (Connection reset by peer)]
ikaros_ has joined #ocaml
iZZy_ has joined #ocaml
ikaros_ has quit [Read error: 104 (Connection reset by peer)]
robocop has quit ["Leaving."]
mbac_ has quit ["leaving"]
mbac has joined #ocaml
Submarine has quit [Remote closed the connection]
seanmcl has quit []
eldragon has joined #ocaml
eldragon has quit [Read error: 104 (Connection reset by peer)]
<thelema>
I have a list, and am iterating through all pairs of elements in the list.
<thelema>
My outer function splits the list at head and gives both parts to the inner function then calls itself on the tail.
<thelema>
My inner function is just a fold_left across the tail
<thelema>
with a function that takes [i] and [j] and does the work.
<thelema>
I must be triggering some major optimization, because when I try to rewrite the fold_left as a nested loop, I get *much* worse performance
<thelema>
hmm, closure creation?
<thelema>
why?
* thelema
makes an example divorced from his actual code...
<sgnb>
thelema: by expanding fold_left definition and inlining your auxiliary function, you might be creating more closures
<sgnb>
with fold_left, it is created only once
<sgnb>
but I am speculating... it's hard to say without having a look at the code
_unK has joined #ocaml
<thelema>
yes, I'm working on a simple test case...
* thelema
realizes it *has* to be a bug in his transformation
<thelema>
the program runs in 6 seconds with the fold, and after 20 minutes, he gives up with the transformed code... Thus something is wrong in the transformation
<thelema>
Okay, now it's "please help me find my bug"
<sgnb>
thelema: your code doesn't see to be self-contained
<thelema>
it's not, sorry.
Yoric has quit []
<thelema>
My program depends on batteries (specifically aaa-batteries) - I don't expect others to be able to compile/run it
* thelema
instruments more with ocamlviz to find the bug
* thelema
is probably messing up an accumulator somehow
<thelema>
The idea behind the code is that one is given a graph of dependencies, and I'm trying to find pairs of nodes such that the dependencies of one node are a subset of the dependencies of the other.
<thelema>
In this case, I can remove the subset node from the graph
<flux>
thelema, so is there a function that works and doesn't work? reduce_nodes1, reduce_nodes2?
<thelema>
reduce_nodes1 works, not reduce_nodes2
<flux>
but you'd like to use reduce_nodes2?
<flux>
it seems so much more complicated :)
<thelema>
I'd like to use code in the style of reduce_nodes2 because I want to stop the fold once I reduce I
<thelema>
I agree - it is much more complex.
<thelema>
Right now, I'm iterating through every pair (i,j) from the list
<thelema>
I want to skip all remaining (i,x) once I've removed i from the graph
<thelema>
oh crap. That's why...
<thelema>
hmmm...
<thelema>
I'm still using iset (the cached list of I's out-neighbors
<thelema>
even after I've removed I...
<sgnb>
thelema: so, that's specific to the library you use?
<thelema>
kind of
<thelema>
Since I'm doing a bunch of (i,j) tests for a single i, I pre-compute (get i) and use the same value over and over
<thelema>
(I'm planning on investigating caching (get i) for all i in the original list, and using an array to lookup instead of the binary tree it's stored in now.
<thelema>
ok, apparently the iset optimization, despite creating extra comparison work after i is removed, does save work overall.
<thelema>
but doesn't matter for my problem...
<thelema>
there's still some fundamental flaw in reduce_nodes2
<thelema>
and I just can't see it...
Alpounet has quit ["``Do what you think you can't do.'' -- E. Roosevelt"]
<thelema>
and I have to go now... I'll check the irc history when I get back.
Yoric has joined #ocaml
Alpounet has joined #ocaml
slash_ has joined #ocaml
Yoric has quit []
seanmcl has joined #ocaml
Yoric has joined #ocaml
spink has joined #ocaml
<spink>
hi, does ocaml have any type of compiler flag directives ? I wanna encapsulate the flag "-pp camlp5o" to my source file.
<Yoric>
In the source file, that's not possible.
<Yoric>
However, in an external file, it's possible.
<Yoric>
(if you're using ocamlbuild, the file is called _tags)
<spink>
I just submit the one file code to an online grading system :( and it's an one file submission
<spink>
so there is no way to even call the preprocessor from inside a file to do a #load ?
jcaose_ has quit [Read error: 110 (Connection timed out)]
<sgnb>
spink: no
<spink>
Thanks for the info guys, I'll send an email to the admins of the system :)
<Camarade_Tux>
spink: you should be able to preprocess it on your computer and send the file once preprocessed though
<Camarade_Tux>
(unless that makes code hard to read...)
<spink>
I use the Genlex to do some parsing that why I need camlp5. Camarade_Tux, what do you mean ?
<Camarade_Tux>
run the preprocessor on your computer and send the result
<Camarade_Tux>
add a link to the unpreprocessed file in comments (or embed the unpreprocessed code in comments)
<ygrek>
Genlex doesn't require camlp5, stream parsers syntax does (can be camlp4o btw)
yziquel has joined #ocaml
<yziquel>
Anyone having a 32bit Debian linux system and willing to pull in dependencies on R?
<spink>
ygrek, you are correct :)
Yoric has quit [Read error: 60 (Operation timed out)]
jcaose_ has joined #ocaml
jcaose_ has quit [Client Quit]
iZZy_ has quit [Client Quit]
slash_ has quit [Client Quit]
<sgnb>
yziquel: you don't?
w00t has joined #ocaml
w00t is now known as Guest24508
Guest24508 is now known as spink2
Yoric has joined #ocaml
<spink2>
Camarade_Tux was fine but is there a way to create source with camlp5o like with camlp4o ? camlp4o file.ml gives source where camlp5o file.ml gives intermediate.
<Camarade_Tux>
hmmm, I'm not very experienced with preprocessing but I would expect camlp5 not to be different (especially considering you can swap one for the other with minimal changes)
<Camarade_Tux>
or can't you use camlp4?
<Camarade_Tux>
yziquel: does it matter if it's not debian?
<yziquel>
sgnb: 64 bit. sorry for the delay.
<yziquel>
Camarade_Tux: there's either SVN, or a *convenient* Debian package.
<yziquel>
Camarade_Tux: with only SVN, there's perhaps some manual handling of files (i.e. META). That's all.
spink has quit [Read error: 110 (Connection timed out)]
<Camarade_Tux>
yziquel: you mean svn for R? and which version?
<yziquel>
Camarade_Tux: no. svn for OCaml-R. on gna.
<yziquel>
Camarade_Tux: you could probably use any R installation which provides a shared libR.so library.
M| has quit ["leaving"]
* Camarade_Tux
feels stupid, he forgot he was on a 64bit machine...
<Camarade_Tux>
yziquel: you need to test something?
<yziquel>
Camarade_Tux: grmpg.
<yziquel>
Cmarade_Tux: I need to test if OCaml-R works on 32bits, when it comes to wrapping R values. Not sure.
<Camarade_Tux>
you can maybe set a qemu virtual machine in a few minutes
M| has joined #ocaml
<sgnb>
yziquel: amd64 can run i386
<yziquel>
sgnb: linux32?
<sgnb>
yziquel: you can build an i386 (p|cow)builder chroot on amd64
<sgnb>
yziquel: with or without linux32
<yziquel>
ouch! OK.
<Camarade_Tux>
right, I had forgotten chroots
<sgnb>
(linux32 happens to be needed by some ./configure scripts)
<yziquel>
but if any has a 32bits Debian, that would help me a lot...
<sgnb>
one can also run an amd64 chroot on i386, btw, provided it has a suitable processor and kernel
<Camarade_Tux>
for the record, I have a new computer that I intend to make available for such things (and more) but it's not ready yet =/
ikaros has quit ["Leave the magic to Houdini"]
Yoric has quit []
Yoric has joined #ocaml
seanmcl has quit []
spink2 has left #ocaml []
<ygrek>
having R installed counts?
<yziquel>
ygrek: try a 'locate libR.so' and tell me...