<\u>
seems to work: ocamlc -custom llvm.cma -cclib -lLLVMCore
\u has quit [Quit: WeeChat 1.0.1]
darkf_ has joined #ocaml
darkf has quit [Ping timeout: 245 seconds]
samuel02 has joined #ocaml
darkf_ is now known as darkf
samuel02 has quit [Ping timeout: 245 seconds]
thomasga has joined #ocaml
zapu has joined #ocaml
emery has quit [Ping timeout: 250 seconds]
emery has joined #ocaml
MercurialAlchemi has joined #ocaml
WraithM has quit [Quit: leaving]
larhat has joined #ocaml
Simn has joined #ocaml
Submarine has joined #ocaml
Submarine has joined #ocaml
ygrek has joined #ocaml
arj has joined #ocaml
cdidd has quit [Ping timeout: 240 seconds]
cdidd has joined #ocaml
AlexRussia has quit [Ping timeout: 245 seconds]
keen___________5 has joined #ocaml
keen___________4 has quit [Ping timeout: 265 seconds]
ygrek has quit [Ping timeout: 255 seconds]
samuel02 has joined #ocaml
samuel02 has quit [Read error: Connection reset by peer]
samuel02 has joined #ocaml
pierpa` has quit [Ping timeout: 250 seconds]
pierpa has quit [Ping timeout: 250 seconds]
samuel02 has quit [Ping timeout: 245 seconds]
samrat has quit [Quit: Computer has gone to sleep.]
ygrek has joined #ocaml
samrat has joined #ocaml
thomasga has quit [Quit: Leaving.]
thomasga has joined #ocaml
samuel02 has joined #ocaml
waneck has quit [Ping timeout: 240 seconds]
waneck has joined #ocaml
cesar_ has joined #ocaml
cesar_ is now known as Guest37300
iZsh has quit [Ping timeout: 240 seconds]
Guest37300 has quit [Ping timeout: 255 seconds]
xorpse has quit [Ping timeout: 244 seconds]
xorpse has joined #ocaml
iZsh has joined #ocaml
ygrek has quit [Ping timeout: 264 seconds]
zpe has joined #ocaml
ebzzry has quit [Ping timeout: 272 seconds]
ebzzry has joined #ocaml
ikaros has joined #ocaml
samuel02 has quit [Remote host closed the connection]
govg has quit [Quit: Lost terminal]
govg has joined #ocaml
avsm has joined #ocaml
George___ has joined #ocaml
jonludlam has quit [Quit: Coyote finally caught me]
avsm has quit [Quit: Leaving.]
pii4 has quit [Ping timeout: 265 seconds]
chambart has joined #ocaml
nojb has joined #ocaml
cesar_ has joined #ocaml
cesar_ is now known as Guest83050
siddharthv has quit [Ping timeout: 258 seconds]
johnelse has quit [Ping timeout: 258 seconds]
Guest83050 has quit [Ping timeout: 264 seconds]
johnelse has joined #ocaml
johnelse is now known as Guest42247
avsm has joined #ocaml
siddharthv has joined #ocaml
AlexRussia has joined #ocaml
siddharthv is now known as siddharthv_away
ygrek has joined #ocaml
<ggole>
"What reason stands for the use of camelCase instead of PascalCase convention for static methods?" "His reason is probably to make it feel more like functional programming"
nojb has quit [Quit: nojb]
* ggole
head asplode
<flux>
troll level: expert :)
avsm has quit [Quit: Leaving.]
pii4 has joined #ocaml
nojb has joined #ocaml
samrat has quit [Ping timeout: 264 seconds]
dsheets_ has joined #ocaml
samrat has joined #ocaml
ousado has quit [Ping timeout: 240 seconds]
_5kg has quit [Ping timeout: 264 seconds]
ousado has joined #ocaml
_5kg has joined #ocaml
<reynir>
what's the context?
companion_cube has quit [Read error: Connection reset by peer]
<adrien>
what these stats show are mostly the amount of trolling
<adrien>
and they only give a partial picture
<Leonidas>
I wonder why there is not more traffic. maybe mailing lists are dying.
<MercurialAlchemi>
reddit effect?
<Leonidas>
because I'd say that ocaml has been picking up momentum lately.
companion_cube has joined #ocaml
libertas has joined #ocaml
<George___>
Guys, does the performance of Hastbl.find also depend on the data type of the key?
<flux>
george___, well, hashing may depend on the data type of the key. in addition, when there are multiple keys in a bucket, they need to be compared to each other, so that may be a factor as well.
jonludlam has joined #ocaml
<George___>
thanks, I have keys that are of type list or arbitrary size. I am wondering if lookups will be faster if I replace list with record, that might have option types when an entry doesn't exist
<George___>
The question can also asked this way: does the speed of hashing the key depends on the key's data type?
Guest42247 is now known as johnelse
<mrvn>
depends on how much sharing is in the list and how often you hash the same key and how cache local that then is.
<mrvn>
and what's in the list?
<mrvn>
Pverall the only thing to do is benchmark.
<mrvn>
s/P/O/
siddharthv_away is now known as siddharthv
chambart has quit [Ping timeout: 256 seconds]
sepp2k has joined #ocaml
companion_cube has quit [Ping timeout: 244 seconds]
<George___>
list contain variant types, and hashing the same key may happen millions of tmes a second
<George___>
any idea now?
companion_cube has joined #ocaml
<ousado>
George___: does that include possibly hashing the same object several times?
<flux>
george___, you should run Hashtbl.hash on those objects to see how many collisions you have
<flux>
though if you have lists as objects, maybe it a Map would serve you better?
<flux>
they need to dig down the lists only as far as to find a difference
<flux>
I suppose that's true of the Hashtbl equality comparison as well, and the depth the hash function travels is fixed
<ousado>
wouldn't that mean that lists with the same prefix for that length/depth end up with the same hash value?
<George___>
no, the same object is not stored in the table several times.
<George___>
But I might look for the same object many times in a sec in the table
dsheets_ has quit [Ping timeout: 264 seconds]
<ousado>
then how about precalculating the/a hash, { key=(hash list); list=list }
<ousado>
and using a specific has for the hashtable that uses only that key
<ousado>
*hash
<George___>
is this hash the same as the hash function used by hashtbl?
<ousado>
you can define your own
<ousado>
you can also use Hashtbl.hash
<George___>
brilliant! So if Hashtbl.hash is the same function that is used by Hashtbl, then I can do some performance test using it
cesar_ has joined #ocaml
lu324 has quit [Ping timeout: 240 seconds]
cesar_ is now known as Guest53575
<ousado>
George___: yes (my main point is that you can use the functorial interface of Hashtbl to provide a function that uses the 'key' field of the record instead of the whole thing, just in case I didn't make that clear)
<George___>
Thanks
Guest53575 has quit [Ping timeout: 265 seconds]
<ggole>
George___: Hashtbl.find invokes a polymorphic equality that can be pretty slow
<ggole>
You might benefit from replacing that with a hand-written or ppx-derived one (even if you keep the same hash)
<George___>
Interesting,!
<ggole>
But before doing any of that I would get a profiler up and running to measure any changes.
<nicoo>
ousado: If you can hashcons your objects, then you can give them very fast comparison and hash operators
<nicoo>
Which would help
<ousado>
that was for George___ I guess
<George___>
what does hashcons do?
ygrek has quit [Ping timeout: 244 seconds]
badon has quit [Read error: Connection timed out]
<nicoo>
George___: Basically, you ensure that there is a single instance of each object in memory. Imagine I'm handling some tree datatype (type 'a tree = Leaf of 'a | Node of 'a tree * 'a tree). Then, instead of constructing a new tree Node(a,b) directly, I would check if I already constructed one. If so, I return than one
badon has joined #ocaml
<nicoo>
It makes it very easy to have fase compare and hash, and can reduce memory usage. On the other hand, it makes constructing a new object a bit slower
<George___>
great, fast lookup is more important for me. I will have a look at that
thomasga has quit [Quit: Leaving.]
lu324 has joined #ocaml
<ousado>
nicoo: so hashconsing also captures the structure of objects?
<nicoo>
ousado: I'm not sure what you mean, but if you have two values with different structure, then they won't be equal an would have different ids if you employ hashconsing
<ggole>
No, it just makes equal constructors have the same identity
<ggole>
If you need to determine some fact about the structure other than equality, you need to look at the parts just as usual.
<ousado>
hmm, my use case is not applicable then, I guess
<ggole>
It also doesn't work well in the presence of mutation.
dsheets_ has joined #ocaml
<ousado>
I'm looking for a way to establish that a serialized blob of data will have a specific structure when deserialized before actually doing it
<ousado>
something like a "structural hash"
samuel02 has joined #ocaml
<ousado>
but that's very hard to do, I think
<ggole>
Hmm... you could compare it to a precomputed digest of a seralized version of your structure
<ggole>
I mean, cerealized.
nojb has quit [Quit: nojb]
siddharthv is now known as siddharthv_away
samuel02 has quit [Client Quit]
<ousado>
heh
lordkryss has joined #ocaml
<Leonidas>
the gmane archive looks strange
<Leonidas>
seems like the messages all expire
<Leonidas>
starting with may 2014
<ousado>
ggole: what I've thought about is serializing in a way so that all size information (like arrays/lists/strings etc.) is put at the beginning of the serialized blob, and checking whether that matches the type and is coherent with the size of the whole blob
<ousado>
(assuming a simple binary, length-prefixed format)
<ggole>
ousado: that won't tell you if the contents are the same, though
<ggole>
(Not sure if that's a dealbreaker or not, though.)
<ousado>
it's not
<ousado>
I don't know the contents beforehand
samrat has quit [Quit: Computer has gone to sleep.]
<ggole>
Unless I misunderstand (very possible), this doesn't seem much like a structural hash
<ousado>
I'm not aware of a/the correct term for this, a 'function that can be used to verify structural properties of the serialized blob' is about what I mean
<ggole>
So you want to hash on the size/arrangement, but not the contents?
<ousado>
yes, it's for deciding whether to even try to unserialize a message of untrusted origin
<ousado>
*deserialize
<ggole>
Hmm, I see.
nojb has joined #ocaml
<ggole>
I guess you aggregate the information that you are interested in and run a digest over that.
<ggole>
(And check for equality if it matches.)
q66[lap] has quit [Read error: Connection reset by peer]
q66[lap] has joined #ocaml
nojb has quit [Read error: Connection reset by peer]
<ousado>
someone here mentioned his implementation of 'authenticated datastructures' (ADS) in ocaml sometime last year I think
nojb has joined #ocaml
<ousado>
that's going much further than what I want, but seems to be a similar direction
<ousado>
anyway, thanks for your feedback, I guess I'll have to go for a more naive solution in the meantime, there's more pressing stuff to do right now :)
<arj>
can I reload a module within utop somehow?
<ggole>
Sort of. You can load it again, but references to the old module will still point there.
<arj>
can I invalidate the old references somehow?
<ggole>
Not without reloading the whole thing.
<arj>
load again results in "disagree over interface"
<ggole>
#mod_use "foo.ml", maybe
<ggole>
But again, references will not be up to date.
<arj>
thanks
<ggole>
My solution to this was to write some automation to make it easy to restart the toplevel with up-to-date source.
<ggole>
It still sucks a bit.
<arj>
ggole: can I script utop somehow to automatically load some modules?
<ggole>
You can, with .ocamlinit
<ggole>
But it's usually easier to use a build system
<Leonidas>
I just used a file that I included in the toplevel
<ggole>
ocamlbuild has somewhat half-hearted support for building toplevels.
<ggole>
For simple cases it's fine.
<ggole>
I usually enhance this by putting something like #use code.ml in a local .ocamlinit, and typing or copying code I want to test into that file.
<ggole>
When you build and start the toplevel, the code runs and you see the effect of your change without having to type anything.
<arj>
ah very nice :)
<arj>
almost feels like cabal repl ;-) Ocaml definitely needs this!
<arj>
oasis repl
nojb has quit [Quit: nojb]
<ggole>
Yeah, it's a bit too clumsy to set up as things stand.
<ggole>
Maybe I should suggest some changes to the tuareg/ocamlbuild people to make it easier.
ysz has joined #ocaml
BitPuffin has joined #ocaml
<arj>
is there some curry uncurry function in Core?
ysz has quit [Client Quit]
<Leonidas>
i'd expect it in Core.Fn, but it doesn't seem to be there.
<ggole>
Easy enough to write.
<Leonidas>
yep, uncurry = BatPervasives.uncurry
<Leonidas>
;)
<arj>
Leonidas: ;-)
<arj>
ggole: every time? in every new project? no...
<ggole>
Embrace the nih!
<ggole>
I write Option.map/iter/get for each project, too.
<arj>
but this should be in the standard library, then!
<ggole>
Lots of things should be in the standard library that are not -_-
<ggole>
But that's what Batteries and the like are for.
<Leonidas>
arj: you could propose a Core(_kernel).Fn.uncurry
samrat has joined #ocaml
<companion_cube>
or use CCFun.uncurry :>
<companion_cube>
(from containers)
<MercurialAlchemi>
companion_cube: what can you store in 'uncurry'?
<MercurialAlchemi>
not curry, that's for sure...
<arj>
Leonidas: yes, I should probably do it.
<Leonidas>
like lisp seems to be designed as a tool to implemt lisps, ocaml is a tool to implement stdlibs.
<arj>
loooooooooool
<MercurialAlchemi>
and make puns
hugomg has joined #ocaml
cesar_ has joined #ocaml
cesar_ is now known as Guest45448
Nahra has joined #ocaml
Guest45448 has quit [Ping timeout: 255 seconds]
ygrek has joined #ocaml
nox_ is now known as nox
rand000 has quit [Quit: Lost terminal]
rand000 has joined #ocaml
ygrek has quit [Ping timeout: 258 seconds]
thomasga has joined #ocaml
sheijk has joined #ocaml
<Unhammer>
If I do "Eliom_pervasives.server_function ~timeout:0.01 Json.t<string> some_slow_function", I would expect the result to be None, but I get "Failure(unwrap_value…)" in the console instead; what's the correct way to catch timeouts?
nojb has joined #ocaml
samuel02 has joined #ocaml
chambart has joined #ocaml
jgjl has quit [Ping timeout: 264 seconds]
jonludlam has quit [Ping timeout: 244 seconds]
samrat has quit [Ping timeout: 240 seconds]
badkins has joined #ocaml
sepp2k has quit [Quit: Konversation terminated!]
zpe has quit [Remote host closed the connection]
samrat has joined #ocaml
jonludlam has joined #ocaml
<arj>
anyone knows how to execute core_kernel's tests?
<arj>
make test doesn't work.
avsm has joined #ocaml
shinnya has joined #ocaml
enitiz has joined #ocaml
bezirg has quit [Ping timeout: 240 seconds]
rand000 has quit [Quit: leaving]
<Leonidas>
arj: which tests?
badon has quit [Disconnected by services]
badon_ has joined #ocaml
badon_ is now known as badon
<arj>
Leonidas: there are lines like TEST inside the source files of core_kernel.
<Unhammer>
how can I return "Some div" from an Eliom_registration.*.register_post_coservice' ?
<Leonidas>
arj: ah. I suppose you get the same error about Core not being found?
jgjl has joined #ocaml
<arj>
Leonidas: no, I tried doing make test. Then it says I should run configure --enable-tests. When I do this and run make test afterwards, nothing happens. It just compiles and stops.
darkf has quit [Quit: Leaving]
ysz has joined #ocaml
<Drup>
Unhammer: I ... have no idea.
<Drup>
do a bug report about the timeout failure.
<Unhammer>
hm, I know I can use Eliom_registration.Ocaml.register_post_coservice' to return "Some string", but returning "Some div" just gives me "expression contains type variables that cannot be generalized"
<nojb>
arj: you probably need to specify which tests to run - look at the documentation of pa_ounit
<Drup>
Unhammer: oh, that's a different issue
<arj>
nojb: thanks for the pointer. But there should be some means of running all of them?
<Drup>
Unhammer: give an explicit type somewhere about it, and it should be ok