<thelema>
fasta_: at the moment, odb can't get version numbers from the server, it can only get version numbers from installed packages. I'll have to bother gildor to do this unless someone with ocsigen experience wants to figure out his oasis-db server code.
<thelema>
fasta_: --info may or may not contact a server; it depends on whether the metadata for that package is local or remote.
<thelema>
fasta_: ah, you think that the description for --info should be more verbose? I can see a note in the returned info saying where the metadata came from, but I don't see much benefit in making the help verbose
<thelema>
_habnabit: for that bucket histogram, is it under 4.00 or 3.x? The hash function details changed between the two; I think for the better on 4.00, but if it's still like this on 4.00... :(
<_habnabit>
thelema, 3.12.1
avsm has quit [Quit: Leaving.]
avsm has joined #ocaml
<thelema>
_habnabit: ok, you may see better hashtbl performance (and bucket distribution) under 4.00
<thelema>
Maybe sometime I'll run some tests to see for myself.
<thelema>
In theory, a simple hash function is good enough if its input has enough entropy / unpredictableness, but the way that ocaml values are hashed, I don't see too much entropy going in
<_habnabit>
when I was looking at the values, it seemed that it might be not probing deeply enough into the structure to get an adequate hash value out
<_habnabit>
I didn't check into that too much, though
<thelema>
yup, ocaml only goes part of the way in. Also, the old hash function had terrible performance on deep structures, because it could get stuck with a hash value of 0
<thelema>
meaning that sufficiently deep structures would all hash to the same bucket
<_habnabit>
yeah
<_habnabit>
this wasn't the 0 bucket, though
<thelema>
for i = 0 to 40 do print_int (Hashtbl.hash(String.make i '0')); print_newline(); done;;
<thelema>
try that in your toplevel
<thelema>
I can't verify it right now (err, I haven't, but I can, I realize), but iirc, it should print 0's after some i
<_habnabit>
looks fine here, actually
<thelema>
hmm, maybe it wasn't strings... one sec while I try to reproduce... maybe it was lists
avsm has quit [Quit: Leaving.]
<thelema>
for i = 0 to 15 do Printf.printf "%d %d\n" i (Hashtbl.hash (Array.to_list (Array.make i 'a'))); done;;
<thelema>
yes, that's right, ocaml has a different hash function just for strings, and this one isn't too bad.
sepp2k has quit [Read error: Connection reset by peer]
jave has quit [Read error: Connection reset by peer]
jave has joined #ocaml
<thelema>
but the code I just pasted goes to 0 and stays at 0 for i>=9
<thelema>
(by just, I mean before the huge blip in my internet connectino)
jamii has quit [Ping timeout: 246 seconds]
gnuvince has joined #ocaml
bzzbzz has joined #ocaml
emmanuelux has quit [Ping timeout: 248 seconds]
abeaulieu has quit [Changing host]
abeaulieu has joined #ocaml
paolooo has joined #ocaml
avsm has joined #ocaml
Yoric has joined #ocaml
ulfdoz has quit [Ping timeout: 272 seconds]
<cacho>
any String.repeat?
testcocoon has quit [Quit: Coyote finally caught me]
BiDOrD_ has joined #ocaml
BiDOrD has quit [Ping timeout: 252 seconds]
eikke has joined #ocaml
testcocoon has joined #ocaml
eikke has quit [Ping timeout: 240 seconds]
<adrien>
I'm looking for an application using lablgtk2 in threaded mode
<adrien>
the mode unfortunately caused CPU usage even when idle and not many programs use it but I need one or two not trivial programs to check they still work with the latest changes (thanks to diml)
avsm has quit [Quit: Leaving.]
Yoric has quit [Ping timeout: 244 seconds]
madroach has quit [Quit: leaving]
madroach has joined #ocaml
jamii has joined #ocaml
osa1 has joined #ocaml
jave has quit [Read error: Connection reset by peer]
jave_ has joined #ocaml
osa1 has quit [Read error: Operation timed out]
cdidd has quit [Ping timeout: 252 seconds]
Tobu has quit [Read error: Connection reset by peer]
Tobu has joined #ocaml
lopex has quit [Remote host closed the connection]
joewilliams has quit [Remote host closed the connection]
bobry has quit [Remote host closed the connection]
emmanuelux has joined #ocaml
<vbmithr>
Hi
<vbmithr>
I can’t compile ocaml 4.00 for armhf
<vbmithr>
make opt.opt fails with error 137
<vbmithr>
the rest works fine
Yoric has joined #ocaml
<vbmithr>
I think it’s a matter of the job used too much cpu time
Yoric has quit [Ping timeout: 244 seconds]
<flux>
too much cpu time?
<vbmithr>
I’ve seen that somewhere on the net, but it seems strange
<vbmithr>
I didn’t know that a user process has a limit in terms of CPU time it can use
<flux>
such a limit can be set, but I imagine it's very rare to do so.
<vbmithr>
I’m using debian on a efika MX machine (cortex A8 CPU, armhf arch)
<vbmithr>
ys
<flux>
and the limit would need to be quite low for ocaml not to compile..