<deech>
Hi all, is the ocaml-beginners Yahoo group still the best place to post questions?
<adrien>
it's a good one
<deech>
It's been a couple of years since I delved into Ocaml and I noticed that traffic on the list was less than before - are there other resources?
gnuvince_ has joined #ocaml
<mbac>
this channel's pretty good
<mbac>
i don't know why beginner questions would be off-topic
<mbac>
i think if you didn't write your thesis on type theory you're probably an ocaml beginner :)
deech has quit [Ping timeout: 240 seconds]
kerneis has quit [Ping timeout: 252 seconds]
diml has quit [Ping timeout: 252 seconds]
patronus has quit [Ping timeout: 255 seconds]
diml has joined #ocaml
Asmadeus has quit [Ping timeout: 265 seconds]
kerneis has joined #ocaml
patronus has joined #ocaml
Asmadeus has joined #ocaml
Edward_ has quit []
joewilliams is now known as joewilliams_away
philtor has joined #ocaml
Amorphous has quit [Ping timeout: 264 seconds]
Amorphous has joined #ocaml
Associat0r has joined #ocaml
philtor has quit [Ping timeout: 245 seconds]
Associat0r has quit [Quit: Associat0r]
joewilliams_away has quit [Excess Flood]
joewilliams has joined #ocaml
joewilliams is now known as joewilliams_away
joewilliams_away has quit [Excess Flood]
joewilliams has joined #ocaml
lamawithonel has quit [Remote host closed the connection]
lamawithonel has joined #ocaml
joewilliams is now known as joewilliams_away
joewilliams_away has quit [Excess Flood]
joewilliams has joined #ocaml
joewilliams is now known as joewilliams_away
joewilliams_away is now known as joewilliams
joewilliams is now known as joewilliams_away
joewilliams_away is now known as joewilliams
joewilliams is now known as joewilliams_away
philtor has joined #ocaml
joewilliams_away is now known as joewilliams
joewilliams is now known as joewilliams_away
joewilliams_away is now known as joewilliams
joewilliams is now known as joewilliams_away
joewilliams_away is now known as joewilliams
hcarty has quit [Ping timeout: 265 seconds]
philtor has quit [Ping timeout: 245 seconds]
joewilliams has quit [Excess Flood]
joewilliams has joined #ocaml
hcarty has joined #ocaml
seafood has quit [Quit: seafood]
seafood has joined #ocaml
joewilliams is now known as joewilliams_away
ygrek has joined #ocaml
init1 has joined #ocaml
hcarty has quit [Ping timeout: 260 seconds]
lpereira has joined #ocaml
seafood has quit [Quit: seafood]
Snark has joined #ocaml
ygrek has quit [Ping timeout: 245 seconds]
avsm has quit [Ping timeout: 252 seconds]
avsm has joined #ocaml
Yoric has joined #ocaml
mcclurmc has joined #ocaml
pikachuyann has joined #ocaml
ygrek has joined #ocaml
hcarty has joined #ocaml
ikaros has joined #ocaml
ulfdoz has joined #ocaml
ygrek_ has joined #ocaml
ygrek has quit [Ping timeout: 245 seconds]
mcclurmc has quit [Read error: Operation timed out]
asmanur has joined #ocaml
asmanur_ has quit [Read error: Operation timed out]
lpereira has quit [Quit: Leaving.]
mcclurmc has joined #ocaml
<ygrek_>
mfp, looks like your enhancement for caml_modify went unnoticed by inria, or are there any drawbacks? Are you going to push it to upstream?
oriba has joined #ocaml
lpereira has joined #ocaml
<mfp>
ygrek_: it's been a long while since I wrote that; AFAIK there's no drawback apart from the code duplication
<ygrek_>
probably you could submit it to bugtracker
<ygrek_>
great work btw
<mfp>
I'd have to measure the (code) size of the current and the modified Modify macros
<mfp>
the latter being bigger, it'll probably use more cache lines BUT they would rarely be evicted if caml_modify is called frequently enough (and if it isn't, we don't care)
<mfp>
so I think it'd be rather hard to come up with a scenario where it is slower than the original one
ftrvxmtrx has quit [Quit: Leaving]
<ygrek_>
it is also used in Array.set
<ygrek_>
other usages (obj and bytecode interpreter) are not interesting
<mfp>
yes, essentially in all the mutations where the compiler cannot prove the type is "immediate" (bool, char, int, type with only constant constructors)
<ygrek_>
actually for cache argument Modify would better be a function, to have one copy of code (guessing) :)
ftrvxmtrx has joined #ocaml
<mfp>
the OCaml code uses caml_modify, the Modify macro is only used in a couple places in the runtime IIRC
<mfp>
just caml_lazy_make_forward, a couple functions in Weak and a couple primitives (caml_array_set_addr, caml_array_unsafe_set_addr) which aren't really used (the native compiler emits the code directly)
<mfp>
ygrek_: ah, just remembered I've got to push your ocaml-cassandra patches... btw, do you want direct access to the github repos?
<mfp>
(sorry for the delay)
<ygrek_>
np
<ygrek_>
would not hurt, but I guess there will be not much patches from me, we've moved to hypertable here
<mfp>
why the move?
<ygrek_>
too resource hungry, too slow, too unpredictable.
<ygrek_>
it cannot do sequential reads at the speed of the disk - this is a show-stopper
<mfp>
mem usage and the unpredictability of GC storms also worried me
<mfp>
how mature is hypertable, though?
<ygrek_>
hypertable has problems too, like corrupting tables when disk is full, and startup is slow (opening 32K files at once, why the hell?), but better
<ygrek_>
yes, GCs for 15s are not funny
<ygrek_>
we moved one project from cassandra to hypertable and got visible LA drop and could handle more clients
<mfp>
the data model is the same as cassandra/bigtable, right?
<ygrek_>
same machine, same setup
<ygrek_>
essentially yes, it is more bigtable-style
<ygrek_>
i.e. no supercolumns, but CFs can be grouped for close physical storage
<ygrek_>
the whole arch is bigtable-alike, with main metadata server, etc
<mfp>
hmmm 2 ppl on #hypertable -> I hope it's the wrong channel :)
<flux>
butbut.. none of those are written in ocaml!
eldragon has joined #ocaml
<mfp>
Hypertable is written in C++, which should make it less bloaty and more predictable than Cassandra if the guys know what they're doing
oriba has quit [Ping timeout: 240 seconds]
<mfp>
Why do Java things have to get so enterprisy and bloated? e.g. when it tried to use ActiveMQ it used like 150 MB at startup and something ridiculous like 30 KB per subscription. I wrote my STOMP broker in OCaml and used uh 3 MB + 150 bytes per subs?
ftrvxmtrx has quit [Quit: Leaving]
ftrvxmtrx has joined #ocaml
oriba has joined #ocaml
<mfp>
ygrek_: are you the ygrek @ github I've found?
<mfp>
(adding you to the collaborator list)
eldragon has quit [Ping timeout: 245 seconds]
<ygrek_>
yes
<mfp>
OK, added
<mfp>
you should be able to push to git@github.com:mfp/ocaml-cassandra.git now
<ygrek_>
hypertable also crashes at shutdown, not a big deal, but makes you remember c++ :)
<mfp>
given the state of the competition, I guess it's OKish as long as it doesn't corrupt data :-/
mcclurmc has quit [Read error: Operation timed out]
mcclurmc has joined #ocaml
joewilliams_away is now known as joewilliams
mcclurmc has quit [Ping timeout: 250 seconds]
Yoric has quit [Quit: Yoric]
ftrvxmtrx has quit [Quit: Leaving]
oriba has quit [Quit: Verlassend]
mcclurmc has joined #ocaml
julm has quit [Quit: leaving]
ztfw has joined #ocaml
ftrvxmtrx has joined #ocaml
julm has joined #ocaml
Smerdyakov has joined #ocaml
avsm has quit [Quit: Leaving.]
eldragon has joined #ocaml
ulfdoz has quit [Remote host closed the connection]
ulfdoz has joined #ocaml
ftrvxmtrx has quit [Read error: Connection reset by peer]
ftrvxmtrx has joined #ocaml
dug_ has joined #ocaml
ccasin has joined #ocaml
Fullma has quit [Ping timeout: 240 seconds]
Fullma has joined #ocaml
rks has joined #ocaml
Anarchos has joined #ocaml
boscop has quit [Ping timeout: 245 seconds]
boscop_ has joined #ocaml
philtor has joined #ocaml
Anarchos has quit [Quit: Vision[0.8.5-0418]: i've been blurred!]
matthiasgoergens has joined #ocaml
<matthiasgoergens>
hi. is there a standard library function to fold (e.g. fold_right or fold_left) over a string?