#ocaml on 2015-05-19 — irc logs at freenode.irclog.whitequark.org

2014-11-22 23:23 ChanServ changed the topic of #ocaml to: Discussions about the OCaml programming language | http://www.ocaml.org | OCaml 4.02.1 announcement at http://ocaml.org/releases/4.02.html | Public channel logs at http://irclog.whitequark.org/ocaml

00:03 ollehar has quit [Read error: Connection reset by peer]

00:06 ollehar has joined #ocaml

00:11 <jonh> hmm

00:12 <jonh> q66: interesting they are calling it a forum, but using irc logs?

00:12 rgrinberg has joined #ocaml

00:15 <dmbaturin> Those logs are misrepresented, I'd say.

00:23 uris77 has joined #ocaml

00:24 <apache2> I like this post you made on the forum, dmbaturin, the one titled: "Well, given how many people write applications in C++, that's still valid. :)"

00:27 <dmbaturin> Precisely the kind of stuff I post on forums I don't know they exist. :)

00:27 <companion_cube> adrien: on a new wheel

00:34 <apache2> hehe :D

00:35 hilquias has quit [Remote host closed the connection]

00:35 uris77 has quit [Ping timeout: 276 seconds]

00:35 <apache2> haha, this format is quite enjoyable to read. like this brilliant forum post by n3ss3s: "oh sweet jesus now utop is not found"

00:37 ir2ivps4 has quit [K-Lined]

00:40 <dmbaturin> Oh, so "member profiles" allow one to see what that person posted on all channels they log.

00:41 <apache2> I wonder what the purpose is. Siderank is quite probable. But then what? ads? malware?

00:42 <dmbaturin> Why not all of these?

00:43 <dmbaturin> They aren't mutually exclusive.

00:43 <apache2> point.

00:48 thomasga has quit [Quit: Leaving.]

00:51 thomasga has joined #ocaml

00:56 psy has quit [Disconnected by services]

00:56 thomasga has quit [Quit: Leaving.]

00:57 psy_ has joined #ocaml

01:04 shinnya has quit [Ping timeout: 272 seconds]

01:10 palomer has joined #ocaml

01:14 The_Mad_Pirate has joined #ocaml

01:17 struktured has joined #ocaml

01:22 jonh has left #ocaml ["WeeChat 0.4.2"]

01:23 Algebr has joined #ocaml

01:48 ygrek has quit [Ping timeout: 264 seconds]

01:53 idegen has quit [Quit: Leaving.]

01:54 kdef has quit [Quit: Leaving]

01:55 manizzle has quit [Ping timeout: 276 seconds]

01:58 manizzle has joined #ocaml

02:07 hilquias has joined #ocaml

02:13 badon has quit [Read error: Connection reset by peer]

02:14 badon has joined #ocaml

02:15 darkf has joined #ocaml

02:18 jabesed has quit [Ping timeout: 258 seconds]

02:28 ilia has joined #ocaml

02:29 idegen has joined #ocaml

02:31 keen_____ has quit [Read error: Connection reset by peer]

02:31 b0mbastus has quit [Ping timeout: 265 seconds]

02:33 keen_____ has joined #ocaml

03:06 <ilia> I have a question regarding debugging: how to trace a function, e.g. to print something in a log?

03:08 <dmbaturin> ilia: In a daemon or other type of non-interactive program?

03:10 <dmbaturin> If stdout/stderr are not closed, it's trivial. let foo x = Printf.fprintf stderr "doing things with %d\n" x; x + 1

03:16 idegen has quit [Quit: Leaving.]

03:18 c74d has quit [Remote host closed the connection]

03:21 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

03:24 c74d has joined #ocaml

03:24 palomer has quit [Quit: palomer]

03:25 manizzle has quit [Ping timeout: 252 seconds]

03:28 b0mbastus has joined #ocaml

03:29 kido1412 has joined #ocaml

03:30 Algebr has quit [Remote host closed the connection]

03:31 <kido1412> Does codes like ```type xxx with sexp``` is a standard OCaml syntax?

03:34 sdothum has quit [Quit: ZNC - 1.6.0 - http://znc.in]

03:37 Denommus has joined #ocaml

03:37 Denommus has quit [Max SendQ exceeded]

03:38 Denommus has joined #ocaml

03:38 Denommus has quit [Max SendQ exceeded]

03:39 Denommus has joined #ocaml

03:44 hay207 has joined #ocaml

03:45 <Drup> kido1412: no, it's a syntax extension

03:46 <dmbaturin> Drup: What does the default Lwt logger do when stderr is closed? Does it automatically switch to syslog, or it does something else?

03:48 <Drup> nothing special, I think

03:49 <Drup> dmbaturin: the code is quite trivial, you can just read it really ;)

03:51 <dmbaturin> I automatically assumed that Lwt code is full of heavy wizardry and didn't even clone it. Maybe I should.

03:51 <Drup> it is actually not.

03:53 hay207 has quit [Ping timeout: 264 seconds]

04:07 milosn has quit [Read error: Connection reset by peer]

04:07 milosn_ has joined #ocaml

04:08 hay207 has joined #ocaml

04:17 swgillespie has joined #ocaml

04:19 swgillespie has quit [Client Quit]

04:19 psy_ has quit [Ping timeout: 276 seconds]

04:21 swgillespie has joined #ocaml

04:37 Submarine_ has joined #ocaml

04:38 <flux> companion_cube, oh, too bad

04:39 <flux> companion_cube, I guess if it'll annoy me too much I'll eventually do a rewrite then :)

04:39 Submarine_ has quit [Client Quit]

04:46 ygrek has joined #ocaml

04:50 tnguyen has joined #ocaml

04:51 tnguyen has quit [Client Quit]

04:57 tnguyen has joined #ocaml

05:02 tnguyen has quit [Quit: tnguyen]

05:04 milosn_ is now known as milosn

05:04 adc90 has quit [Read error: Connection reset by peer]

05:04 adc90 has joined #ocaml

05:13 MercurialAlchemi has joined #ocaml

05:19 <flux> you shouldn't really close stderr or stdout, should you? if you do that, you risk reopening a new file handle and it being assigned 1 or 2. instead, you should open /dev/null to their fds.

05:20 <flux> and in that case any log writer doesn't need to care.

05:20 tnguyen has joined #ocaml

05:22 <dmbaturin> flux: True. I was just searching for the least headache way to log my stuff.

05:23 <dmbaturin> I guess I'll actually have to resort to reading the source.

05:23 hay207 has quit [Remote host closed the connection]

05:27 hay207 has joined #ocaml

05:33 ilia has quit [Quit: Leaving]

05:42 hay207 has quit [Ping timeout: 272 seconds]

05:45 manizzle has joined #ocaml

05:47 <apache2> flux: correct me if I'm wrong, but if the log writer has an open handle, and you close the fd, wouldn't that yield a broken pipe on the log writer end too?

05:50 <flux> apache2, yes, but I mean the case if the writing process closes the fd it's writing to

05:50 <flux> so the the fd (say stdout = fd 1) is released for other use

05:51 <flux> and if you open any file, it will get assigned fd 1

05:51 matason has joined #ocaml

05:51 <flux> and if you then, say, start an external application that writes to stdout, it will get into that file

05:55 projective-pyon is now known as degenerate-pyon

05:59 <dmbaturin> apache2: Oh, so I'm not the only one who's having a log problem today?

06:02 <apache2> flux: oh, so children of your process. right.

06:03 <apache2> dmbaturin: I don't keep logs, I don't care :)

06:06 matason has quit []

06:13 Haudegen has quit [Ping timeout: 276 seconds]

06:16 tmtwd has quit [Quit: Leaving]

06:19 <flux> apache2, or even just some 'stdout' object that references to the said fd.. but I guess you can be mostly certain you've closed them all within your process.

06:21 <apache2> yeah, you're right

06:24 ingsoc has joined #ocaml

06:25 <flux> at least ocaml sets fd to '-1' when closing an out_channel

06:25 <flux> but it does happily try to write there.. and then gets an error from the kernel interface.

06:26 <apache2> :-/

06:26 <apache2> wait, why would it try to write to a closed descriptor?

06:26 <apache2> runtime checks for "is this fd == -1" too slow?

06:27 kido1412 has quit [Quit: Page closed]

06:27 <apache2> that probably makes sense - most of the time it would be open, then the check would be overhead

06:31 <adrien> companion_cube: congrats!

06:32 <flux> apache2, sure, why not, the kernel is going to check it anyway

06:35 MercurialAlchemi has quit [Remote host closed the connection]

06:35 djellemah has joined #ocaml

06:35 MercurialAlchemi has joined #ocaml

06:37 Haudegen has joined #ocaml

06:37 MercurialAlchemi has quit [Remote host closed the connection]

06:37 MercurialAlchemi has joined #ocaml

06:48 SIGILL has quit [Read error: Connection reset by peer]

06:48 SIGILL has joined #ocaml

06:53 Bhavya has joined #ocaml

06:55 muzzle has joined #ocaml

06:59 oscar_toro has quit [Ping timeout: 256 seconds]

07:07 matason has joined #ocaml

07:09 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

07:09 adc90 has quit [Read error: Connection reset by peer]

07:09 adc90 has joined #ocaml

07:11 jonludlam has quit [Ping timeout: 252 seconds]

07:17 Gama11 has joined #ocaml

07:18 oscar_toro has joined #ocaml

07:19 thomasga has joined #ocaml

07:38 <struktured> anyone aware of a pure ocaml curl lib?

07:40 tmtwd has joined #ocaml

07:40 <MercurialAlchemi> struktured: a web client?

07:41 <struktured> MercurialAlchemi: I just want to download a http or fttp resource in pure ocaml, using sys.command + wget atm, trying to find something better

07:42 <MercurialAlchemi> struktured: cohttp has a client part

07:43 muzzle has quit [Quit: Leaving.]

07:45 <MercurialAlchemi> struktured: don't think it does ftp though...

07:46 <struktured> MercurialAlchemi: I don't need ftp, but not sure if this is best choice either, need to understand api's scope better. was playing with curl bindings earlier but really want something pure if it's out there

07:49 <MercurialAlchemi> struktured: if all you need is to get a response, you should be covered

07:49 <MercurialAlchemi> it can use lwt or async

07:49 martintrojer has quit [Max SendQ exceeded]

07:49 martintrojer has joined #ocaml

07:50 <struktured> yeah I saw that, is there a client only package, as you hinted at earlier too? rather not pull in the server libs

07:51 muzzle has joined #ocaml

07:51 <MercurialAlchemi> don't think so

07:51 <MercurialAlchemi> opam doesn't think so either

07:52 <struktured> well it claims to be light weight, and I have to give them kudos for supporting both lwt and async, something i'm trying to do myself lately

07:52 <reynir> hi struktured! Any luck with opam on solaris? :)

07:52 <MercurialAlchemi> it's a lot lighter than ocamlnet, that's for sure

07:53 mort___ has joined #ocaml

07:53 <struktured> reynir: I pushed back the urgency of it, as I don't need to immediately run it on anything solaris. I will revisit it for like 30 minutes tommorow and review with you the problem's I'm running into. Also, I could sometimes be blocked by my ticketing system of unfortunately..

07:54 <struktured> reynir: last I remember, the configur was barfing on a some malformed "test" command. I might overlay gnu tools into ~/local and adjust the path. screw solaris build tools

07:54 <reynir> heh

07:55 <reynir> I was just curious :)

07:55 <struktured> I tihnk key to victory is rocking gnu as far as you are willing

07:55 <struktured> still can be solaris target binaries of course

07:57 <reynir> Yea

08:01 gperetin has quit [Ping timeout: 272 seconds]

08:01 adr has joined #ocaml

08:01 adr has quit [Remote host closed the connection]

08:04 manizzle has quit [Ping timeout: 256 seconds]

08:10 dav- has quit [Read error: Connection reset by peer]

08:11 hilquias has quit [Remote host closed the connection]

08:14 rgrinberg has quit [Ping timeout: 252 seconds]

08:14 badon has quit [Disconnected by services]

08:14 badon_ has joined #ocaml

08:15 badon_ is now known as badon

08:15 muzzle has quit [Quit: Leaving.]

08:16 axiles has quit [Quit: Quitte]

08:16 dav has joined #ocaml

08:17 <companion_cube> flux: maybe the right idea is to use a matrix of codepoints

08:17 <companion_cube> rather than string typs

08:18 <companion_cube> types

08:18 <companion_cube> (if you look at the code)

08:19 badon has quit [Disconnected by services]

08:21 jwatzman|work has joined #ocaml

08:21 maurer has quit [Ping timeout: 240 seconds]

08:25 <Enjolras> hi

08:25 <Enjolras> I made a few tries and i am starting to understand what you mean about subtyping and GADT not working

08:26 <Enjolras> for instance, an existential type like this type test = Test [< `A | `B | `C ] t -> test doesn't work as expected

08:26 <Enjolras> if i try to unwrap test, and match with something of type [`A] t, it fails

08:27 badon_ has joined #ocaml

08:28 badon_ is now known as badon

08:29 <Enjolras> still the variance issue i guess

08:30 avsm has joined #ocaml

08:30 badon has quit [Remote host closed the connection]

08:31 yomimono has joined #ocaml

08:31 <flux> companion_cube, aren't there combined characters as well?

08:32 <flux> leading to just using a matrix of strings

08:32 <flux> also, works with ansi codes as well

08:33 <companion_cube> well, if you want to go general I guess you need a sum type then

08:33 <flux> functor!

08:33 <companion_cube> type point = UChar of int | Ansi of ansi_code * point | ...

08:33 <companion_cube> erf

08:34 <flux> what's the point, in the end?

08:34 dsheets has quit [Ping timeout: 256 seconds]

08:34 <flux> you're not going to do much processing to that data regardless its contents

08:34 <flux> it just needs to fits :)

08:34 thomasga has quit [Quit: Leaving.]

08:36 <companion_cube> indeed

08:36 <companion_cube> it needs to be of size 1

08:36 <companion_cube> then the main issue (which made me give up yesterday) is: given a string, how to split it into 'points' (of size 1)

08:37 <companion_cube> unicode lexing is a first step, of course

08:37 ygrek has quit [Ping timeout: 272 seconds]

08:37 <companion_cube> flux: if you're willing to functorize and provide a default instance for unicode strings, then yay!

08:38 badon has joined #ocaml

08:41 <flux> why does it need to split it into anything? perhaps I should read the code to see the problem :). given a string foo, it needs to know that the cursor shall move string_length str characters to the right. and that's it? and then there can be a matrix of strings and one can put whichever stuff into it. \e[1m;| could be one string in a cell, "h" "e" "l" "o" could be four adjacent strings.

08:41 <flux> and the basic element is a "character" that's actually a string, or perhaps "code" or "element" would be a better term..

08:43 thomasga has joined #ocaml

08:44 <companion_cube> oh, you can split into individual strings, indeed

08:44 <companion_cube> it might be costly, but anyway

08:44 <companion_cube> it would be nice also to be able to render to text, or to other formats (html?)

08:46 hay207 has joined #ocaml

08:46 adc90 has quit [Read error: Connection reset by peer]

08:46 <flux> I think a HTML renderer would be a completely different problem :)

08:46 adc90 has joined #ocaml

08:46 <flux> possibly with the same user-facing interface

08:46 <companion_cube> well yes, I meant with the same combinators

08:46 <companion_cube> otoh printbox needs to remain reasonably simple imho

08:46 maurer has joined #ocaml

08:51 mort___ has left #ocaml [#ocaml]

08:54 dsheets has joined #ocaml

08:59 asQuirreL has joined #ocaml

08:59 avsm has quit [Quit: Leaving.]

09:05 michatotol_ has quit [Remote host closed the connection]

09:08 rand000 has joined #ocaml

09:10 rgrinberg has joined #ocaml

09:13 thomasga has quit [Quit: Leaving.]

09:18 kushal has joined #ocaml

09:21 bobry has joined #ocaml

09:22 hay207 has quit [Ping timeout: 244 seconds]

09:22 hay207 has joined #ocaml

09:27 Simn has joined #ocaml

09:31 wildsebastian has joined #ocaml

09:34 maurer has quit [Ping timeout: 246 seconds]

09:34 Kakadu has joined #ocaml

09:34 thomasga has joined #ocaml

09:36 maurer has joined #ocaml

09:40 Bhavya has quit [Quit: Quit the channel]

09:45 lordkryss has joined #ocaml

09:48 sdothum has joined #ocaml

09:50 Haudegen has quit [Ping timeout: 276 seconds]

10:05 avsm has joined #ocaml

10:06 octachron has joined #ocaml

10:16 octachron has quit [Ping timeout: 264 seconds]

10:18 dav- has joined #ocaml

10:20 dav has quit [Ping timeout: 272 seconds]

10:27 oscar_toro has quit [Ping timeout: 264 seconds]

10:30 thomasga has quit [Quit: Leaving.]

10:35 octachron has joined #ocaml

10:35 Haudegen has joined #ocaml

10:36 thomasga has joined #ocaml

10:41 thomasga has quit [Quit: Leaving.]

10:42 rand000 has quit [Quit: leaving]

10:42 octachron has quit [Ping timeout: 250 seconds]

10:42 thomasga has joined #ocaml

10:43 sepp2k has joined #ocaml

10:45 b0mbastus has quit [Ping timeout: 265 seconds]

10:46 dav has joined #ocaml

10:49 dav- has quit [Ping timeout: 264 seconds]

10:50 jonludlam has joined #ocaml

10:51 asQuirreL has quit [Ping timeout: 240 seconds]

10:54 jwatzman|work has quit [Quit: jwatzman|work]

10:58 _andre has joined #ocaml

10:58 axiles has joined #ocaml

11:03 s1n4 has quit [Ping timeout: 246 seconds]

11:06 thomasga has quit [Quit: Leaving.]

11:09 Denommus has quit [Ping timeout: 244 seconds]

11:13 asQuirreL has joined #ocaml

11:13 asQuirreL has quit [Read error: Connection reset by peer]

11:13 asQuirreL has joined #ocaml

11:14 oscar_toro has joined #ocaml

11:22 octachron has joined #ocaml

11:22 asQuirreL has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

11:27 TheLemonMan has joined #ocaml

11:30 tmtwd has quit [Ping timeout: 252 seconds]

11:34 jabesed has joined #ocaml

11:41 ggole has joined #ocaml

11:41 thomasga has joined #ocaml

11:42 sdothum has quit [Quit: ZNC - 1.6.0 - http://znc.in]

11:43 sdothum has joined #ocaml

11:43 tmtwd has joined #ocaml

11:51 yomimono has quit [Ping timeout: 244 seconds]

11:56 arj has joined #ocaml

11:56 thomasga has quit [Quit: Leaving.]

11:57 Haudegen has quit [Ping timeout: 265 seconds]

12:03 jonludlam has quit [Ping timeout: 265 seconds]

12:09 Hannibal_Smith has joined #ocaml

12:10 Denommus has joined #ocaml

12:15 jabesed has quit [Ping timeout: 264 seconds]

12:19 Haudegen has joined #ocaml

12:20 thomasga has joined #ocaml

12:21 BitPuffin|osx has quit [Ping timeout: 276 seconds]

12:23 siddharthv_away is now known as siddharthv

12:34 jabesed has joined #ocaml

12:36 thomasga has quit [Quit: Leaving.]

12:36 milosn has quit [Quit: leaving]

12:42 djellemah_ has joined #ocaml

12:44 thomasga has joined #ocaml

12:46 djellemah has quit [Ping timeout: 256 seconds]

13:01 thomasga has quit [Quit: Leaving.]

13:01 boegel has joined #ocaml

13:01 <boegel> oh hai!

13:02 badon has quit [Disconnected by services]

13:02 badon_ has joined #ocaml

13:02 <boegel> quick question, is there an equivalent in OCaml for something like "python -c 'import numpy'", to check whether an OCaml package is available (or not)?

13:02 <ggole> ocamlfind query, maybe

13:03 <boegel> ggole: that requires findlib, right?

13:04 badon__ has joined #ocaml

13:04 <companion_cube> well, why would you not have findlib ? :p

13:04 <ggole> Yeah, but you should probably have it

13:05 <ggole> (You may need to install it if you have a fresh install or switch.)

13:05 badon__ is now known as badon

13:07 <boegel> ggole: I'm compiling OCaml myself, and I want to look into a 'batteries included' type of installation, i.e. OCaml + findlib + packages in a single location

13:07 badon_ has quit [Ping timeout: 256 seconds]

13:07 <companion_cube> aww

13:08 <ggole> boegel: is opam not suitable?

13:09 <boegel> ggole: I'm installing OCaml such that it's version and package versions are fixed, and not updated later (unless OCaml (or selected packages) get reinstalled to another prefix)

13:09 <boegel> so I don't think opam is what I need

13:10 <ggole> I don't think doing it yourself would be any easier

13:12 <boegel> ggole: isn't the prime goal of opam keeping things up-to-date easily?

13:13 idegen has joined #ocaml

13:14 thomasga has joined #ocaml

13:14 <smondet> boegel: you can also use opam to fix the versions of everything (with `opam pin` or you own opam-repo)

13:15 <smondet> (s/you/your/

13:15 <ggole> boegel: it makes installing things (like findlib) dead simple

13:16 <ggole> You can also switch between different compiler versions, which is nice if you want to check out new language features that you may not be able to otherwise use

13:16 <boegel> smondet: opam compiles things from source, or installs binary packages?

13:16 <boegel> ggole: to try a new OCaml version, I'd just install everything in a new prefix (I'm not installing it for myself, but for our users on our supercomputer)

13:17 lordkryss has quit [Quit: Connection closed for inactivity]

13:17 <ggole> (Isn't that a lot of work?)

13:17 <ggole> Well, up to you.

13:18 <theblatte> boegel: it compiles from source

13:18 <smondet> boegel: everything from source

13:18 Hannibal_Smith has quit [Quit: Leaving]

13:18 <boegel> ggole: not really, since we automate things, just need to figure it out once (see http://hpcugent.github.io/easybuild)

13:18 <boegel> ok, then I guess I should reconsider opam

13:19 <boegel> I'm guessing opam itself is written in OCaml? :)

13:19 <thizanne> yup

13:20 <boegel> ok, great :P

13:22 <theblatte> I find this not ideal as you need to install ocaml to install opam to install ocaml :p

13:24 <boegel> theblatte: well, I'm fine with building OCaml (+ opam) from source

13:25 <boegel> just for fun, this is one of the reasons I need to install OCaml: http://users.ugent.be/~kehoste/QIIME.pdf (one of the dependencies is pplacer, which is written in OCaml)

13:27 jonludlam has joined #ocaml

13:29 <boegel> "Run `make lib-ext` as advertised by `./configure` if you don't have the dependencies installed and only need the opam binary (not the libs)."

13:29 <boegel> how do I know if I need the libraries?

13:32 <reynir> hrm, opam in debian testing (or unstable?) is still at 1.2.0

13:38 thomasga has quit [Quit: Leaving.]

13:39 thomasga has joined #ocaml

13:44 tmtwd has quit [Ping timeout: 246 seconds]

13:45 thomasga has quit [Quit: Leaving.]

13:46 darkf_ has joined #ocaml

13:46 darkf has quit [Disconnected by services]

13:47 darkf_ is now known as darkf

13:48 axiles has quit [Ping timeout: 244 seconds]

13:49 thomasga has joined #ocaml

13:51 thomasga has quit [Client Quit]

13:53 BitPuffin has joined #ocaml

13:56 ollehar1 has joined #ocaml

13:56 shinnya has joined #ocaml

14:00 tmtwd has joined #ocaml

14:00 axiles has joined #ocaml

14:06 <zozozo> is there a direct access to keyboard events in lablgtk ? I can't seem to find any

14:09 <adrien_znc> zozozo: in gdk or something

14:10 dav- has joined #ocaml

14:11 dav has quit [Ping timeout: 252 seconds]

14:11 freling has joined #ocaml

14:12 axiles has quit [Quit: Quitte]

14:13 <zozozo> adrien_znc: thx

14:21 kdef has joined #ocaml

14:29 TheLemonMan has quit [Quit: "It's now safe to turn off your computer."]

14:32 <boegel> so, binaries for stuff installed with opam go to $OPAMROOT/system/bin ?

14:32 jabesed has quit [Ping timeout: 250 seconds]

14:34 jabesed has joined #ocaml

14:39 jabesed has quit [Ping timeout: 258 seconds]

14:40 <boegel> the opam --help is broken; it mentions a 'config' subcommand, but that just yields "opam: Missing subcommand. Valid subcommands are ..."

14:43 palomer has joined #ocaml

14:46 bobry has quit [Quit: Connection closed for inactivity]

14:46 <Drup> config is not a subcommand, it's a command

14:46 <Drup> or, alternatively, config takes subsubcommands ...

14:49 agarwal1975 has joined #ocaml

14:50 <boegel> Drup: you mean, I should run 'config' rather than 'opam config'?

14:51 <boegel> Drup: ah, no, I see now, my bad

14:51 <boegel> I should be using 'opam config <subcmd>'

14:51 * boegel whacks himself in the head

14:51 avsm has quit [Quit: Leaving.]

14:52 darkf has quit [Quit: Leaving]

14:53 thomasga has joined #ocaml

14:54 octachron has quit [Quit: Leaving]

15:03 Cyanure has joined #ocaml

15:04 thomasga has quit [Quit: Leaving.]

15:05 wildsebastian has quit [Remote host closed the connection]

15:11 thomasga has joined #ocaml

15:12 Cyanure has quit [Read error: Connection reset by peer]

15:18 struktured has quit [Ping timeout: 265 seconds]

15:19 <boegel> is there a way to control where OPAM downloads source tarballs too, and/or to provide locations where they were downloaded before (so they don't get redownloaded)?

15:19 milosn has joined #ocaml

15:20 <Drup> 1) opam caches tarballs, so no re-download

15:20 <boegel> Drup: caches them where?

15:20 <Drup> 2) you could create your own opam-repository to really control both location and versions

15:20 dav- is now known as dav

15:21 <Drup> .opam/archives, iirc

15:22 <boegel> hmm, ok

15:22 <boegel> thx

15:26 palomer has quit [Quit: palomer]

15:29 Cyanure has joined #ocaml

15:33 sepp2k has quit [Ping timeout: 256 seconds]

15:34 asQuirreL has joined #ocaml

15:41 palomer has joined #ocaml

15:43 A1977494 has joined #ocaml

15:46 asQuirreL has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

15:47 A19774941 has joined #ocaml

15:49 A1977494 has quit [Ping timeout: 246 seconds]

15:50 ollehar1 has quit [Ping timeout: 250 seconds]

15:51 rgrinberg has quit [Ping timeout: 265 seconds]

15:54 rgrinberg has joined #ocaml

15:55 <reynir> Drup: re furl, I think maybe the logic for Slash/NoSlash is reversed? NoSlash seems to add a slash at the end :)

15:56 <Drup> oh ?

15:56 <reynir> at least when I Furl.eval it

15:56 <Drup> ahah, you are right

15:56 <reynir> heh :)

15:57 <Drup> should MaybeSlash put a slash at the end or not ?

15:57 <reynir> Maybe :P

15:57 <Drup> ಠ_ಠ

15:57 <reynir> Sorry, I don't have an opinion on that :)

15:58 A19774941 has quit [Ping timeout: 245 seconds]

15:59 <Drup> fixed

15:59 <reynir> ^^

16:00 jeffmo has joined #ocaml

16:01 <reynir> hrm, is it safe to upgrade opam?

16:01 <Drup> yeah, I haven't wrote the test suite yet ...

16:03 A1977494 has joined #ocaml

16:04 <Drup> upgrade opam to what ? The lastest stable ? Sure

16:04 <reynir> yes

16:05 sepp2k has joined #ocaml

16:05 A19774941 has joined #ocaml

16:07 madroach has quit [Ping timeout: 264 seconds]

16:07 madroach has joined #ocaml

16:08 A1977494 has quit [Ping timeout: 256 seconds]

16:10 dubosec has quit [Ping timeout: 265 seconds]

16:11 jabesed has joined #ocaml

16:13 slash^ has joined #ocaml

16:15 thegameg has joined #ocaml

16:17 mcclurmc has quit [Ping timeout: 276 seconds]

16:20 dav- has joined #ocaml

16:21 tmtwd has quit [Ping timeout: 252 seconds]

16:22 avsm has joined #ocaml

16:22 dav has quit [Ping timeout: 255 seconds]

16:30 freling has quit [Quit: Leaving.]

16:34 tmtwd has joined #ocaml

16:35 ncthom91 has joined #ocaml

16:35 ollehar1 has joined #ocaml

16:35 <ollehar1> test

16:35 <ollehar1> can't install ppx_tools with 4.02+modular-implicits

16:35 <Drup> that is not very surprising

16:35 <ollehar1> no?

16:35 <Drup> I think there is a patch in ppx_tools's bugtracker

16:36 <ncthom91> hey all. Has anybody here used ocamllex/ocamlyacc for parsing HTML? I know HTML is a particularly unusual grammar, but if you ignore the error-friendliness you can simplify things considerably

16:37 <Drup> there are things

16:37 <Drup> not of them are done with ocamllex/ocamlyacc, though

16:38 <Drup> there is a sloppy html parser in ocamlnet, but it's for html 4

16:38 <Drup> and there are things based on xmlm, done by dsheets

16:40 <dsheets> doesn't really parse html... more like xml/html polyglot

16:40 <Drup> yes I know, but well ...

16:40 <dsheets> i.e. can't use it to build a crawler

16:41 <dsheets> html isn't context-free so i don't think yacc will work...

16:42 <Drup> you mean, the sloppy version ?

16:42 <dsheets> all versions of xml/html/sgml are context-free

16:43 <dsheets> erm *not* context-free

16:45 avsm has quit [Quit: Leaving.]

16:45 ollehar has quit [Remote host closed the connection]

16:50 <ncthom91> dsheets what does it mean for a grammar to be non context-free?

16:53 <ncthom91> ah http://en.wikipedia.org/wiki/Context-sensitive_grammar ;)

16:53 <ncthom91> didn't know it by that name

16:53 ollehar has joined #ocaml

16:53 tane has joined #ocaml

16:54 <dsheets> after looking, i retract my statement. well-formed sgml, html, and xml are probably context-free

16:54 <ollehar1> syntastic in vim can't handle [@@deriving]?

16:54 <dsheets> if you want to crawl, you can't rely on well-formedness, though

16:55 ollehar has quit [Remote host closed the connection]

16:56 ollehar has joined #ocaml

16:56 <ollehar1> test

16:56 jwatzman|work has joined #ocaml

16:56 <ollehar1> (sorry, problem with IRC)

16:56 <dsheets> ncthom91, http://stackoverflow.com/questions/5175840/is-html-a-context-free-language

16:56 jwatzman|work has quit [Client Quit]

16:57 <Drup> dsheets: I was really wondering how would xml be non context free

16:57 <Drup> (well formed)

16:57 <ncthom91> dsheets right, I was only curious about well-formed HTML. Interested in building a toy browser engine for fun + learning, definitely dont' need to support HTML to spec :)

16:58 <dsheets> i was just confused... when i tried to find a counterexample, i also failed and then started googling

16:59 <Drup> well, for valid html, heeeh, not sure

17:00 <dsheets> what is 'valid' when html5 defines broken junk as well-defined?

17:00 <Drup> I mean valid as "right attributes on the right tags"

17:00 <Drup> but you don't usually consider that as part of the grammar.

17:01 ingsoc1 has joined #ocaml

17:01 <bernardo1pc> /bin/sh: 1: ocamlfind: not found -> beuh

17:01 <bernardo1pc> opam does list findlib 1.5.5 installed :/

17:01 <Drup> bernardo1pc: the eval dance ?

17:02 ingsoc has quit [Ping timeout: 265 seconds]

17:03 arj has quit [Quit: Leaving.]

17:03 <bernardo1pc> right

17:03 <bernardo1pc> why is this the first time this backfires on me ?

17:03 <bernardo1pc> pure luck ?

17:04 <Drup> I don't know. You switched ?

17:05 <Drup> you do have the eval thingy in your .bashrc/.profile/.stuffrc ?

17:05 <bernardo1pc> no, I did not switch

17:05 <bernardo1pc> yes, I have eval in my .rc

17:05 <bernardo1pc> I just upgraded zarith + mirage-nox

17:05 <Drup> I don't know.

17:06 <bernardo1pc> and the first thing I tried to compile after was fubar

17:06 ollehar has quit [Remote host closed the connection]

17:07 ollehar has joined #ocaml

17:07 <bernardo1pc> well, I this happens again, I will try to make a test case

17:07 srcerer has quit [Quit: ChatZilla 0.9.91.1 [Firefox 37.0.2/20150415140819]]

17:07 ygrek has joined #ocaml

17:08 srcerer has joined #ocaml

17:09 Sim_n has joined #ocaml

17:09 ollehar has quit [Remote host closed the connection]

17:09 dsheets has quit [Ping timeout: 258 seconds]

17:10 ollehar1 has quit [Quit: ollehar1]

17:10 ollehar has joined #ocaml

17:10 Simn has quit [Ping timeout: 252 seconds]

17:12 ollehar has quit [Remote host closed the connection]

17:13 ollehar has joined #ocaml

17:14 simn__ has joined #ocaml

17:15 Sim_n has quit [Ping timeout: 252 seconds]

17:18 Simn has joined #ocaml

17:20 Simn has quit [Client Quit]

17:20 simn__ has quit [Ping timeout: 258 seconds]

17:22 agarwal1975_ has joined #ocaml

17:23 palomer has quit [Quit: palomer]

17:23 agarwal1975 has quit [Quit: agarwal1975]

17:23 agarwal1975_ is now known as agarwal1975

17:24 Hannibal_Smith has joined #ocaml

17:24 andreypopp has joined #ocaml

17:26 tmtwd has quit [Ping timeout: 256 seconds]

17:27 ingsoc1 has quit [Quit: Leaving.]

17:27 ingsoc has joined #ocaml

17:30 kushal has quit [Ping timeout: 252 seconds]

17:38 tmtwd has joined #ocaml

17:39 ollehar1 has joined #ocaml

17:39 <ollehar1> using two menhir projects in one project is not possible? having problems with Lexing module

17:39 <ollehar1> and types

17:40 ncthom91 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

17:43 <ollehar1> Lexing wants to produce Parser.token, by I want MyParser.token :P

17:43 <ollehar1> or MyLexing, for that matter

17:44 tmtwd has quit [Ping timeout: 256 seconds]

17:46 ncthom91 has joined #ocaml

17:48 <reynir> ollehar1: regarding syntastic and [@@deriving], IIRC it uses camlp4 for its syntax checker by default. I recommend using the merlin checker instead

17:49 <ollehar1> reynir: good point

17:50 <reynir> the option is maybe let g:syntastic_ocaml_checkers=merlin

17:50 <flux> ollehar1, hmm, I don't have anything called Parser in my project, though it uses ocamlyacc instead of menhir, but as it should be drop-in-replacement, I don't think there should be any issue..

17:50 b0mbastus has joined #ocaml

17:50 <ollehar1> flux: ok

17:50 <ollehar1> hm

17:50 dubosec has joined #ocaml

17:50 <ollehar1> I will try again, might have been issue with all my renaming :P

17:51 penryu has left #ocaml ["WeeChat 1.1-dev"]

17:51 TheLemonMan has joined #ocaml

17:57 wwilly has joined #ocaml

17:58 MrScout has joined #ocaml

17:58 tmtwd has joined #ocaml

18:01 ollehar1 has quit [Ping timeout: 252 seconds]

18:03 Kakadu has quit [Quit: Page closed]

18:09 BitPuffin has quit [Ping timeout: 246 seconds]

18:14 swgillespie has joined #ocaml

18:26 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

18:27 Denommus` has joined #ocaml

18:28 Denommus has quit [Ping timeout: 245 seconds]

18:28 swgillespie has joined #ocaml

18:28 Denommus` is now known as Denommus

18:43 tobiasBora has quit [Ping timeout: 258 seconds]

18:44 sgnb has quit [Read error: Connection reset by peer]

18:44 sgnb has joined #ocaml

18:48 tobiasBora has joined #ocaml

18:51 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

18:54 <TheLemonMan> silly question here, I have a module where I've defined an 'interface' and it includes a type I don't want to repeat in the modules, how should I go ?

18:56 mcclurmc has joined #ocaml

18:57 degenerate-pyon is now known as terminal-pyon

18:58 swgillespie has joined #ocaml

18:58 <Unhammer> I think you may have to repeat it :/

18:59 Kakadu has joined #ocaml

18:59 <Unhammer> (unless you want it to be hidden, in which case you can do "type t" in the interface but "type t = something" in the implementation)

19:00 <Unhammer> http://stackoverflow.com/questions/3238509/redundancy-in-ocaml-type-declaration-ml-mli

19:01 <Unhammer> "A common idiom is to have a module that is dedicated to type declarations, and make it have only a .mli"

19:01 thomasga1 has joined #ocaml

19:02 thomasga has quit [Ping timeout: 272 seconds]

19:03 thomasga1 has quit [Client Quit]

19:04 leafac has joined #ocaml

19:05 rand000 has joined #ocaml

19:06 ggole has quit []

19:08 slash^ has quit [Read error: Connection reset by peer]

19:08 willy_ has joined #ocaml

19:11 wwilly has quit [Ping timeout: 255 seconds]

19:19 asQuirreL has joined #ocaml

19:21 matason has quit []

19:22 <TheLemonMan> heh, that was unexpected

19:22 mcclurmc has quit [Remote host closed the connection]

19:24 thomasga has joined #ocaml

19:26 rgrinberg has quit [Ping timeout: 256 seconds]

19:27 A19774941 has left #ocaml [#ocaml]

19:28 jabesed has quit [Ping timeout: 272 seconds]

19:29 A1977494 has joined #ocaml

19:31 MercurialAlchemi has quit [Remote host closed the connection]

19:31 MercurialAlchemi has joined #ocaml

19:32 A1977494 has left #ocaml [#ocaml]

19:33 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

19:34 ncthom91 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

19:38 Simn has joined #ocaml

19:43 matason has joined #ocaml

19:49 ollehar has quit [Remote host closed the connection]

19:50 ingsoc1 has joined #ocaml

19:51 ingsoc has quit [Ping timeout: 255 seconds]

19:57 freling has joined #ocaml

20:03 leafac has quit [Ping timeout: 276 seconds]

20:03 mcclurmc has joined #ocaml

20:04 thomasga has quit [Quit: Leaving.]

20:05 thomasga has joined #ocaml

20:09 b0mbastus has quit [Ping timeout: 256 seconds]

20:10 ollehar has joined #ocaml

20:12 _andre has quit [Quit: leaving]

20:13 Algebr has joined #ocaml

20:15 <Algebr> how can I get opam to install a local opam package? this package does not exist in opam but I do have an opam file

20:16 <Drup> opam pin

20:16 <ygrek> if you keep it together with sources then `opam pin ` should do

20:18 <Algebr> ah right, I forgot about pin

20:18 jabesed has joined #ocaml

20:23 rgrinberg has joined #ocaml

20:23 mcclurmc has quit [Remote host closed the connection]

20:23 <Drup> rgrinberg: please review that https://github.com/ocsigen/ocsigenserver/commit/70ce7ba752cb9169426f89e44999ef4604e59d47

20:26 swgillespie has joined #ocaml

20:28 <Drup> wow, this code is so ridiculously bad written

20:28 <Drup> seriously dinosaure u_u'

20:28 Cyanure has quit [Remote host closed the connection]

20:32 <rgrinberg> Drup: where is he again

20:33 <rgrinberg> is there a particular reason you want me to review this commit?

20:34 <Drup> it's supposed to fix some tls stuff, but it's doing a bunch of studd with conduit/cohttp that I don't understand

20:34 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

20:36 <rgrinberg> it looks fine but i'll be honest and say that I don't know that much about conduit

20:36 <Drup> rgrinberg: it's to fix the first of jerome's issues: "Ocsigen_request_info.get_server_address is not implemented but is used by the reverse proxy;"

20:38 <rgrinberg> Drup: so where's teh implementation of get_server_address now

20:38 <Drup> are there no other way to do that ? It seems quite convoluted

20:38 <Drup> https://github.com/ocsigen/ocsigenserver/compare/cohttp_rebased...fix_get_server_address look at 4 last commit

20:39 swgillespie has joined #ocaml

20:42 <rgrinberg> so where is server_addr actually being set

20:42 <Drup> in the last commit

20:44 <rgrinberg> i don't see it :/

20:44 <Drup> https://github.com/ocsigen/ocsigenserver/commit/70ce7ba752cb9169426f89e44999ef4604e59d47#diff-af00eddd5261fa6f6ab200223ead43f5R212

20:44 <Drup> address matched here

20:45 <Drup> https://github.com/ocsigen/ocsigenserver/commit/70ce7ba752cb9169426f89e44999ef4604e59d47#diff-af00eddd5261fa6f6ab200223ead43f5R219

20:45 <Drup> and passed here

20:45 <Drup> yes, he's using the same name for two different things :]

20:45 <rgrinberg> it's called ai_addr

20:47 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

20:49 terminal-pyon has quit [Ping timeout: 255 seconds]

20:49 <Drup> rgrinberg: wouldn't be better if server_address was a Unix.sockaddr ?

20:50 swgillespie has joined #ocaml

20:50 <Drup> this whole code doesn't really make sense to me, why going through conduit and all that to get the inet adresse ?!

20:51 <rgrinberg> well conduit is the one that establishes the tcp connection i think

20:51 pyon has joined #ocaml

20:52 <rgrinberg> Drup: i must be retarded btw. the field that was added is called `server_addr` yet i don't see it set anywhere

20:52 <rgrinberg> your link showed ai_addr

20:53 <rgrinberg> oh this is the magic trick

20:53 <rgrinberg> https://github.com/ocsigen/ocsigenserver/commit/93418b5e5a614b413b9e9af50432f9f67ad5c24a

20:53 <rgrinberg> nvm

20:54 <Drup> ok, you found it before I typed the whole answer :D

20:54 <rgrinberg> punning conditioned me to read code a certain way :/

20:54 <Drup> basically, it ended up being passe through the argument ~address of the handler function

20:55 <Drup> passed*

20:55 <rgrinberg> should this server_addr even be request scoped?

20:56 freling has quit [Quit: Leaving.]

20:56 willy_ has left #ocaml ["Leaving"]

20:56 Algebr has quit [Remote host closed the connection]

20:56 <rgrinberg> Drup: could you rebase that branch though? I want to see the whole diff

20:56 <Drup> sure

20:57 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

20:57 swgillespie has joined #ocaml

20:59 s1n4 has joined #ocaml

21:00 <Drup> rgrinberg: https://github.com/ocsigen/ocsigenserver/compare/cohttp_rebased...fix_get_server_address

21:02 hay207 has quit [Ping timeout: 264 seconds]

21:03 MercurialAlchemi has quit [Ping timeout: 240 seconds]

21:04 Simn has quit [Quit: Leaving]

21:04 srcerer has quit [Read error: Connection timed out]

21:04 <rgrinberg> i want to see how get_server_address was implemented before

21:05 <rgrinberg> still, i think this stuff will work

21:05 tane has quit [Quit: Verlassend]

21:05 <rgrinberg> Drup: it probably doesn't make much sense to worry about aesthetics too much here

21:06 srcerer has joined #ocaml

21:06 <rgrinberg> Drup: maybe you've heard of the phrase "pissing in the wind"...

21:06 matason has quit []

21:06 <Drup> asthetics is not a problem

21:07 <Drup> code base littered with assert false waiting to explode on your face is.

21:08 <rgrinberg> yeah but in this case it works. obviously the assert false it totally out of place...

21:09 <Drup> well, the first one is just shitty code

21:09 <Drup> the second one is more worrying, because I don't really know how to remove it

21:10 <rgrinberg> s/assert false/invalid_arg "fuck you dinosaure"/ perhaps?

21:10 <rgrinberg> s/invalid_arg/failwith/

21:11 <Drup> well, no

21:11 <Drup> I mean, this one: https://github.com/ocsigen/ocsigenserver/compare/cohttp_rebased...fix_get_server_address#diff-af00eddd5261fa6f6ab200223ead43f5R372

21:12 <rgrinberg> yeah i've been staring at it

21:12 <rgrinberg> the only thing i've realized is that i'm getting old

21:12 <rgrinberg> b/c i can't even fucking tell which match statement it belongs to

21:12 <Drup> ahahah

21:12 <rgrinberg> or is it the indentation that's messed up?

21:12 <Drup> https://github.com/ocsigen/ocsigenserver/compare/cohttp_rebased...fix_get_server_address#diff-af00eddd5261fa6f6ab200223ead43f5R347

21:15 <Drup> rgrinberg: I push the code cleanup commit

21:16 rand000 has quit [Ping timeout: 252 seconds]

21:17 rand000 has joined #ocaml

21:18 jonludlam has quit [Ping timeout: 264 seconds]

21:19 <rgrinberg> yeah i'm not totally sure here..

21:20 <rgrinberg> sorry Drup my unix beard is failing me

21:20 <rgrinberg> i like dinosaure's bravery though

21:20 <rgrinberg> except i'm not sure why he didn't just use List.hd

21:21 <Drup> rgrinberg: to extract the content of ai_addr directly, I guess

21:22 <rgrinberg> i'm definitely getting old...

21:23 Denommus` has joined #ocaml

21:23 <rgrinberg> I'd still name the darn thing and then put the pattern in the let e.g. let ... = List.hd xxx

21:24 tsumetai has quit [Ping timeout: 244 seconds]

21:24 <Drup> that's a style choice, I don't care.

21:24 <Drup> ok, pushing on the main branch

21:25 <rgrinberg> it's not a style choice when you add an assert false to the pattern

21:25 <rgrinberg> but anyway it's off topic

21:25 Denommus has quit [Ping timeout: 272 seconds]

21:26 <Drup> rgrinberg: you will have an assert false anyway, there are two variants in ai_addr

21:26 <Drup> that is the main issue

21:26 <Drup> the list is barely a problem

21:26 tmtwd has quit [Ping timeout: 246 seconds]

21:28 <rgrinberg> Drup: oh yeah

21:28 Denommus` has quit [Ping timeout: 264 seconds]

21:28 <rgrinberg> Drup: does ocsigen support proxying on a unix domain socket?

21:29 <Drup> dinosaure claim that it doesn't

21:30 Denommus has joined #ocaml

21:32 manizzle has joined #ocaml

21:32 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

21:32 asQuirreL has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

21:32 ygrek has quit [Ping timeout: 256 seconds]

21:33 axiles has joined #ocaml

21:34 swgillespie has joined #ocaml

21:34 <Drup> the remaining bullet point is the infamous awake parameter

21:37 ingsoc1 has quit [Ping timeout: 256 seconds]

21:39 <rgrinberg> i thought i addressed it

21:39 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

21:39 <rgrinberg> there's so much other dead code

21:39 <Drup> btw

21:39 <Drup> does anyone know is where is alain frish's dead code analyzer and if it's still working ?

21:39 <Drup> -is

21:43 tmtwd has joined #ocaml

21:43 <Drup> https://github.com/alainfrisch/dead_code_analyzer

21:43 <Drup> "This project has just been started. Nothing is ready yet."

21:43 <Drup> well.

21:45 tinhead has joined #ocaml

21:48 mcclurmc has joined #ocaml

21:53 ingsoc has joined #ocaml

21:54 rgrinberg has quit [Ping timeout: 256 seconds]

21:55 Hannibal_Smith has quit [Quit: Leaving]

21:56 badon has quit [Quit: Leaving]

21:57 swgillespie has joined #ocaml

21:58 jonludlam has joined #ocaml

21:59 ontologiae_ has joined #ocaml

21:59 Gama11 has quit [Read error: Connection reset by peer]

22:01 jabesed has quit [Ping timeout: 264 seconds]

22:03 TheLemonMan has quit [Quit: "It's now safe to turn off your computer."]

22:05 milosn has quit [Quit: Reconnecting]

22:05 milosn has joined #ocaml

22:07 s1n4 has quit [Ping timeout: 250 seconds]

22:08 tsumetai has joined #ocaml

22:09 Kakadu has quit [Remote host closed the connection]

22:11 swgillespie has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]

22:11 jabesed has joined #ocaml

22:16 <dmbaturin> Lwt_log adventures, ep2: on a machine with normal syslog, info and debug messages created with the syslog logger don't make it to the log either. logger -p user.info something works fine.

22:19 <dmbaturin> * it works on the journald machine as well.

22:26 <Drup> are they printed on stderr with the default logger ?

22:29 jeffmo has quit [Quit: jeffmo]

22:29 dav has joined #ocaml

22:29 <dmbaturin> Drup: Just checked. No, they aren't.

22:30 <Drup> dmbaturin: read http://ocsigen.org/lwt/2.4.8/api/Lwt_log_core#TYPEsection

22:30 <Drup> (I should have directed you there sooner ~~)

22:30 dav- has quit [Ping timeout: 264 seconds]

22:30 Algebr has joined #ocaml

22:31 ygrek has joined #ocaml

22:33 sepp2k has quit [Quit: Leaving.]

22:34 Denommus` has joined #ocaml

22:34 ontologiae_ has quit [Ping timeout: 256 seconds]

22:36 Denommus has quit [Ping timeout: 272 seconds]

22:36 tmtwd has quit [Read error: Connection reset by peer]

22:36 struktured has joined #ocaml

22:36 <dmbaturin> Drup: Thanks, looks like Lwt_log.add_rule "*" Lwt_log.Debug did the trick.

22:37 <Algebr> If multicore does indeed come, will lwt/async become irrelevant or need a massive refactor?

22:42 BitPuffin|osx has joined #ocaml

22:43 <ygrek> no

22:43 <ygrek> is orthogonal

22:46 <Drup> well, yes and no

22:46 <Drup> Lwt is basically incompatible with fibers :(

22:46 thomasga has quit [Quit: Leaving.]

22:49 jabesed has quit [Ping timeout: 250 seconds]

22:51 rgrinberg has joined #ocaml

22:52 freling has joined #ocaml

22:52 <companion_cube> Algebr: people use lightweight threads or continuations even in multicore languages

22:53 tmtwd has joined #ocaml

22:55 <ygrek> Drup, what that can possible mean?

22:55 <ygrek> lwt is a way to organize fibers

22:57 <ygrek> one problem wrt multicore in lwt is that it doesn't support multiple threads - but I believe it is fixable

22:59 <Drup> ygrek: I meant multicore's fibers, as in the one that migrates across domains

22:59 <Drup> well, it's rather equivalent to what you said

23:03 <ygrek> you mean native OS thread?

23:04 <Drup> yeah

23:05 <ygrek> I don't feel it's a problem, lwt for example could just retain current model, but allow to run many event loops in many threads - one event loop per thread

23:05 <ygrek> another option is to go fancy and to manage all threads hidden behind lwt

23:12 freling has quit [Quit: Leaving.]

23:13 ncthom91 has joined #ocaml

23:15 dubosec has quit [Quit: Leaving]

23:17 swgillespie has joined #ocaml

23:21 mfp has quit [Read error: Connection reset by peer]

23:28 badkins has joined #ocaml

23:31 mfp has joined #ocaml

23:33 avsm has joined #ocaml

23:34 madroach has quit [Ping timeout: 264 seconds]

23:35 madroach has joined #ocaml

23:48 Denommus has joined #ocaml

23:50 rgrinberg has quit [Ping timeout: 264 seconds]

23:51 tinhead has quit [Remote host closed the connection]

23:51 <yminsky> We were discussing with the OCaml Labs folk recently, and I agree with ygrek. Lwt (and Async) can just have one fiber per domain that runs the scheduler. You wouldn't use much of the fiber framework, but the ability to have multiple fibers in different domains would be useful in an Lwt/Async world.

23:51 <yminsky> Now, it may be that people feel less need for Lwt or Async once we have fibres.... I'm less sure about that.

23:53 <Algebr> What do fibers exactly mean?

23:54 <yminsky> The thing that Async gives you (and to a lesser degree Lwt does the same) is a clear understanding of what interleavings are possible. That's precisely what you'd lose going straight to fibers. You'd have the usual threading story.

23:54 rgrinberg has joined #ocaml

23:54 <yminsky> Fibers are the thread-like things that are what gets multiplexed over each domain. Think of domain as the heavy-weight context, with a s

23:55 <yminsky> Think, one domain per processor, many fibers per domain.

23:55 <Algebr> yminsky: thank you

23:55 <yminsky> (and to deal with crappy non-blocking semantics on UNIX, multiple threads per domain, but one lock.)

23:58 <ygrek> one runtime lock you mean?

23:58 <yminsky> They're called fibers to avoid confusion with threads. But some confusion on this point is I think inevitable...

23:58 adc90 has quit [Read error: Connection reset by peer]

23:58 <yminsky> Yes, one runtime lock per domain.

23:58 adc90 has joined #ocaml

23:58 <ygrek> ok, now that makes sense

23:58 <ygrek> terminology

23:58 <yminsky> Indeed.

23:59 <yminsky> The interesting bit of the design is that the fiber scheduler is actually completely up to user-land.

23:59 <yminsky> (by which I mean, not mandated by the runtime.)

23:59 <ygrek> but there will be one in runtime?

23:59 <yminsky> Algebraic effects (think, souped-up exceptions) are used for interrupting and restarting fibers.

23:59 <ygrek> I guess it could be more efficient - no?

23:59 <yminsky> I'm sure they're going to have some example schedulers. But there will be nothing built in.