00:12 <hcube> yes, this seems a networking problem

00:28 <rixed_> hcube: related to DNS ?

00:30 <hcube> i dont test it deeply. just disabled image fetching but added 200 new test threads which prints a message repeatedly the list scrolling was smooth.

00:30 <hcube> i'll check this issue in depth tomorrow

00:30 <hcube> thanks for the help

09:10 <thomasga> what can I do to make ocamlbuild handle dependencies in sub-sub directories ?

09:11 <thomasga> (ie. in sub-directories, everything seems to work fine, but when I go on directory up, dependencies are computed correctly, but files are not built in the right order)

09:11 <gildor> "sub/sub": include

09:12 <thomasga> <sub/sub>: inclue or "sub/sub": include ?

09:12 <gildor> "sub/sub": include

09:12 <gildor> <sub/sub> means that there can be expansion

09:12 <gildor> but this is not the case

09:13 <thomasga> I've replace the < > by "" but I still have the problem

09:14 <thomasga> _build/sub/sub/a.ml.depends is there and contains the right information

09:15 <thomasga> but ocamlbuild doesn't try to build the dependencies ...

09:15 <gildor> source code ?

09:16 <thomasga> I haven't committed that code anywhere yet, I am just trying to build it before pushing to github :-)

09:16 <thomasga> what do you need ?

09:17 <gildor> something to reproduce the bug !

09:18 <thomasga> :-)

09:20 <thomasga> gildor: see your PM

09:27 <thomasga> ok, ocamlbuild doesn't like symbolic links in sub-sub directories ...

09:27 <thomasga> thx gildor :-)

09:28 <jsk> morning

13:05 <thomasga> any way to add a manual dependency with ocamlbuild ?

13:24 <gildor> thomasga: use the "dep" function into your myocamlbuild.ml

13:25 <gildor> thomasga: dep ["link"; "ocaml"; "myfunnytag"] ["src/mystrangefile.c"]

13:29 <thomasga> thx!

14:15 <thomasga> If I have A/a.ml, A/b.ml and B/a.ml and B/b.ml and I want to pack A/a.ml and B/b.ml into A/a.cmxa and B/a.ml and B/b.ml into B/b.cmxa; is there an easy way (using _tags) to do that while keeping ocamldep happy ? I've tried to include A et B into a root _tags, but it doesn't work as name clashes so I'm a bit stuck

14:15 <thomasga> (and let say A/a.ml and B/a.ml depends on C/c.ml for example)

14:28 <gildor> thomasga: don't see an obvious solution

14:34 <adrien> I have a project were subdirectories are kept independant but have files with the same names, I ended up using several _tags files and cd'ing into each directory

14:36 <thomasga> I think I found a solution <A/a.*>: include and <B/b.*>: include seem to work ...

14:36 <thomasga> (well <A/a.*> or <B/b.*>: include)

14:45 <gildor> thomasga: tricky

14:49 <thomasga> actually, not sure it does really what I want ...

16:01 <thomasga> again an ocamlbuild question: I have a file named pervasives.ml that I want to use to compile my project, instead of the one in the standard distrib - so I use nopervasives and I've added a dep ["ocaml";"compile"] ["[...]/pervasives.cmi"]; all of that seems to work - however when ocamlbuild then try to compile pervasives.mli it finds a circular dependency ... I've tried to use non_dependency "[..].pervasives" "Pervasives" in myocamlbuild.ml but that does

16:05 <f[x]> -nostdlib?

16:05 <thomasga> yea I added it already

16:06 <thelema> thomasga: you added the dependency on pervasived.cmi to compiling pervasives.cmi, of course there's a circle

16:08 <thomasga> well, I would like to add the dependency for every files apart pervasives.ml{,i}, do you know how to do that ?

16:09 <thomasga> ha maybe using 'not <[...]/pervasives.*>: pervasive_depend' in _tags

16:11 <thomasga> cool, that works

17:09 <thelema> http://www.infoq.com/interviews/functional-langs <- interesting video

17:58 <flux> hm, did batteris have a module for nice map of sets?

17:58 <flux> handling the empty case etc

19:12 <flux> I put this together, but I guess it isn't general enough: http://www.modeemi.fi/~flux/software/ocaml/mapSet.mli http://www.modeemi.fi/~flux/software/ocaml/mapSet.ml

19:13 <flux> an optimal solution would allow constructing "n-dimensional" trees by using MapSet again in a simple fashion

19:14 <flux> hm, "n-degree" instead?

19:15 <flux> I think those kind of data structures (multiple values for one key) pop up relatively often

19:18 <hcarty> flux: That does look usefule

19:18 <hcarty> useful

19:21 <flux> I shall be motivated to finalize it a bit, then :)

19:22 <flux> hcarty, how do you feel about the lack of full genericity of the module?

19:47 <hcarty> flux: I'm not sure - I'd probably have to play with it a bit

19:49 <hcarty> flux: I haven't looked at the implementation in detail - how difficult would it be to make this a generic multi-value map?

19:50 <hcarty> Set, List, etc?

19:50 <hcarty> Or is that the lack of genericity you're talking about? :-)

19:50 <flux> hcarty, yes :)

19:51 <flux> I'm thinking there would need to be a module type with at least empty, add, remove, enum, find, compare, equal as a functor parameter

19:52 <adrien> hcarty: completely unrelated: I started working on support for tiling gtk widgets, that is actually a generalization of vertical and horizontal boxes and it should provide support for constructs like window#add `vertical [ menu_bar; toolbar; some_content; status_bar ]

19:52 <hcarty> adrien: Sounds very nice

19:52 <adrien> (the implementation uses a zipper, which is actually what xmonad does with windows)

19:53 <adrien> it's delayed by some other work for now however ;p

19:53 <hcarty> adrien: That's understandable - I have a rather huge backlog of PLplot tasks I'd like to accomplish

19:54 <adrien> ;-)

19:54 <adrien> I'm almost done with some exams here and this tiling support is something which is actually quite high priority for me so it shouldn't be too long before it's available

19:55 <adrien> (and I'll also merge support for React in lablgtk and some patches, and, and, and, and I'll port lablgtk to gtk3, and I'll add seamless Qt support, and EFL (E17)... :-) )

21:52 <Georgey> hi

21:52 <Georgey> are questions to polyml allowed in here?

22:06 <Georgey> http://pastebin.de/11893 how can I call the function inside of the let block?

22:33 <ski> Georgey : actually, PolyML is an implementation of SML, whose channel is #sml

22:34 <Georgey> mhkay

22:34 <ski> (SML and OCaml are sibling languages)

22:35 <Georgey> so you won't kick me if I ask here?