Alpounet has quit [Read error: 104 (Connection reset by peer)]
Alpounet has joined #ocaml
<lpjhjdh>
so I'm trying to get lablgtkmathview installed and it's saying ocamlfind query gdome2 fails
<lpjhjdh>
I have gdome2 but I don't see an entry in ld.conf and there is no META file
<lpjhjdh>
I've got gdome2 v 0.8.1
<palomer>
oh lord
Submarine has quit [Read error: 110 (Connection timed out)]
Pimm has quit [Read error: 110 (Connection timed out)]
Mr_Awesome_ has quit [Read error: 110 (Connection timed out)]
infoe_ has joined #ocaml
eldragon has joined #ocaml
lpjhjdh has left #ocaml []
julm has quit [Remote closed the connection]
julm has joined #ocaml
lanaer_ is now known as lanaer
julm___ has joined #ocaml
julm has quit [Read error: 104 (Connection reset by peer)]
julm___ is now known as julm
struktured has quit [Read error: 110 (Connection timed out)]
tmaedaZ0 is now known as tmaedaZ
ikaros_ has quit ["Leave the magic to Houdini"]
bogen has joined #ocaml
<bogen>
hello
ofaurax has quit ["Leaving"]
<bogen>
well, I have a coworker who is very interested in ocaml, but he is a big proponent of being able to connect with a debugger to a running app after it has launched
<bogen>
and won't really consider ocaml's use other wise
<bogen>
I've played around with it some and have not been able to figure out if one can do that
<bogen>
without starting the debugger first
r0bby_ is now known as r0bby
<bogen>
without his backing I'll have a hard time promoting it's use
joewilliams has quit [Remote closed the connection]
tmaedaZ is now known as tmaedaZ0
Alpounet has quit ["``Do what you think you can't do.'' -- E. Roosevelt"]
tmaedaZ0 is now known as tmaedaZ
<thelema>
bogen: well, there's ocamlviz, but that's not quite a full gdb
<thelema>
is there any non-VM language that one can attack debuggers to at runtime?
<bogen>
ocamlviz is graphical
<bogen>
native code languages that you can attach a debugger at runtime?
<bogen>
gdb with natively compiled C/C++, SwiftX with natively compiled Forth
<bogen>
natively compiled is what you mean by non-VM?
<bogen>
personally I don't see as much of need for a full gdb type debugger with a language like ocaml
<bogen>
one is not going to run into all the long list of problems one has with C
<thelema>
you can use gdb with ocaml
<thelema>
I didn't realize gdb could attach to c at runtime
<thelema>
It seems that it should be as easy to attach it to a natively compiled ocaml program as C
<bogen>
you can? and you to step through machine code? or ocaml code?
<bogen>
I'll give it a try
<thelema>
ocaml code
<bogen>
hmm
<bogen>
cool
<bogen>
it works
<thelema>
Well, now you can sell ocaml properly.
<bogen>
yeah
<bogen>
a couple things I need to figure out first
<bogen>
as far as symbol naming
<bogen>
etc
<bogen>
thanks thelema, it did not even occur to me that gdb would support ocaml
<bogen>
for some reason....
<thelema>
there's an ocamldebug program for bytecode, but yes, ocaml uses gdb for native debugging.
<bogen>
ok, well, I'm going to have to figure out how to set the breakpoints in gdb
<bogen>
I know how to do it with c code
<bogen>
but it is not acting the same with ocaml code
<thelema>
sorry, not a gdb expert
<bogen>
no problem
<bogen>
I just googled a bit for it, others have having similar problems
* bogen
is not a gdb expert either, just knows enough to be dangerous...
<mfp>
thelema: why did you remove extlibcompat.ml?
<thelema>
it's in Batteries.ml
<olegfink>
has anyone played with metagene or can tell me something about ml->c++ template language translation? :-)
<mfp>
wasn't the plan to allow open Extlibcompat for source-level compat?
<mfp>
ah OK
<thelema>
I think
<olegfink>
(and don't ask why I need that)
<thelema>
yes, Batteries.Extlib
<gareth_0>
hi, is there a defacto standard for ocaml gui?
<thelema>
gareth_0: gtk
<thelema>
olegfink: sorry - no idea.
<mfp>
thelema: README needs to be updated
<gareth_0>
thelema, should i use lablgtk1 or lablgtk2?
<EliasAmaral>
i think that mldonkey is the most used ocaml graphical app (or one of most useds), and seems to have a community
<EliasAmaral>
i think it uses wxwidgets, that is compatible with gtk
<thelema>
gareth_0: 2
<thelema>
EliasAmaral: unison is up there, and it uses lablgtk2
gareth_0 has quit [Remote closed the connection]
_unK has joined #ocaml
<thelema>
s/up there/commonly used/
<EliasAmaral>
don't know about them. but i really favour gtk
<EliasAmaral>
i am still confused about what is exactly a peanut gallery, i am reading wiktionary
Pimm has joined #ocaml
<EliasAmaral>
maybe it's just a idiom
<palomer>
it's an idiom, for sure
gareth_0 has joined #ocaml
<EliasAmaral>
palomer, i would love to use your library. in fact i would love to like some libraries out there, that use lots of objects that doesn't really fits on my head
<thelema>
it's definitely an idiom, it's a pejorative term for one's audience
<EliasAmaral>
i seem actually isolated with my functional style
<EliasAmaral>
ops.
<EliasAmaral>
thelema,
<mfp>
is this true? -> Iceweasel can't find the server at forge.ocamlcore.org. ... "It's just you. http://forge.ocamlcore.org is up."
<EliasAmaral>
(i was talking about batteries; never saw it but i will look for sure later)
<palomer>
I'm about to release another version in the coming weeks
<mfp>
thelema: have you replaced ocamlinit already? I remember I found a reasonable solution, but not what it was or who was going to implemented it :)
<thelema>
yes, I remember you coming up with a solution that sounded good... Where's those #ocaml logs?
<EliasAmaral>
hmmmm. i was thinking of using ocamllex/ocamlyacc to build a minilanguage, i could use oset instead? (I don't know how to use ocamllex / ocamlyacc yet)
<EliasAmaral>
palomer, you have setup.. youtube videos for your examples?
<thelema>
hmm, too much findlib noise...
<EliasAmaral>
palomer, maybe you should write down in the page (with proper highlighting) the snippet code, i have no flash here
<mfp>
wow the new lwt.top shows keywork completions as you type o.O
<EliasAmaral>
palomer, in fact you have a nice manual but.. do you have a online browseable api?
<mfp>
it takes over rlwrap and doesn't let me use vim bindings, though :-(
<thelema>
mfp: yes, I'd really like to get the lwt.top in batteries
<palomer>
EliasAmaral, you can pretty much stick to the lone function SEditable.input
<palomer>
EliasAmaral, and then use the "with seditable" directive
<palomer>
(or grammar, if you want to get fancy)
<thelema>
mfp: but I have the feeling that'd require using all of lwt as a dependency
<mfp>
I think so
<EliasAmaral>
palomer, i am not with ocaml here, i would just like to peek the library interface
<thelema>
mfp: this isn't necessarily bad, just difficult
<mfp>
it'd be problematic since Lwt is moving very fast
<palomer>
EliasAmaral, there is no real interface, really
<gildor>
mfp, thelema: what is the problem with the DNS of forge.ocamlcore.org ?
<EliasAmaral>
palomer, Hmmm what do you mean? the semantics is always-changing?
<mfp>
gildor: must be a problem with my ISP's DNS server
<gildor>
mfp: I seem to also have this problem
<palomer>
EliasAmaral, nope, but there is only 1 function and 3 methods you can use. the 3 methods are generated by camlp4
<mfp>
oh, then some European DNS server :?
<thelema>
gildor: no problem from my end - forge.ocamlcore.org works fine
<palomer>
EliasAmaral, as for the internal API, it's a mess
<EliasAmaral>
palomer, anyway, a documentation on this stuff would be useful for people like me -- a tiny interface is actually better than a larger one, if it does the job
<palomer>
EliasAmaral, the whole interface is 1 function.
<thelema>
they're the labeled versions of the libraries
<Camarade_Tux>
lablgtk makes *extensive* use of labels and polymorphic variants
<mfp>
maybe because they were originally developed for olabl?
<gareth_0>
hmm thx Camarade_Tux
<Camarade_Tux>
that's how *I* understand it but currently all labl* libraries are from the same guy
<mfp>
at some point labels existed only in an OCaml branch named OLabl IIRC
<mfp>
which was merged into OCaml around 2.x
<EliasAmaral>
is this true that all ocaml programmers pick a % of the language that they like and use and the rest they treat as deprecated? i mean this happens in every programming language, but on some this is more noticeable. (c++ being the obvious example)
<EliasAmaral>
(I myself don't like objects very much.. there are some people that don't like mutable data.. ok i don't like mutable data sometimes.. and so on)
<thelema>
EliasAmaral: do you use the entire english language when you speak/write (or whatever your native language is)?
<thelema>
I'd argue that you don't, you use a "complete enough" subset of it
<thelema>
it's the same thing in programming
<EliasAmaral>
thelema, heh. C folks often says that their language is optimal in this aspect. it's incredible for me, the language is so small
<EliasAmaral>
the best c programmers can argably use all and every feature of C and C preprocessor
<EliasAmaral>
which is not very much. ocaml programmers shouldn't
* Camarade_Tux
goes back to his lablwebkit code :P
<EliasAmaral>
lablwebkit?
<thelema>
sure. And the best writers will use a large variety of the language they write in.
<EliasAmaral>
webkit-gtk renders it into html?
<EliasAmaral>
no, it is a html rendering engine
<EliasAmaral>
now i am confused
<Camarade_Tux>
hahaha, bindings to webkit-gtk
* thelema
chukles at the "OCaml is broken!" post on caml-list
<Camarade_Tux>
I did that too this morning actually
<EliasAmaral>
thelema, print_bool. I wonder why this isn't in pervasives yet. and many other things, like identity.. OCaml isn't a very "stable" language..
<thelema>
mfp: was our conversion strategy from extlib really "open Batteries.Extlib"?
<Camarade_Tux>
actually it reminds me of someone who posted on this channel: he had tried ocaml but said ocaml automatically use the appropriate packages when building, and left
ygrek has quit [Remote closed the connection]
<mfp>
thelema: I thought it was just open Extlibcompat
<gildor>
mfp, thelema: sdns1.ovh.net back online
<mfp>
which is more or less equiv
<gildor>
works fine now
<thelema>
I seem to have put that in batteries. I'll pull it back out. I like extlibcompat better, somehow
<EliasAmaral>
i wrote an ad-hoc input_file many times. with Buffer.t mainly. and it's somewhat error-prone
<thelema>
gildor: ok, still working for me.
<gildor>
mfp, thelema: sorry the problem comes from OVH
<thelema>
but with the way DNS caches, that shouldn't be a surprise
<mfp>
yup, working again for me
<gildor>
mfp, thelema: I was just filling an incident report when it comes back to life
<thelema>
gildor: no problem.
<gildor>
thelema: I was not using cache to test things: dig @sdns1.ovh.net forge.ocamlcore.org returned a timeout
<gildor>
thelema: now it is answering correctly
<thelema>
gildor: n/m. Thanks for fixing it.
<thelema>
hmm, when I run that dig, I get a "no servers could be reached" message
<gildor>
thelema: but it only last less than 1 hour, almost undetectable for 80% of users, some which have flushed their DNS cache, get strange answer
<gildor>
thelema: I think something has hit sdns1.ovh.net and things get back on their way, but it could take time
<thelema>
ah, it worked now.
<gildor>
still quite unstable
<Camarade_Tux>
I can't git-push ='(
<thelema>
??? unzip not in batteries?
<mfp>
what?
<gildor>
Camarade_Tux: DNS problem at OVH, it will come back to normal (but it is not in my hand)
<thelema>
mfp: the unzip module from extlib doesn't have an equivalent in batteries
<mfp>
uh never used Extlib's Unzip
<thelema>
apparently few have, as there's never been any complaint
<thelema>
I'm only noticing it as I build extlibcompat.ml
<mfp>
it inflates IO.input hmm
<mfp>
I'm sure there's something in batteries that does that
<thelema>
we have gzip
<mfp>
only gzip? I thought we also had deflate
<thelema>
apparently not
<mfp>
Unzip's got its own deflate implementation, it doesn't use Zlib (!)
<thelema>
yup
<thelema>
no external dependencies
<mfp>
what are we going to do with camlzip/zip and (full) Batteries' gzip support?
<thelema>
I want to drop them for this RC
<mfp>
you removed them a while ago, didn't you?
<EliasAmaral>
.. if you are colleting a wishlist, bz2 might be useful too (maybe with a optional compile-time configure..)
<mfp>
what about post-RC?
<thelema>
post rc, I want all compression/decompression following the Compressors interface
<thelema>
err, Codec
<thelema>
If we can get Xavier's cryptolib in there too, I'd be happy
<Camarade_Tux>
gildor: yeah, I saw, was only joking ;-) and I'm used not to git-push during long periods now =)
<mfp>
thelema: sounds doable. So cryptokit & camlzip would become (hard) dependencies?
<EliasAmaral>
somebody will consider doing high-level network code in batteries? many language nowdays comes with their own http mini client..
<mfp>
making them optional represents some devel cost for us, and they're available everywhere (Fedora, Debian, GODI...)
<thelema>
yes. They're very stable modules, and I'm an admin for camlzip, and have sent in patches for cryptokit, so may be able to make them easier for batteries to use
<thelema>
probably by adding findlib support to camlzip peopre
<thelema>
*proper
<mfp>
camlzip doesn't have a META?
<thelema>
no, each distro that packages it makes one.
<mfp>
so the one I've been using was writing by Debian uh
<thelema>
which is where the findlib naming problem scame from
<mfp>
well, no diff in practice besides the name pb then
<thelema>
Camarade_Tux: yes, I'm now an admin for that project, I just need to check in the patch I made.
<thelema>
and hope that it doesn't fail too badly for someone.
<gildor>
Camarade_Tux, thelema: hope you copy it from Debian
<gildor>
(the META i mean)
<thelema>
gildor: IIRC, that was the consensus when rwmjones and someone else tried to figure out the best naming
<bogen>
thelema: regarding GDB and OCaml, it works minimally at the moment, but there is a lot of work to be done. (I'll try to help out in that regard as my time permits)
<thelema>
bogen: great. Ocaml can definitely use better debugging tools.
<gildor>
thelema: ok great
<gildor>
thelema: will you also use ocamlfind to install the library ?
<thelema>
I'll have to also add a makefile target for installation
<thelema>
I think this is pretty straightforward, though
<bogen>
thelema: I'll need to research if GDB works with other cactus stack languages. Well, hmm... GCC allows cactus stacks, in a limited manner.
<bogen>
(well, more to the point, GCC allows for nested functions and closures)
<gildor>
ocamlfind install zip META *.cmi *.cma $(wildcard *.cmxa) ...
<thelema>
is install-findlib an appropriate makefile target, or should it use findlib on a default install?
<gildor>
does install target already exist ?
<thelema>
yes
<gildor>
install-findlib seems fine, so
<thelema>
and installopt
<thelema>
but I'll install everything
<thelema>
on findlib install
<gildor>
if you use the wildcard trick, yes you can
<thelema>
I just wondered if there was a standard already - I know GODI has requirements for makefiles
<gildor>
there is a GODIVA packaging policy
ikaros has quit ["Leave the magic to Houdini"]
<thelema>
is there any reason not to install *.mli?
<gildor>
it is better to install *.mli
<gildor>
except if *.mli is younger than matching .cmi
<gildor>
in this case, this is not good
<thelema>
ok, *.a have to be installed, no problem.
<thelema>
libcamlzip.a and zip.a
<thelema>
but what about this dllcamlzip.so?
<gildor>
yes you need to install it
BiDOrD has quit [Read error: 110 (Connection timed out)]
* thelema
copies the code from the install target, as it doesn't seem to go into the ocamlfind dir
fx___ has quit [Read error: 110 (Connection timed out)]
<thelema>
does it?
jcaose_ has joined #ocaml
BiDOrD has joined #ocaml
jcaose has quit [Read error: 110 (Connection timed out)]
gareth_000 has joined #ocaml
gareth_0 has quit [Remote closed the connection]
gareth_000 has quit [Remote closed the connection]
gareth_000 has joined #ocaml
_zack has joined #ocaml
bogen has quit [Read error: 110 (Connection timed out)]
bogen has joined #ocaml
TaXules_ has quit [Remote closed the connection]
TaXules has joined #ocaml
johnnowak has joined #ocaml
ttamttam has joined #ocaml
hugin_ is now known as hugin
_zack has quit ["Leaving."]
caligula has joined #ocaml
ttamttam has quit ["Leaving."]
Pimm has quit [Read error: 110 (Connection timed out)]
Amorphous has quit [Read error: 110 (Connection timed out)]
Amorphous has joined #ocaml
<thelema>
can someone check out svn of camlzip and verify that the install-findlib target works for them?
<thelema>
svn://scm.ocamlcore.org/svnroot/camlzip
<Camarade_Tux>
any way to check without actually installing it?
<Camarade_Tux>
hmmm, -destdir?
<thelema>
:( I don't know of any
<thelema>
you'd have do kludge the makefile at least
<Camarade_Tux>
seems to work (I used "ocamlfind install -destdir")
<thelema>
thank you very much
<thelema>
at least I didn't make a huge mistake in my checkin
* thelema
isn't a SVN expert
bogen has quit [Read error: 110 (Connection timed out)]
<Camarade_Tux>
np :-)
bogen has joined #ocaml
<Camarade_Tux>
it's been a long time I used svn, I was happy to remember "svn co" ;-)
johnnowak has quit []
<thelema>
is there a real reason for Batteries to depend on stdlib's map implementation and go through hoops to get at the internals?
<thelema>
I'm thinking of having a Batteries.Trees package with various types of trees implemented (splay, Red-black, AVL, OCaml's modified AVL, etc) and allow set/map to work over them...
<thelema>
kind of like the relationship between Camomile's ISet/IMap and AvlTree
<thelema>
TODO, I think.
tmaeda0 is now known as tmaedaZ
<mfp>
thelema: binary compatibility
<thelema>
huh? is there ever binary compatibility in the ocaml world?
<mfp>
in this case
<thelema>
what's the use case?
<mfp>
you can use ExtSet functions with sets created by 3rd party libs that use the std Set
<thelema>
hmmm...
struktured has joined #ocaml
<thelema>
Is that a big deal? You can always just use regular set functions on those sets.
<mfp>
but not the extra ones
<mfp>
so you'd have to change your libs to use AAA's sets
<thelema>
There's all sorts of functions you can't use
<thelema>
You'd have to do that to get the functions that AAA adds, but I can't say we've got anything particularly compelling, except maybe conversion to enum
<mfp>
mostly enum, print, keys, values, filter, and ExceptionLess
<mfp>
how would you implement Batteries.Trees's sets/maps? by functorizing over a tree?
<thelema>
keys/values is very nice for maps...
<thelema>
look at AvlTree with ISet
<thelema>
now batAvlTree, batISet
<thelema>
but there'd need to be a functorized version too to allow pluggability
<thelema>
There'd be one Set module with all the set functions, created by including the default functor parameter
ski_ has quit ["Lost terminal"]
<thelema>
and a submodule that (I guess would have to have a duplicate of all the code :( ) that'd be functorized to use whatever tree one wanted
<thelema>
also, the impacts of the 3.12 changes on batteries' design haven't been thought out.
<thelema>
But that's another TODO
bogen has quit [Read error: 113 (No route to host)]
bogen has joined #ocaml
ttamttam has joined #ocaml
gareth_0 has joined #ocaml
<thelema>
mfp: I'm planning on releasing batteries as 0.9.1 per http://semver.org/
ski_ has joined #ocaml
<thelema>
hmm, we lost some pretty in the rebuild of aaa's docs
<thelema>
I'm going to try to get it back
gareth_000 has quit [Remote closed the connection]
alp_ has joined #ocaml
ulfdoz has quit [Read error: 110 (Connection timed out)]
Narrenschiff has joined #ocaml
Alpounet has quit [Connection timed out]
ikaros has joined #ocaml
jcaose_ has quit [Read error: 110 (Connection timed out)]
ttamttam has quit ["Leaving."]
<gareth_0>
does anyone know how to get lablgtk2 working with ocamlbuild?
<thelema>
gareth_0: step 1: install lablgtk2 with findlib
<gareth_0>
"ocamlbuild -libs lablgtk -cflags -I,+lablgtk2 -lflags -I,+lablgtk2 guitest.byte" <-- builds but segfaults on running
<thelema>
step 2: use findlib-aware ocamlbuild
<thelema>
err, findlib-aware myocamlbuild.ml
<gareth_0>
i already have lablgtk2 installed
<gareth_0>
i dont have findlib
<gareth_0>
compiling using ocamlc works as expected
<Camarade_Tux>
gareth_0: run "gdb yourapp.native", then in gdb, type "run", when it segfaults, type "bt", you can quit with Ctrl+D twice in a row
<Camarade_Tux>
gareth_0: hmmmm, looks like my problem a few hours ago :P
<Camarade_Tux>
gareth_0: run ocamlbuild with -classic-display and pastebin the commands it run
ikaros has quit [Read error: 104 (Connection reset by peer)]
<Camarade_Tux>
but the point is you have to link against gtkInit.cmx, and you probably don't (surprising how same issues come out at the same time :p )
<Camarade_Tux>
gareth_0: also, why no ocamlfind?
<gareth_0>
thanks Camarade_Tux, gdb output is here: http://pastebin.com/m7b77bb51 , ocamlbuild with -classic-display doesnt print anything (not even the usual line)
<Camarade_Tux>
gareth_0: 'rm -r _build'?
<gareth_0>
that did it
<gareth_0>
i dont know what ocamlfind is or why i dont have it
<Camarade_Tux>
gareth_0: on which platform are you? how did you install ocaml? any specific constraint?
ulfdoz has joined #ocaml
<gareth_0>
im on ubuntu 9.10 (uname -r is "2.6.31-16-generic"), i installed ocaml from synaptic, constraints on what ?
<Camarade_Tux>
and ocamlfind is a way to ease compilation and ocaml package management
<Camarade_Tux>
constraint on where you're going to run the app, but if you're on ubuntu I guess you probably don't have many
<Camarade_Tux>
thelema: aren't you running ubuntu?
<thelema>
yes
<thelema>
gareth_0: sudo apt-get install findlib
<thelema>
gareth_0: if you can share your code, we may be able to help - the problem doesn't seem to be in compilation.
<thelema>
I admit that lablgtk shouldn't segfault, but apparently it does.
<Camarade_Tux>
thelema: it's in the C and I think the gtk error messages are exactly the same I got when I was using ld-ocaml without linking against gtkInit.cmx
<thelema>
yup, you're not compiling with gtkInit.cmo
<thelema>
add the line [let _ = GtkMain.Main.init ()] to the beginning of your source
<thelema>
after the [open]s
<gareth_0>
wow, works
<gareth_0>
thxx!
<Camarade_Tux>
\o/
<thelema>
The gtkInit.cmo thing is a bit of a hack.
<thelema>
My opinion is that as much as possible should be in your source file, and as little as possible on the command line to compile
<thelema>
which is one major objection I have to camlp4
<Camarade_Tux>
ocamlfind helps however: I have "pkg_mikmatch_pcre, syntax_camlp4o" in my _tags and that's it
<gareth_0>
(i think it should just work without writing it anywhere...)
<Camarade_Tux>
(we should all be rich and happy without doing anything ;-) )
<gareth_0>
(what i mean is, the information is already on my system, isnt it, so it should be intelligent and find it)
<thelema>
ocamlfind suffers from the same problem - putting the information on dependencies outside the file. ocamlbuild tags should be able to be part of the file
<thelema>
a magic comment if nothing else is possible.
<thelema>
I admit there are reasons to use project-wide tags, and having those in _tags is ok.
<thelema>
But especially camlp4, which changes the meaning of the contents, should be specified at the beginning, so you know what context the following will be evaluated in
<Camarade_Tux>
I'm not really for all-automatic, often fails
<Camarade_Tux>
as for _tags, three easy lines sounds ok
<thelema>
it's not the size of tags, it's that tags aren't possible to put in files, so the _tags file isn't even needed in the first place
Pimm has joined #ocaml
<gareth_0>
all automatic relies upon convention, which needs to be specified early on before lots of people develop their own...
<thelema>
Well, OCaml is still a small language.
mjsor has joined #ocaml
<Camarade_Tux>
I see how one could want something like in python but what ocaml has right now is perfectly fine for me and I don't fancy python's way at all
<thelema>
python's way of specifying dependencies?
<thelema>
I don't know python that well - I do appreciate Perl's dependency handling
<Camarade_Tux>
afaik, "import" serves that purpose
<thelema>
you don't like specifying dependencies at the top of your file?
<Camarade_Tux>
I know python's "import" because it's at the top of files so I read it before giving up on trying to understand python code :P (I can't read python, doesn't make sense to me)
<thelema>
s/file/source/
<Camarade_Tux>
thelema: no, not that I don't like:
<Camarade_Tux>
I see how nice it can be but I really don't mind
<thelema>
you don't mind specifying dependencies on command line or in _tags
<thelema>
or in Makefile
<Camarade_Tux>
command-line is ok for me but _tags is nicer (much more concise, readable and editable)
* thelema
likes self-contained pieces, and when you have even tiny pieces of information external (in tags, needed on command line, etc) you lose that
<thelema>
I wouldn't mind it if the tags could be put in the header of the file
<thelema>
especially if most myocamlbuild.ml tag extensions could be system-wide, installed with findlib/etc.
<Camarade_Tux>
I wouldn't either but I find what we have currently is ok, actually with some sed-magic you could emulate that
bzzbzz has quit ["leaving"]
<thelema>
but then you'd need sed-magic in your build system
<thelema>
another external piece of information
<thelema>
if the sed-magic was built in to ocamlbuild, that'd work for me.
bzzbzz has joined #ocaml
<Camarade_Tux>
olegfink: btw, short tutorials quickly grow and limiting a tutorial to reading is pretty hard ;p