<adrien>
(what I hate is having to mix libraries and everything built with different toolchains)
sebz has joined #ocaml
dahc has quit [Remote host closed the connection]
yroeht has quit [Ping timeout: 240 seconds]
yroeht has joined #ocaml
sepp2k has joined #ocaml
avsm has joined #ocaml
dnolen has joined #ocaml
<f[x]>
overbld is all-in-one
<f[x]>
what different toolchains?
<adrien>
f[x]: for gtk stuff
<adrien>
f[x]: all-in-one, with other libraries? C libraries? mingw?
<f[x]>
not C libraries
sebz has quit [Quit: Computer has gone to sleep.]
<adrien>
I have a few changes to do that will take me a few minutes and then I'll be busy compiling on windows =/ (lablgtk stuff, with archimedes using natdynlink)
mjonsson has quit [Remote host closed the connection]
larhat1 has quit [Quit: Leaving.]
larhat has joined #ocaml
<thelema>
does anyone know when -modules was added to ocamldep argument list?
<adrien>
thelema: 2006 I think
<adrien>
thelema: git log -S-modules
<adrien>
looks like it
<adrien>
(the commit message doesn't mention ocamldep however)
<adrien>
actually, git log and look for '-modules'
iratsu has quit [Ping timeout: 250 seconds]
<adrien>
is there no way to get a lighter notation for 'fun x -> record.field <- x'? I wish I could use '(<-) record.field'
<thelema>
adrien: nope, gotta write it out. caml-light had an lvalue feature that made things like this easier, but ocaml didn't inherit it
<thelema>
adrien: thanks for the pointer
<adrien>
ok, I wouldn't have that much use for it, only a quick interface hopefully
<adrien>
and you're welcome =)
<thelema>
hmm, that commit is before 3.10, now why is someone having problems with [ocamlfind] ocamldep -module with ocaml 3.11.0...
<thelema>
maybe findlib is too old at 1.1.2 to support it
gildor has joined #ocaml
blinky- has joined #ocaml
avsm has quit [Read error: Operation timed out]
dnolen has quit [Quit: dnolen]
ankit9 has quit [Quit: Leaving]
rjsimmon has quit [Ping timeout: 252 seconds]
<adrien>
has oasis been tested a lot on windows?
rjsimmon has joined #ocaml
<adrien>
gildor: oasis hardcodes the use of C:\\Program\ Files ?
<adrien>
gildor: oasis hardcodes the use of C:\\Program\ Files\\Objective\ Caml ?
lopex has quit []
larhat has quit [Quit: Leaving.]
larhat has joined #ocaml
<adrien>
it seems it hardcodes it but is able to use it from somewhere else as long as it's in PATH
<sgnb>
adrien: isn't it `ocamlc -where`?
<sgnb>
(or anything else in `ocamlc -config`)
<adrien>
sgnb: I have C:/desk/ocaml/bin in PATH but I also needed to have everything in C:/Program\ Files/Objective\ Caml
<adrien>
and it hardcodes a few things for msvc too
<sgnb>
at worst, it should be configurable (by ./configure-like flag)
<adrien>
I believe it's an error that some path got hardcoded there, it seems to be only for some initial stuff
<adrien>
or maybe that was because I had a folder there even though it was empty
<adrien>
so, some things are hardcoded for msvc but only if not on Cygwin, I need to try cross-compilation from cygwin again, might be the best solution in the end when cross-compilation from linux is not possible
<gildor>
adrien: it has been tested on windows
<gildor>
adrien: it runs the same testsuite, which includes building most examples
<gildor>
adrien: but it remains probably less tested than on Linux
<gildor>
adrien: I don't think I hardcode any C:\\Program Files, at least it is not where I put my ocaml binaries
<adrien>
gildor: is this path part of some default sreach path?
<adrien>
search*
<gildor>
adrien: we use PATH env for search, AFAIR
<adrien>
and it's picking up gcc (under cmd.exe: no cygwin, no msys) but using msvc switches (/Fe, /L, /I)
<adrien>
hmm, weird, I'll have to try again
<f[x]>
ocaml hardcodes path to stdlib
<f[x]>
you can override it with OCAMLLIB
<adrien>
gildor: things have to be linked like: gcc foo.c -LFOO -IFOO -lcairo, using gcc -LFOO -IFOO -lcairo foo.c (note the position of foo.c) creates undefined references
<adrien>
unless --as-needed is available (iirc, ask NaCl :P ), the order of symbol resolution is from right-to-left so the element with the most dependencies on symbols must be leftmost
<adrien>
and using -mno-cygwin is now an error (the transition is going to be a lot of fun)
<gildor>
adrien: well I don't manage gcc with oasis, that something dealt at ocamlbuild level
<adrien>
and, and maybe that my issue is that I have the OCAMLLIB environment variable pointing to programfiles/*
<adrien>
gildor: not CFLAGS and LDFLAGS?
<gildor>
adrien: we are managing CCOpt and CCLib
<gildor>
(I follow ocamlc convention)
<adrien>
I have get_cairo_cflags in setup.ml, is it generated by oasis or by something only in cairo2?
<gildor>
something added by cairo2 I think
<adrien>
ah :-)
<adrien>
and ocamlfind is dependency of oasis?
<adrien>
I have to say I'm pretty scared right now: need lablgtk2, cairo2 and archimedes (which uses natdynlink) on windows
<gildor>
well ocamlfind is required to detect ocaml library installation
<thelema>
gildor: welcome back
yroeht has quit [Ping timeout: 246 seconds]
<gildor>
you can see if you can split lablgtk2 packages into a separate library, but I think you'll need cairo2 anyway
<gildor>
thelema: thx
<gildor>
adrien: but you'll have to submit a patch to upstream about that
<adrien>
I need the whole lablgtk2 actually (ok, maybe not librsvg)
<gildor>
adrien: do you need lablgtk2 for your own use or for archimedes ?
<thelema>
gildor: archimeded needs cairo which needs lablgtk2
<thelema>
(at least according to deps on oasis-db
<adrien>
archimedes needs cairo2 which can (or needs?) lablgtk2 to be precise
<adrien>
but I use archimedes in a lablgtk application anyway
<gildor>
so maybe there is something to split in cairo, don't know precisely if it is possible or nort
<gildor>
adrien: ok, so not a real big deal, you are just stuck with windows installation in general
* gildor
know this kind of pain
ttamttam has joined #ocaml
yroeht has joined #ocaml
rjsimmon has quit [Ping timeout: 252 seconds]
<adrien>
overbld's ocamlc -where is hardcoded to C:\Program Files\Objective Caml
* f[x]
can only repeat :
<f[x]>
<f[x]> ocaml hardcodes path to stdlib
<f[x]>
<f[x]> you can override it with OCAMLLIB
lopex has joined #ocaml
<adrien>
hah, sorry, I had missed your message
<adrien>
reminds me the time I was compiling mldonkey
larhat has quit [Quit: Leaving.]
<adrien>
ocamlfind's Makefiles won't run on windows without changing everything so I need to decide whether I'm going to use msys (bluh) or cygwin (bah)
<thelema>
all: should `make install` remove previous ocamlfind installation?
<hcarty>
thelema: Probably not
<hcarty>
thelema: But a "make uninstall" target is (always?) welcome
<thelema>
hcarty: this means that oasis and batteries are doing the wrong thing, and that odb needs to step up to uninstall a package before reinstalling
<hcarty>
thelema: I don't think oasis removes existing installs
<hcarty>
thelema: I've used it for a few libraries and I get the standard "this library already exists" error from ocamlfind if I try to "make install" twice without "make uninstall" between
<thelema>
hcarty: oops, you're right.
<thelema>
I was testing some packages in oasis-db/unstable to promote them to testing, and saw oasis configure
<thelema>
but the packages were actually using omake
<hcarty>
A user may want to install a library which exists in a system location, but install it somewhere else where they have permission.
<thelema>
in this case, make uninstall should fail to remove the system version
<gildor>
thelema: there is no check on the URL provided
<thelema>
gildor: yes, I'm asking for at least a simple one - that the URL ends in a filename that looks like a tarball
<gildor>
thelema: there is an ocaml setup.ml -uninstall target if needed, but it will uninstall only what has been installed
<thelema>
gildor: my mistake for libeling oasis for uninstall-on-install behavior
<hcarty>
thelema: It would fail, which would cause the entire make process to fail. If this uninstall is a prerequisite of the install target then they would be stuck.
<gildor>
thelema: there should be a check, but probably based on a real download and comparison with provided tarball AND/OR consistency of the tarball (is ocaml-archive able to expand it)
* gildor
put one more thing on the TODO list
<thelema>
hcarty: failing to uninstall shouldn't cause the entire make process to fail
<hcarty>
thelema: Doesn't make stop once an error result is returned?
<thelema>
:( ocaml-posix-resource provides the package posix_resource
<gildor>
thelema: wthe problem posix_resource ?
<thelema>
yes
<thelema>
although I guess the only counter-examples in my current ocamlfind db are type-conv and num-top
<thelema>
but mostly the problem is that odb package names aren't guessable from oasis package names
<thelema>
although bin_prot seems to follow the underscore convention
<thelema>
xdg-basedir vs. posix_resource
blinky- has quit [Quit: /quat]
avsm has joined #ocaml
<gildor>
thelema: - and _ are allowed, there is no convention
<thelema>
it might not hurt to have one.
bobry has quit [Quit: Leaving.]
<adrien>
time to try my new META file for lablgtk2 xP
<adrien>
oasis will compile using ocamlfind during compilation, right?
<hcarty>
adrien: Yes
<adrien>
gildor: on windows (in msys), [ ocamlfind query bigarray ] returns "\"foo\"", meaning that Sys.file_exists/is_directory will fail, do you think it would be good to automatically strip surrounding quotation marks when reading a program's stdin? that sounds a bit too much but on windows... or maybe an optional argument could help
<adrien>
hcarty: thanks, so I can't hack around it in the configure stage :P
<hcarty>
adrien: You may be able to work around it in the _oasis file
Boscop has quit [Quit: OutOfTimeException: Allocation of TimeFrame failed due to lack of time. Free up time by cancelling unimportant events.]
<adrien>
hcarty: well, I've fixed these issues
<adrien>
now, ocamlbuild tells me that it can't find some lib in the ocamlbuild -where directory and this error is actually scary
ikaros has joined #ocaml
<adrien>
think I'll have to give up with "error while reaading liblablgtk2.a: Sys_error(\"Invalid argument\")" =/
<_habnabit>
aaaaaaaaaargh. so apparently this system has a working libncurses and a broken libcurses. however, ocaml tries -lcurses before -lncurses
<_habnabit>
which wouldn't be an issue except that this is godi.
lopex has quit [Ping timeout: 252 seconds]
lopex has joined #ocaml
sepp2k has quit [Ping timeout: 258 seconds]
sepp2k has joined #ocaml
avsm has quit [Quit: Leaving.]
shachaf_ has joined #ocaml
<thelema>
data structure question:
george_z0rwell has joined #ocaml
zzz_` has joined #ocaml
edwin1 has joined #ocaml
edwin has quit [Disconnected by services]
edwin1 is now known as edwin
bitbckt_ has joined #ocaml
<thelema>
given an ordered sequence, I want to find the maximal element and the subsequences of elements before that element and after that element
<thelema>
repeatedly and efficiently.
wishi_ has joined #ocaml
<thelema>
I guess I want to turn the list into a kind of tree with the largest element as the root
emias_ has joined #ocaml
<thelema>
but I have to keep the order of the sequence
Anarchos has joined #ocaml
<thelema>
anyone know a data structure that'll do this well?
joewilliams_ has joined #ocaml
Cyanure has quit [Ping timeout: 252 seconds]
bobry has joined #ocaml
haelix_ has joined #ocaml
joewilliams_ is now known as joewilliams_away
Morphous has quit [*.net *.split]
Reaganomicon has quit [*.net *.split]
bitbckt has quit [*.net *.split]
emias has quit [*.net *.split]
wishi has quit [*.net *.split]
shachaf has quit [*.net *.split]
zzz_ has quit [*.net *.split]
haelix has quit [*.net *.split]
joewilliams has quit [*.net *.split]
bitbckt_ is now known as bitbckt
bitbckt has quit [Changing host]
bitbckt has joined #ocaml
emias_ is now known as emias
joewilliams_away is now known as joewilliams
Morphous has joined #ocaml
<flux>
thelema, you could maybe put nodes of a doubly-linked list into a tree..
<flux>
but
<flux>
it's difficult to compose such a data structure out of existing ones
edwin has quit [Remote host closed the connection]
metasyntax|work has quit [Quit: WeeChat [quit]]
ttamttam has quit [Ping timeout: 252 seconds]
ttamttam has joined #ocaml
sebz has joined #ocaml
ttamttam has quit [Quit: Leaving.]
avsm has joined #ocaml
Smerdyakov has joined #ocaml
shachaf_ is now known as shachaf
sebz has quit [Quit: Computer has gone to sleep.]
<_habnabit>
how is it that installing ocaml-sqlite3 can succeed, but then trying to use it gives an error 'ld: cannot find -lsqlite3' ?
<Smerdyakov>
Is there a separate dev package?
<taupin>
did you add -linkpkg to the compiler ?
<taupin>
+command
<_habnabit>
taupin, I added pkg_sqlite3 to my _tags. I thought it was supposed to do the rest for me.
<_habnabit>
taupin, ocamlbuild seems to add -linkpkg to the ocamlopt flags, though.
<taupin>
_habnabit: to be honest, I don't know how this works
<_habnabit>
hmm, odd. it only fails when I build a .native.
<_habnabit>
doing a .byte succeeds just fine.
<taupin>
when I use sqlite3, I do `ocamlfind ocamlc -thread -package sqlite3 -linkpkg` (or `ocamlfind ocamlopt` etc.)
<_habnabit>
of note is I'm using ocamlbuild.
<taupin>
noticed.
Smerdyakov has quit [Quit: Leaving]
Morphous has quit [Ping timeout: 264 seconds]
<avsm>
check your $(ocamlc -where)/ld.conf and make sure its got an entry for your sqlite lib directory. but it sounds like the linker cant find the actual sqlite dll, not the ocamlopt stubs
zsparks has quit [Ping timeout: 252 seconds]
<_habnabit>
actually found out that the issue is that -static is being added to the flags.
sepp2k has quit [Read error: Connection reset by peer]