ocamlbuild is VERY nifty

ocamlbuild is VERY nifty

[Joel Reymont]

I asked around before settling on a build system since I'm new to  
OCaml. I looked at OMake, OCamlMakefile and found them complex and  
confusing. Then William Neuman pointed me to ocamlbuild and Nicolas  
Pouillard helped me hook it up to external libraries via a plugin.

Deep into my first project I confirm that ocamlbuild is VERY nifty. I  
can automatically rebuild when any parts of my project change and  
then run my unit tests.

All I need to do is:

ocamlbuild test.byte --

*** my project ***

      189 easy.mli
      282 easy_lexer.mll
      420 easy_parser.mly
       17 myocamlbuild.ml
       14 parser_util.ml
       30 symtab.ml
        8 test.ml
      284 parser_test.ml
     1355 total

*** test.ml ***

open OUnit

let suite = "Morpher" >:::
   [Parser_test.suite]

let _ =
   run_test_tt_main suite

*** myocamlbuild.ml ***

open Ocamlbuild_plugin;;
open Command;;

let packages = "ounit" (* "pkg1,pkg2,..." *);;

let ocamlfind cmd =
S[A"ocamlfind"; A cmd; A"-package"; A packages];;

flag ["ocaml"; "link"] (A"-linkpkg");;

dispatch begin function
| After_options ->
      Options.ocamlc := ocamlfind "ocamlc";
      Options.ocamlopt := ocamlfind "ocamlopt";
| _ -> ()
end

--
http://wagerlabs.com/

Re: [Caml-list] beta-test of OCaml 3.10.0

[Jon Harrop]

On Wednesday 07 March 2007 15:53, Joel Reymont wrote:
> Deep into my first project I confirm that ocamlbuild is VERY nifty.

Excellent news. Make just can't hack it when things get non-trivial and I'm 
bored of recompiling everything all the time, even if ocamlopt is a million 
times faster than gcc. :-)

Here are my results (timings are old vs new) for my 2.2GHz AMD64x2:

OCaml itself still doesn't parallel build but serial builds in around 4mins.

Ray tracer builds fine: 3.963s vs 4.000s run time

n-th nearest neighbour example from OFS builds and runs fine: 8.142s vs 8.320s

Smoke builds fine and runs fine: 12.39s vs 11.56s build time, 68.5fps vs 
72.5fps running speed

Presenta builds fine: 39.354s vs 39.524s

I've only had one problem and I'm not sure if that is due to the new version 
of ocamlopt: Presenta screws up font rendering when using the experimental 
stencil buffer code instead of the usual GLU tesselator code.

So my conclusion is that 3.10 looks fine. Compile times are slightly faster in 
some cases. Running times are basically indifferent.

-- 
Dr Jon D Harrop, Flying Frog Consultancy Ltd.
OCaml for Scientists
http://www.ffconsultancy.com/products/ocaml_for_scientists

Re: [Caml-list] ocamlbuild is VERY nifty

[Aleksey Nogin]

On 07.03.2007 07:53, Joel Reymont wrote:

> I asked around before settling on a build system since I'm new to OCaml. 
> I looked at OMake, OCamlMakefile and found them complex and confusing.

Well, it should have been equally easy to use OMake in this case - while 
it is true that OMake is complex, the complexity is only there to help 
you out when you need it; you can safely ignore it until then.

If I understand what you are doing here correctly, it would have been 
sufficient to just use the following OMakefile:

---------
# Optional - use bytecode compilation even when ocamlopt is available
NATIVE_ENABLED = false
BYTE_ENABLED = true

USE_OCAMLFIND = true
OCAMLPACKS[] = ounit

.DEFAULT: $(OCamlProgram test, test easy easy_lexer easy_parser 
parser_util symtab parser_test)
-------

and then simply running "omake" would do what you want.

> Then William Neuman pointed me to ocamlbuild and Nicolas Pouillard 
> helped me hook it up to external libraries via a plugin.
> 
> Deep into my first project I confirm that ocamlbuild is VERY nifty. I 
> can automatically rebuild when any parts of my project change and then 
> run my unit tests.
> 
> All I need to do is:
> 
> ocamlbuild test.byte --
> 
> *** my project ***
> 
>      189 easy.mli
>      282 easy_lexer.mll
>      420 easy_parser.mly
>       17 myocamlbuild.ml
>       14 parser_util.ml
>       30 symtab.ml
>        8 test.ml
>      284 parser_test.ml
>     1355 total
> 
> *** test.ml ***
> 
> open OUnit
> 
> let suite = "Morpher" >:::
>   [Parser_test.suite]
> 
> let _ =
>   run_test_tt_main suite
> 
> *** myocamlbuild.ml ***
> 
> open Ocamlbuild_plugin;;
> open Command;;
> 
> let packages = "ounit" (* "pkg1,pkg2,..." *);;
> 
> let ocamlfind cmd =
> S[A"ocamlfind"; A cmd; A"-package"; A packages];;
> 
> flag ["ocaml"; "link"] (A"-linkpkg");;
> 
> dispatch begin function
> | After_options ->
>      Options.ocamlc := ocamlfind "ocamlc";
>      Options.ocamlopt := ocamlfind "ocamlopt";
> | _ -> ()
> end
> 
> -- 
> http://wagerlabs.com/
> 
> 
> 
> 
> 
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
> 

Re: [Caml-list] ocamlbuild is VERY nifty

[Joel Reymont]

Aleksey,

I'm trying to build my project as you suggested but I get this error

*** omake error:
    File /usr/local/lib/omake/build/OCaml.om: line 1027, characters 8-53
    Do not know how to build "easy.cmo" required for "test.run"

I only have easy.mli and no easy.ml, should I exclude easy from the  
build list below?

	Thanks, Joel

On Mar 13, 2007, at 6:38 PM, Aleksey Nogin wrote:

> # Optional - use bytecode compilation even when ocamlopt is available
> NATIVE_ENABLED = false
> BYTE_ENABLED = true
>
> USE_OCAMLFIND = true
> OCAMLPACKS[] = ounit
>
> .DEFAULT: $(OCamlProgram test, test easy easy_lexer easy_parser  
> parser_util symtab parser_test)

--
http://wagerlabs.com/

Re: [Caml-list] ocamlbuild is VERY nifty

[Aleksey Nogin]

Joel,

I think this was already brought up on the Caml list - it is the basic 
OCaml assumption, I believe, that you should not have an .mli without 
.ml. OMake makes the same assumption. If you do need the Easy module, 
you should have easy.ml. If your easy.mli only defines new types and/or 
signatures, you should be able to simply rename easy.mli into easy.ml.

Hope this helps.

Aleksey

On 16.03.2007 06:46, Joel Reymont wrote:

> Aleksey,
> 
> I'm trying to build my project as you suggested but I get this error
> 
> *** omake error:
>    File /usr/local/lib/omake/build/OCaml.om: line 1027, characters 8-53
>    Do not know how to build "easy.cmo" required for "test.run"
> 
> I only have easy.mli and no easy.ml, should I exclude easy from the 
> build list below?
> 
>     Thanks, Joel
> 
> On Mar 13, 2007, at 6:38 PM, Aleksey Nogin wrote:
> 
>> # Optional - use bytecode compilation even when ocamlopt is available
>> NATIVE_ENABLED = false
>> BYTE_ENABLED = true
>>
>> USE_OCAMLFIND = true
>> OCAMLPACKS[] = ounit
>>
>> .DEFAULT: $(OCamlProgram test, test easy easy_lexer easy_parser 
>> parser_util symtab parser_test)
> 
> -- 
> http://wagerlabs.com/
> 
> 
> 
> 
> 

Re: [Caml-list] beta-test of OCaml 3.10.0

[Oliver Bandel]

On Wed, Mar 07, 2007 at 02:10:38PM +0100, Xavier Leroy wrote:
[...]
> - Unix: new function Unix.isatty.
[...]

Yes, that's nice.
I waited for this function for a longer time; is good for a lot of things.
For I tool I just write, it would be good to have this function
available. :)

When will OCaml 3.10 be available as a well tested, stable release?

Ciao,
   Oliver

Re: [Caml-list] beta-test of OCaml 3.10.0

[Nicolas Pouillard]

On 3/17/07, Peter Halacsy <peter@halacsy.com> wrote:
>
>
> On Mar 7, 2007, at 2:10 PM, Xavier Leroy wrote:
>
>
> - A new tool: ocamlbuild, a compilation manager for OCaml projects.
> Hi,
> I've downloaded and succesfuly compiled and installed 3.10 (on Intel MacBook
> Pro). Ocamlbuild is a great tool!
>
> I only wonder how I can build a cmxa file from several ml files. Should I
> define a custom target or something like that?

Just write a .mllib file with the contents of your library.

$ cat foo.mllib
Bar
Baz

$ ocamlbuild foo.cma foo.cmxa

If you have a main module (Foo) and want a library from it, it will
make a library with the sorted transitive closure from Foo.

$ ocamlbuild foo.cma foo.cmxa

-- 
Nicolas Pouillard

Re: [Caml-list] beta-test of OCaml 3.10.0

[Peter Halacsy]

[-- Attachment #1: Type: text/plain, Size: 371 bytes --]


On Mar 7, 2007, at 2:10 PM, Xavier Leroy wrote:

> - A new tool: ocamlbuild, a compilation manager for OCaml projects.

Hi,
I've downloaded and succesfuly compiled and installed 3.10 (on Intel  
MacBook Pro). Ocamlbuild is a great tool!

I only wonder how I can build a cmxa file from several ml files.  
Should I define a custom target or something like that?

peter



[-- Attachment #2: Type: text/html, Size: 928 bytes --]

Re: [Caml-list] beta-test of OCaml 3.10.0

[Olivier Andrieu]

On 3/16/07, Mike Lin <mikelin@mit.edu> wrote:
> On 3/7/07, Xavier Leroy <Xavier.Leroy@inria.fr> wrote:
> >
> > - Much work on Camlp4, including a new API.
>
>
> Has this broken a lot of stuff?? e.g. ocamlgsl has a camlp4 extension inside
> it and will not compile under 3.10 ('Unbound value Quotation.add'). what are
> we to do with existing extensions?

I'll be rewriting this part soon.

Re: [Caml-list] beta-test of OCaml 3.10.0

[Gabriel Kerneis]

[-- Attachment #1: Type: text/plain, Size: 692 bytes --]

Le Fri, 16 Mar 2007 04:54:28 -0400, "Mike Lin" <mikelin@mit.edu> a
écrit :
> On 3/7/07, Xavier Leroy <Xavier.Leroy@inria.fr> wrote:
> > - Much work on Camlp4, including a new API.
> Has this broken a lot of stuff?? 

Yes, but once you understand the changes, it's much easier to use.

> e.g. ocamlgsl has a camlp4 extension
> inside it and will not compile under 3.10 ('Unbound value
> Quotation.add'). what are we to do with existing extensions?

Rewrite them. See http://gallium.inria.fr/~pouillar/camlp4-changes.html
for the common pitfalls, it helps a lot in moving from old to new
camlp4.

I guess Nicolas Pouillard will give you more details.

Regards,
-- 
Gabriel

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [Caml-list] beta-test of OCaml 3.10.0

[Nicolas Pouillard]

On 3/16/07, Mike Lin <mikelin@mit.edu> wrote:
>
> On 3/7/07, Xavier Leroy <Xavier.Leroy@inria.fr> wrote:
> >
> > - Much work on Camlp4, including a new API.
>
> Has this broken a lot of stuff?? e.g. ocamlgsl has a camlp4 extension inside
> it and will not compile under 3.10 ('Unbound value Quotation.add'). what are
> we to do with existing extensions?

When Camlp4 is used as a library yes. The API is completly different.

-- 
Nicolas Pouillard

Re: [Caml-list] beta-test of OCaml 3.10.0

[Mike Lin]

[-- Attachment #1: Type: text/plain, Size: 296 bytes --]

On 3/7/07, Xavier Leroy <Xavier.Leroy@inria.fr> wrote:
>
>
> - Much work on Camlp4, including a new API.


Has this broken a lot of stuff?? e.g. ocamlgsl has a camlp4 extension inside
it and will not compile under 3.10 ('Unbound value Quotation.add'). what are
we to do with existing extensions?

[-- Attachment #2: Type: text/html, Size: 590 bytes --]

Re: [Caml-list] beta-test of OCaml 3.10.0

[Sven Luther]

On Wed, Mar 07, 2007 at 02:10:38PM +0100, Xavier Leroy wrote:
> - Two new 64-bit ports: to Windows 64/AMD64 and MacOS X/PowerPC 64.

What about 64bit support for linux/powerpc64 ? 

Friendly,

Sven Luther

Re: [Caml-list] beta-test of OCaml 3.10.0

[Paul Snively]

[-- Attachment #1: Type: text/plain, Size: 1365 bytes --]

Dr. Leroy:

First, my deepest thanks to you and your colleagues for your  
excellent, hard work on the Objective Caml system and its related  
tools. I'm simultaneously learning a tremendous amount from  
programming in O'Caml and having great fun doing so.

I'm curious as to whether there's any chance of having (some variant  
of) the attached patch applied to the O'Caml 3.10.0 codebase. The  
attached patch was provided to me by Alon Kama, maintainer of the  
Ensemble process group communication system, in support of enabling  
Ensemble 2.01 to work correctly on Mac OS X. As you know, Mac OS X is  
a BSD-derived platform, hence makes the distinction between  
SO_REUSEADDR and SO_REUSEPORT in socket options related socket  
rebinding--a distinction that affects much server-oriented socket  
code. So this patch should benefit any BSD-derived platform, not only  
Mac OS X. Having said that, I cannot honestly say that I fully  
understand what changes, particularly conditionalization, might be  
necessary to ensure its proper use on all supported O'Caml platforms;  
nor, of course, can I say whether making that determination with your  
available staff, hardware, OS, and time resources is realistic at  
all, let alone in the time remaining to deliver O'Caml 3.10.0. So I  
appreciate your forbearance.

Many thanks and best regards,
Paul Snively


[-- Attachment #2: ocaml-3.08.3.the_patch.txt --]
[-- Type: text/plain, Size: 2831 bytes --]

diff -Naur ocaml-3.08.3.orig/otherlibs/threads/unix.ml ocaml-3.08.3.patched/otherlibs/threads/unix.ml
--- ocaml-3.08.3.orig/otherlibs/threads/unix.ml	Sat Nov  6 11:14:58 2004
+++ ocaml-3.08.3.patched/otherlibs/threads/unix.ml	Sun Jul 10 17:54:35 2005
@@ -544,6 +544,7 @@
     SO_DEBUG
   | SO_BROADCAST
   | SO_REUSEADDR
+  | SO_REUSEPORT
   | SO_KEEPALIVE
   | SO_DONTROUTE
   | SO_OOBINLINE
diff -Naur ocaml-3.08.3.orig/otherlibs/unix/sockopt.c ocaml-3.08.3.patched/otherlibs/unix/sockopt.c
--- ocaml-3.08.3.orig/otherlibs/unix/sockopt.c	Mon Jan 17 19:10:36 2005
+++ ocaml-3.08.3.patched/otherlibs/unix/sockopt.c	Sun Jul 10 17:56:34 2005
@@ -35,6 +35,9 @@
 #ifndef SO_REUSEADDR
 #define SO_REUSEADDR (-1)
 #endif
+#ifndef SO_REUSEPORT
+#define SO_REUSEPORT (-1)
+#endif
 #ifndef SO_KEEPALIVE
 #define SO_KEEPALIVE (-1)
 #endif
@@ -76,7 +79,7 @@
 #endif
 
 static int sockopt_bool[] = {
-  SO_DEBUG, SO_BROADCAST, SO_REUSEADDR, SO_KEEPALIVE,
+  SO_DEBUG, SO_BROADCAST, SO_REUSEADDR, SO_REUSEPORT, SO_KEEPALIVE,
   SO_DONTROUTE, SO_OOBINLINE, SO_ACCEPTCONN };
 
 static int sockopt_int[] = {
diff -Naur ocaml-3.08.3.orig/otherlibs/unix/unix.ml ocaml-3.08.3.patched/otherlibs/unix/unix.ml
--- ocaml-3.08.3.orig/otherlibs/unix/unix.ml	Sat Nov  6 11:14:58 2004
+++ ocaml-3.08.3.patched/otherlibs/unix/unix.ml	Sun Jul 10 17:57:40 2005
@@ -436,6 +436,7 @@
     SO_DEBUG
   | SO_BROADCAST
   | SO_REUSEADDR
+  | SO_REUSEPORT
   | SO_KEEPALIVE
   | SO_DONTROUTE
   | SO_OOBINLINE
diff -Naur ocaml-3.08.3.orig/otherlibs/unix/unix.mli ocaml-3.08.3.patched/otherlibs/unix/unix.mli
--- ocaml-3.08.3.orig/otherlibs/unix/unix.mli	Wed Dec 22 17:11:13 2004
+++ ocaml-3.08.3.patched/otherlibs/unix/unix.mli	Sun Jul 10 18:05:16 2005
@@ -988,6 +988,7 @@
     SO_DEBUG       (** Record debugging information *)
   | SO_BROADCAST   (** Permit sending of broadcast messages *)
   | SO_REUSEADDR   (** Allow reuse of local addresses for bind *)
+  | SO_REUSEPORT
   | SO_KEEPALIVE   (** Keep connection active *)
   | SO_DONTROUTE   (** Bypass the standard routing algorithms *)
   | SO_OOBINLINE   (** Leave out-of-band data in line *)
diff -Naur ocaml-3.08.3.orig/otherlibs/unix/unixLabels.mli ocaml-3.08.3.patched/otherlibs/unix/unixLabels.mli
--- ocaml-3.08.3.orig/otherlibs/unix/unixLabels.mli	Fri Jul  2 11:37:17 2004
+++ ocaml-3.08.3.patched/otherlibs/unix/unixLabels.mli	Sun Jul 10 17:58:36 2005
@@ -1003,6 +1003,7 @@
     SO_DEBUG               (** Record debugging information *)
   | SO_BROADCAST           (** Permit sending of broadcast messages *)
   | SO_REUSEADDR           (** Allow reuse of local addresses for bind *)
+  | SO_REUSEPORT
   | SO_KEEPALIVE           (** Keep connection active *)
   | SO_DONTROUTE           (** Bypass the standard routing algorithms *)
   | SO_OOBINLINE           (** Leave out-of-band data in line *)

[-- Attachment #3: Type: text/plain, Size: 4787 bytes --]


On Mar 7, 2007, at 5:10 AM, Xavier Leroy wrote:

> Hello,
>
> The long-awaited release 3.10 of Objective Caml is making good
> progress and is now ready for beta-testing.  A source-only
> beta distribution can be downloaded from
>
>     http://caml.inria.fr/pub/distrib/ocaml-3.10/ocaml-3.10.0 
> +beta.tar.gz
>
> We would be grateful if some of you could compile it, give it a
> good workout on your favorite Caml programs, and report any problems
> you might encounter via the Caml bug tracker.
>
> The main novelties in 3.10 are:
>
> - A new tool: ocamlbuild, a compilation manager for OCaml projects.
> - Much work on Camlp4, including a new API.
> - A long-awaited feature: stack backtraces on uncaught exceptions
>   that work for native-code as well as for bytecode.
> - Two new 64-bit ports: to Windows 64/AMD64 and MacOS X/PowerPC 64.
>
> Other changes are listed below.
>
> The documentation is not up to date yet, but some preliminary
> documentation on ocamlbuild and the new Camlp4 can be found on
> Nicolas Pouillard's Web page, http://gallium.inria.fr/~pouillar/
>
> Enjoy,
>
> - Xavier Leroy, for the Caml development team
>
> ----------------------------------------------------------------------
>
> Objective Caml 3.10.0:
> ----------------------
>
> (Changes that can break existing programs are marked with a "*"  )
>
> Language features:
> - Added virtual instance variables in classes "val virtual v : t"
> * Changed the behaviour of instance variable overriding; the new
>   definition replaces the old one, rather than creating a new
>   variable.
>
> New tools:
> - ocamlbuild: compilation manager for OCaml applications and  
> libraries.
>   See draft documentation at http://gallium.inria.fr/~pouillar/
> * Camlp4: heavily revised implementation, new API.
>
> New ports:
> - MacOS X PowerPC 64 bits.
> - MS Windows 64 bits (x64) using the Microsoft PSDK toolchain.
> - MS Windows 32 bits using the Visual Studio 2005 toolchain.
>
> Compilers:
> - Faster type-checking of functor applications.
> - Referencing an interface compiled with -rectypes from a module
>     not compiled with -rectypes is now an error.
> - Revised the "fragile matching" warning.
>
> Native-code compiler:
> - Print a stack backtrace on an uncaught exception.
>   (Compile and link with ocamlopt -g; execute with OCAMLRUNPARAM=b.)
>   Supported on Intel/AMD in 32 and 64 bits, PPC in 32 and 64 bits.
> - Stack overflow detection on MS Windows 32 bits (courtesy O.  
> Andrieu).
> - Stack overflow detection on MacOS X PPC and Intel.
> - Intel/AMD 64 bits: generate position-independent code by default.
> - Fixed bug involving -for-pack and missing .cmx files (PR#4124).
> - Fixed bug causing duplication of literals  (PR#4152).
>
> Run-time system:
> - C/Caml interface functions take "char const *" arguments
>   instead of "char *" when appropriate.
> - Faster string comparisons (fast case if strings are ==).
>
> Standard library:
> - Refined typing of format strings (type format6).
> - Printf, Format: new function ifprintf that consumes its arguments
>     and prints nothing (useful to print conditionally).
> - Scanf:
>     new function format_from_string to convert a string to a format  
> string;
>     new %r conversion to accomodate user defined scanners.
> - Filename: improved Win32 implementation of Filename.quote.
> - List: List.nth now tail-recursive.
> - Sys: added Sys.is_directory.  Some functions (e.g. Sys.command) that
>     could incorrectly raise Sys_io_blocked now raise Sys_error as  
> intended.
>
> Other libraries:
> - Bigarray: mmap_file takes an optional argument specifying
>     the start position of the data in the mapped file.
> - Dynlink: now defines only two modules, Dynlink and Dynlinkaux  
> (internal),
>     reducing risks of name conflicts with user modules.
> - Labltk under Win32: now uses Tcl/Tk 8.4 instead of 8.3 by default.
> - VM threads: improved performance of I/O operations (less polling).
> - Unix: new function Unix.isatty.
> - Unix emulation under Win32:
>     fixed incorrect error reporting in several functions (PR#4097);
>     better handling of channels opened on sockets (PR#4098);
>     fixed GC bug in Unix.system (PR#4112).
>
> Documentation generator (OCamldoc):
> - correctly handle '?' in value names (PR#4215)
> - new option -hide-warnings not to print ocamldoc warnings
>
> Lexer generator (ocamllex): improved error reporting.
>
> License: fixed a typo in the "special exception" to the LGPL.
>
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs

Re: [Caml-list] beta-test of OCaml 3.10.0

[Nicolas Pouillard]

On 3/7/07, Vu Ngoc San <san.vu-ngoc@ujf-grenoble.fr> wrote:
> I've noticed that standard compilation (make world.opt) uses 333 Mb of
> hard disk, while ./build/fastworld.sh uses much more: 443 Mb

Yep that's due to copying into the build directory.

-- 
Nicolas Pouillard

Re: [Caml-list] beta-test of OCaml 3.10.0

[Vu Ngoc San]

I've noticed that standard compilation (make world.opt) uses 333 Mb of 
hard disk, while ./build/fastworld.sh uses much more: 443 Mb

anyhow both work :)

(Linux debian 2.6.18-4-686)

Re: [Caml-list] beta-test of OCaml 3.10.0

[Joel Reymont]

On Mar 7, 2007, at 3:01 PM, Nicolas Pouillard wrote:

> You have this problem even after a full clean (make clean &&
> ./build/distclean.sh) ?

That works fine, thanks!

--
http://wagerlabs.com/

Re: [Caml-list] beta-test of OCaml 3.10.0

[Nicolas Pouillard]

On 3/7/07, Joel Reymont <joelr1@gmail.com> wrote:
>
> On Mar 7, 2007, at 2:08 PM, micha wrote:
>
> > building with "./build/fastworld.sh" doesn't work here (Etch,
> > 2.6.20.1,
> > Athlon XP)
>
> You need to do 'make clean' beforehand but I still get this on Mac
> OSX 10.4.8 Intel:

You have this problem even after a full clean (make clean &&
./build/distclean.sh) ?

>
> Error while linking bytecomp/symtable.cmo:
> Reference to undefined global `Runtimedef'
> Exit code 2 while executing this command:
>    boot/ocamlrun ./ocamlc -nostdlib -g -I stdlib
> myocamlbuild_config.cmo utils/config.cmo bytecomp/bytesections.cmo
> utils/misc.cmo bytecomp/dll.cmo parsing/linenum.cmo utils/
> terminfo.cmo utils/warnings.cmo parsing/location.cmo parsing/
> longident.cmo typing/ident.cmo typing/path.cmo typing/primitive.cmo
> typing/types.cmo typing/btype.cmo typing/predef.cmo typing/
> datarepr.cmo utils/tbl.cmo typing/subst.cmo utils/clflags.cmo utils/
> consistbl.cmo typing/env.cmo bytecomp/lambda.cmo bytecomp/
> instruct.cmo bytecomp/opcodes.cmo bytecomp/meta.cmo bytecomp/
> symtable.cmo utils/ccomp.cmo bytecomp/bytelink.cmo bytecomp/
> bytelibrarian.cmo bytecomp/switch.cmo bytecomp/bytegen.cmo bytecomp/
> printlambda.cmo typing/ctype.cmo typing/typedtree.cmo bytecomp/
> typeopt.cmo typing/parmatch.cmo bytecomp/matching.cmo bytecomp/
> translobj.cmo bytecomp/translcore.cmo bytecomp/translclass.cmo typing/
> mtype.cmo typing/oprint.cmo typing/printtyp.cmo bytecomp/
> translmod.cmo bytecomp/emitcode.cmo typing/includeclass.cmo typing/
> includecore.cmo typing/includemod.cmo typing/stypes.cmo typing/
> typetexp.cmo typing/typecore.cmo typing/typedecl.cmo typing/
> typeclass.cmo typing/typemod.cmo bytecomp/bytepackager.cmo bytecomp/
> printinstr.cmo bytecomp/simplif.cmo driver/pparse.cmo parsing/
> syntaxerr.cmo parsing/parser.cmo parsing/lexer.cmo parsing/parse.cmo
> parsing/printast.cmo typing/unused_var.cmo driver/compile.cmo driver/
> errors.cmo driver/main_args.cmo driver/main.cmo -o driver/main.byte
>
> --
> http://wagerlabs.com/
>
>
>
>
>
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
>


-- 
Nicolas Pouillard

Re: [Caml-list] beta-test of OCaml 3.10.0

[Nicolas Pouillard]

On 3/7/07, micha <micha-1@fantasymail.de> wrote:
> Am Wed, 7 Mar 2007 14:12:32 +0000
> schrieb Joel Reymont <joelr1@gmail.com>:
>
> >
> > On Mar 7, 2007, at 2:08 PM, micha wrote:
> >
> > > building with   doesn't work here (Etch,
> > > 2.6.20.1,
> > > Athlon XP)
> >
> > You need to do 'make clean' beforehand but I still get this on Mac
> > OSX 10.4.8 Intel:
>

Switching from one mode to another require this:

$ make clean
$ ./build/distclean.sh

Then running configure is required.

-- 
Nicolas Pouillard

Re: [Caml-list] beta-test of OCaml 3.10.0

[micha]

Am Wed, 7 Mar 2007 14:12:32 +0000
schrieb Joel Reymont <joelr1@gmail.com>:

> 
> On Mar 7, 2007, at 2:08 PM, micha wrote:
> 
> > building with   doesn't work here (Etch,  
> > 2.6.20.1,
> > Athlon XP)
> 
> You need to do 'make clean' beforehand but I still get this on Mac  
> OSX 10.4.8 Intel:

so make clean and then "./build/fastworld.sh" gives this error:

+ boot/ocamlrun boot/myocamlbuild.boot -just-plugin -install-lib-dir
_build/ocamlbuild -byte-plugin boot/ocamlrun boot/ocamlc -I boot
-nostdlib
-I /home/micha/ocaml-3.10.0+beta/_build/ocamlbuild /home/micha/ocaml-3.10.0+beta/_build/ocamlbuild/ocamlbuildlightlib.cma
myocamlbuild_config.mli myocamlbuild_config.ml
myocamlbuild.ml /home/micha/ocaml-3.10.0+beta/_build/ocamlbuild/ocamlbuildlight.cmo
-o myocamlbuild
>> Fatal error: cannot open pervasives.cmi
Fatal error: exception Misc.Fatal_error
(Program not linked with -g, cannot print stack backtrace)
Exit code 2 while executing this command:
  boot/ocamlrun boot/ocamlc -I boot -nostdlib
-I /home/micha/ocaml-3.10.0+beta/_build/ocamlbuild /home/micha/ocaml-3.10.0+beta/_build/ocamlbuild/ocamlbuildlightlib.cma
myocamlbuild_config.mli myocamlbuild_config.ml
myocamlbuild.ml /home/micha/ocaml-3.10.0+beta/_build/ocamlbuild/ocamlbuildlight.cmo
-o myocamlbuild

Re: [Caml-list] beta-test of OCaml 3.10.0

[Joel Reymont]

On Mar 7, 2007, at 2:08 PM, micha wrote:

> building with "./build/fastworld.sh" doesn't work here (Etch,  
> 2.6.20.1,
> Athlon XP)

You need to do 'make clean' beforehand but I still get this on Mac  
OSX 10.4.8 Intel:

Error while linking bytecomp/symtable.cmo:
Reference to undefined global `Runtimedef'
Exit code 2 while executing this command:
   boot/ocamlrun ./ocamlc -nostdlib -g -I stdlib  
myocamlbuild_config.cmo utils/config.cmo bytecomp/bytesections.cmo  
utils/misc.cmo bytecomp/dll.cmo parsing/linenum.cmo utils/ 
terminfo.cmo utils/warnings.cmo parsing/location.cmo parsing/ 
longident.cmo typing/ident.cmo typing/path.cmo typing/primitive.cmo  
typing/types.cmo typing/btype.cmo typing/predef.cmo typing/ 
datarepr.cmo utils/tbl.cmo typing/subst.cmo utils/clflags.cmo utils/ 
consistbl.cmo typing/env.cmo bytecomp/lambda.cmo bytecomp/ 
instruct.cmo bytecomp/opcodes.cmo bytecomp/meta.cmo bytecomp/ 
symtable.cmo utils/ccomp.cmo bytecomp/bytelink.cmo bytecomp/ 
bytelibrarian.cmo bytecomp/switch.cmo bytecomp/bytegen.cmo bytecomp/ 
printlambda.cmo typing/ctype.cmo typing/typedtree.cmo bytecomp/ 
typeopt.cmo typing/parmatch.cmo bytecomp/matching.cmo bytecomp/ 
translobj.cmo bytecomp/translcore.cmo bytecomp/translclass.cmo typing/ 
mtype.cmo typing/oprint.cmo typing/printtyp.cmo bytecomp/ 
translmod.cmo bytecomp/emitcode.cmo typing/includeclass.cmo typing/ 
includecore.cmo typing/includemod.cmo typing/stypes.cmo typing/ 
typetexp.cmo typing/typecore.cmo typing/typedecl.cmo typing/ 
typeclass.cmo typing/typemod.cmo bytecomp/bytepackager.cmo bytecomp/ 
printinstr.cmo bytecomp/simplif.cmo driver/pparse.cmo parsing/ 
syntaxerr.cmo parsing/parser.cmo parsing/lexer.cmo parsing/parse.cmo  
parsing/printast.cmo typing/unused_var.cmo driver/compile.cmo driver/ 
errors.cmo driver/main_args.cmo driver/main.cmo -o driver/main.byte

--
http://wagerlabs.com/

Re: [Caml-list] beta-test of OCaml 3.10.0

[micha]

Am Wed, 07 Mar 2007 14:10:38 +0100
schrieb Xavier Leroy <Xavier.Leroy@inria.fr>:

> 
> - A new tool: ocamlbuild, a compilation manager for OCaml projects.

building with "./build/fastworld.sh" doesn't work here (Etch, 2.6.20.1,
Athlon XP)

> ./configure -cc "gcc -march=athlon-xp -O3"
> make world ..
> ./build/fastworld.sh

after some compilation gives errors:
...
+ cd yacc
+ make
make: Für das Ziel »all« ist nichts zu tun.
+ ./boot/ocamlrun boot/myocamlbuild.boot boot/stdlib.cma
boot/std_exit.cmo
SANITIZE: a total of 664 files that should probably not be in your
source
  tree has been found. A script shell file "sanitize.sh" is being
created.
  Check this script and run it to remove unwanted files or use other
options
  (such as defining hygiene exceptions or using the -no-hygiene option).
IMPORTANT: I cannot work with leftover compiled files.
ERROR: Leftover Ocaml compilation files:
  File lambda.cmo in bytecomp has suffix .cmo
  File printlambda.cmo in bytecomp has suffix .cmo
  File typeopt.cmo in bytecomp has suffix .cmo
  File switch.cmo in bytecomp has suffix .cmo
... many more...
ERROR: Leftover ocamllex-generated files:
  Files lexer.mll and lexer.ml should not be together in debugger
ERROR: Leftover ocamlyacc-generated files:
  Files parser.mly and parser.ml should not be together in debugger
  Files parser.mly and parser.mli should not be together in debugger
ERROR: Leftover Ocaml compilation files:
  File dynlink.cmo in debugger has suffix .cmo
 ...many more...
Exiting due to hygiene violations.



cheers,
 Michael