ocamlnet-devel — Developer discussion

Re: [Ocamlnet-devel] Multithreaded https requests in ocamlnet netclient

[Mike L. <ml...@ml...>]

Thanks, Gerd!

FWIW I could not reproduce the crash by using ocaml-ssl's blocking
operations directly.

https://gist.github.com/4152047#file_ssl_threads.ml

This works fine- so, perhaps something nasty arises from using nonblocking
I/O on ssl sockets from multiple threads. I'm sure if there is any other
critical difference with how netclient/equeue-ssl and my example use
ocaml-ssl.

I also don't have time to pursue this much further, so I will try to put
all of my http operations on one thread as your example suggests.

Best,
Mike


On Tue, Nov 27, 2012 at 5:03 AM, Gerd Stolpmann <in...@ge...>wrote:

> Hi Mike,
>
> I can confirm the problem, but I'm not able to track it down.
>
> First of all, I'm quite sure that you cannot share SSL contexts between
> threads. But even if I fix this, I get crashes - usually early during
> SSL_connect.
>
> I'm wondering whether ocaml-ssl initializes openssl correctly for
> multi-threading. In my version of openssl the prototype is
>
> int CRYPTO_THREADID_set_callback(**void (*threadid_func)(CRYPTO_**THREADID
> *));
>
> but in ocaml-ssl the CRYPTO_THREADID pointer is ignored (it assumes void
> here). So, maybe this is the reason. I currently don't have time to follow
> this idea.
>
> If everything fails, there is a way to use Netclient in threaded
> applications so that only one thread is doing the networking stuff, see
> this example:
>
> https://godirepo.camlcity.org/**svn/lib-ocamlnet2/trunk/code/**
> examples/netclient/simple/**http_mt.ml<https://godirepo.camlcity.org/svn/lib-ocamlnet2/trunk/code/examples/netclient/simple/http_mt.ml>
>
> As Netclient is event-driven, the requests are processed in parallel (with
> one exception, though: DNS lookups are synchronous, but this can be worked
> around, as the DNS lookup function is settable in Netclient).
>
> Gerd
>
> Am 27.11.2012 03:54:16 schrieb(en) Mike Lin:
>
>  Hello Gerd, all,
>>
>> I'm trying to find the right incantations to make parallel https requests
>> from multiple threads using Http_client. (I know the library is capable of
>> multiple requests from one thread -- I have other strong reasons to use
>> multithreading.)
>>
>> https://gist.github.com/**4152047 <https://gist.github.com/4152047>
>>
>> This program generally segfaults for me at some random point during the 10
>> requests.  Compiled as in the comment in line 2 (ubuntu quantal,
>> libocamlnet-ocaml-dev 3.5.1, ocaml 3.12.1).
>>
>> Permutations attempted:
>>
>> 1) Change https to http in google URL -- works as expected
>> 2) Serialize requests by uncommenting line 35 -- works as expected
>> 3) Move invocations of Ssl.init, Ssl.create_context,
>> HTTPS.https_transport_channel_**type in and out of fresh_pipeline (which
>> runs
>> in the worker threads) -- no help
>>
>> Any suggestions? Thanks!
>>
>> Mike Lin
>>
>>
>
>
> --
> ------------------------------**------------------------------
> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
> Creator of GODI and camlcity.org.
> Contact details:        http://www.camlcity.org/**contact.html<http://www.camlcity.org/contact.html>
> Company homepage:       http://www.gerd-stolpmann.de
> ------------------------------**------------------------------

Re: [Ocamlnet-devel] Multithreaded https requests in ocamlnet netclient

[Gerd S. <in...@ge...>]

Hi Mike,

I can confirm the problem, but I'm not able to track it down.

First of all, I'm quite sure that you cannot share SSL contexts between  
threads. But even if I fix this, I get crashes - usually early during  
SSL_connect.

I'm wondering whether ocaml-ssl initializes openssl correctly for  
multi-threading. In my version of openssl the prototype is

int CRYPTO_THREADID_set_callback(void (*threadid_func)(CRYPTO_THREADID  
*));

but in ocaml-ssl the CRYPTO_THREADID pointer is ignored (it assumes  
void here). So, maybe this is the reason. I currently don't have time  
to follow this idea.

If everything fails, there is a way to use Netclient in threaded  
applications so that only one thread is doing the networking stuff, see  
this example:

https://godirepo.camlcity.org/svn/lib-ocamlnet2/trunk/code/examples/netclient/simple/http_mt.ml

As Netclient is event-driven, the requests are processed in parallel  
(with one exception, though: DNS lookups are synchronous, but this can  
be worked around, as the DNS lookup function is settable in Netclient).

Gerd

Am 27.11.2012 03:54:16 schrieb(en) Mike Lin:
> Hello Gerd, all,
> 
> I'm trying to find the right incantations to make parallel https  
> requests
> from multiple threads using Http_client. (I know the library is  
> capable of
> multiple requests from one thread -- I have other strong reasons to  
> use
> multithreading.)
> 
> https://gist.github.com/4152047
> 
> This program generally segfaults for me at some random point during  
> the 10
> requests.  Compiled as in the comment in line 2 (ubuntu quantal,
> libocamlnet-ocaml-dev 3.5.1, ocaml 3.12.1).
> 
> Permutations attempted:
> 
> 1) Change https to http in google URL -- works as expected
> 2) Serialize requests by uncommenting line 35 -- works as expected
> 3) Move invocations of Ssl.init, Ssl.create_context,
> HTTPS.https_transport_channel_type in and out of fresh_pipeline  
> (which runs
> in the worker threads) -- no help
> 
> Any suggestions? Thanks!
> 
> Mike Lin
> 



-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------

[Ocamlnet-devel] Multithreaded https requests in ocamlnet netclient

[Mike L. <ml...@ml...>]

Hello Gerd, all,

I'm trying to find the right incantations to make parallel https requests
from multiple threads using Http_client. (I know the library is capable of
multiple requests from one thread -- I have other strong reasons to use
multithreading.)

https://gist.github.com/4152047

This program generally segfaults for me at some random point during the 10
requests.  Compiled as in the comment in line 2 (ubuntu quantal,
libocamlnet-ocaml-dev 3.5.1, ocaml 3.12.1).

Permutations attempted:

1) Change https to http in google URL -- works as expected
2) Serialize requests by uncommenting line 35 -- works as expected
3) Move invocations of Ssl.init, Ssl.create_context,
HTTPS.https_transport_channel_type in and out of fresh_pipeline (which runs
in the worker threads) -- no help

Any suggestions? Thanks!

Mike Lin

Re: [Ocamlnet-devel] [ocamlnet 3.6] SOLVED Netplex broken on Mac OS X 10.4 PPC

[Gerd S. <in...@ge...>]

Am 28.09.2012 21:55:39 schrieb(en) Matías Giovannini:
> Well, after a long detour I found and *fixed* the issue. After many  
> hours
> trying to backport to Ocamlnet 2.x, I found that the root cause is  
> the one
> identified by Joel Reymont back in 2007:
> http://caml.inria.fr/pub/ml-archives/caml-list/2007/05/0a6a0d3113993e5eefccb7cc3aa3eebe.en.html
> 
> Here are my diffs against 3.6:
> ~/src/ocamlnet-3.6/src/netsys> diff netsys.ml  
> netsys.ml.orig125,127d124
> <     | Unix.Unix_error(Unix.EAFNOSUPPORT,_,_) ->
> <     (* On older Mac OS X, getpeername on a pipe return EAFNOSUPPORT;
> assume *)
> < `Read_write

No, returning `Recv_send_implied is not wrong: We are not using a pipe  
here, but a socketpair. The error is in Uq_engines, where at one point  
`Recv_send_implied is not supported where it should be.

My tentative fix is in svn.

Gerd

> ~/src/ocamlnet-3.6/src/equeue> diff uq_engines.ml uq_engines.ml.orig
> 3132c3132
> <   try match Netsys.getpeername s with
> ---
> >   match Netsys.getpeername s with
> 3137,3138d3136
> <   with Unix.Unix_error (Unix.EAFNOSUPPORT, _, _)
> <     `Sock_unix(stype, "")
> 3143c3141
> <   try match Netsys.getpeername s with
> ---
> >   match Netsys.getpeername s with
> 3148,3149d3145
> <   with Unix.Unix_error (Unix.EAFNOSUPPORT, _, _)
> < None
> 
> 
> On Wed, Sep 26, 2012 at 9:15 PM, Matías Giovannini  
> <mgi...@gm...>wrote:
> 
> > Hi,
> >
> > I send out in the hopes that, despite the ancient software and  
> platform
> > I'm having problems with, someone could shed some light on this.  
> Currently
> > netplex is broken on this machine, failing with errors of the  
> following
> > sort:
> >
> > fd style: Recv_send_implied
> > fd style: Recv_send_implied
> > Netplex Catastrophic Error: Unable to send log message - exception
> > Rpc_client.Communication_error(File "uq_engines.ml", line 2576,
> > characters 6-12: Assertion failed)
> > Log message is: poll: Exception Rpc_client.Communication_error(File  
> "
> > uq_engines.ml", line 2576, characters 6-12: Assertion failed)
> > [Wed Sep 26 21:10:39 2012] [netplex.controller] [alert] Disabling  
> service
> > after 10 startup failures
> >
> > (the "fd_style" line is a strategically placed print statement). I  
> am at a
> > loss in following uq_engines.ml, so I don't know where to look.
> >
> > I configured ocamlnet with:
> >
> > ./configure -with-nethttpd -enable-zip -disable-tcl -prefer-netcgi2
> > -enable-pcre
> >
> > and the result is:
> >
> > able-tcl -prefer-netcgi2 -enable-pcre
> > Welcome to Ocamlnet version 3.6
> > Checking operating system... Generic
> > Checking for findlib... found
> > Checking multi-threading support... posix (ok)
> > Checking word size... 32 bit
> > Checking endianess... big
> > Checking for GPROF... found
> > Checking for PCRE... found
> > Checking for win32... no
> > Checking whether Ocaml has Printexc.register_printer... found
> > Checking for POSIX pthread... found
> > Checking whether Ocaml has fancy page tables... found
> > Checking for POSIX shared memory... found
> > Checking for POSIX semaphores (anonymous)... not found
> > Checking for POSIX semaphores (named)... found
> > Checking for POSIX spawn... not found
> > Checking for POSIX fadvise... not found
> > Checking for POSIX fallocate... not found
> > Checking for POSIX memalign... not found
> > Checking for syslog... found
> > Checking for POSIX functions like openat... not found
> > Checking for fchdir... found
> > Checking for fdopendir... not found
> > Checking for realpath... not found
> > Checking for grantpt (System V style PTYs)... found
> > Checking for posix_openpt (System V style PTYs)... found
> > Checking for initgroups...found
> > Checking for POSIX clocks...not found
> > Checking for POSIX timers...not found
> > Checking for eventfd (Linux)...not found
> > Checking for timerfd (Linux)...not found
> > Checking for epoll (Linux)...not found
> > Checking for compare_and_swap (GCC)...not found
> > Checking for GPROF... found
> > Checking for O_SHARE_DELETE... yes
> > Checking for getpeereid... found
> > Checking for getpeerucred... not found
> > Checking for camlzip... found
> > Preprocessor for ocamlrpcgen: cpp
> >
> > Effective options:
> >     -enable-pcre
> >     -disable-gtk
> >     -disable-gtk2
> >     -disable-tcl
> >     -disable-ssl
> >     -enable-zip
> >     -disable-crypto
> >     -disable-apache
> >     -with-nethttpd
> >     -without-rpc-auth-dh
> >     -bindir /opt/ocaml/bin
> >     -datadir /opt/ocaml/lib/site-lib/netstring
> >     -cpp cpp
> >
> > uname -a is "Darwin Blue.local 8.11.0 Darwin Kernel Version 8.11.0:  
> Wed
> > Oct 10 18:26:00 PDT 2007; root:xnu-792.24.17~1/RELEASE_PPC Power  
> Macintosh
> > powerpc", ocaml -version is "The OCaml toplevel, version 4.00.0"  
> and gcc
> > --version gives:
> >
> > powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc.  
> build
> > 5370)
> > Copyright (C) 2005 Free Software Foundation, Inc.
> > This is free software; see the source for copying conditions.   
> There is NO
> > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR  
> PURPOSE.
> >
> > (prehistoric, I know). If anyone has any ideas, I'd be grateful  
> (except
> > for suggestions to run Linux! ;-)
> >
> > Best regards,
> > Matías.
> >
> 

------Zitierte Anlage------
> ------------------------------------------------------------------------------
> Got visibility?
> Most devs has no idea what their production app looks like.
> Find out how fast your code is with AppDynamics Lite.
> http://ad.doubleclick.net/clk;262219671;13503038;y?
> http://info.appdynamics.com/FreeJavaPerformanceDownload.html

------Zitierte Anlage------
> _______________________________________________
> Ocamlnet-devel mailing list
> Oca...@li...
> https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
> 



-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------

Re: [Ocamlnet-devel] PPC detected as little endian (!)

[Gerd S. <in...@ge...>]

Thanks Matías,

I've changed the test in svn.

Gerd

Am 27.09.2012 00:25:55 schrieb(en) Matías Giovannini:
> Hello,
> 
> after seeing ocamlnet 3.6 fail spectacularly on a Power Mac, I tried
> recompiling and watching closely the configuration options. Lo and  
> behold,
> configure proudly proclaims:
> 
> Checking endianess... little
> 
> which is of course completely wrong. The problem is that pointer  
> aliasing
> via unions is undefined behavior in C (barring packing pragmas and  
> other
> compiler wizardry), and GCC/LLVM do whatever they want with that  
> code. I've
> found that I can distinguish endianness reliably with the following  
> check:
> 
> value check(value d) {
>   int i = 1;
>   char *s = (char*) &i;
>   return (s[0] == 0 ? Val_true : Val_false);
> }
> 
> Regards,
> Matías.
> 

------Zitierte Anlage------
> ------------------------------------------------------------------------------
> How fast is your code?
> 3 out of 4 devs don\\\'t know how their code performs in production.
> Find out how slow your code is with AppDynamics Lite.
> http://ad.doubleclick.net/clk;262219672;13503038;z?
> http://info.appdynamics.com/FreeJavaPerformanceDownload.html

------Zitierte Anlage------
> _______________________________________________
> Ocamlnet-devel mailing list
> Oca...@li...
> https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
> 



-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------

Re: [Ocamlnet-devel] [ocamlnet 3.6] SOLVED Netplex broken on Mac OS X 10.4 PPC

[Matías G. <mgi...@gm...>]

Well, after a long detour I found and *fixed* the issue. After many hours
trying to backport to Ocamlnet 2.x, I found that the root cause is the one
identified by Joel Reymont back in 2007:
http://caml.inria.fr/pub/ml-archives/caml-list/2007/05/0a6a0d3113993e5eefccb7cc3aa3eebe.en.html

Here are my diffs against 3.6:

~/src/ocamlnet-3.6> diff configure configure.orig
450,452c450,455
<   int i = 1;
<   char* s = (char*) &i;
<   return (s[0] == 0 ? Val_true : Val_false);
---
>   union {
>     int i;
>     char s[4];
>   } u;
>   u.i = 0x0001010101;
>   return (u.s[0] == 0 ? Val_true : Val_false);

~/src/ocamlnet-3.6/src/netzip> diff netgzip.ml netgzip.ml.orig
326c326
<     ~encoder:(new deflating_pipe ~level:6)
---
>     ~encoder:(new deflating_pipe)

~/src/ocamlnet-3.6/src/netsys> diff netsys.ml netsys.ml.orig125,127d124
<     | Unix.Unix_error(Unix.EAFNOSUPPORT,_,_) ->
<     (* On older Mac OS X, getpeername on a pipe return EAFNOSUPPORT;
assume *)
< `Read_write

~/src/ocamlnet-3.6/src/equeue> diff uq_engines.ml uq_engines.ml.orig
3132c3132
<   try match Netsys.getpeername s with
---
>   match Netsys.getpeername s with
3137,3138d3136
<   with Unix.Unix_error (Unix.EAFNOSUPPORT, _, _)
<     `Sock_unix(stype, "")
3143c3141
<   try match Netsys.getpeername s with
---
>   match Netsys.getpeername s with
3148,3149d3145
<   with Unix.Unix_error (Unix.EAFNOSUPPORT, _, _)
< None


On Wed, Sep 26, 2012 at 9:15 PM, Matías Giovannini <mgi...@gm...>wrote:

> Hi,
>
> I send out in the hopes that, despite the ancient software and platform
> I'm having problems with, someone could shed some light on this. Currently
> netplex is broken on this machine, failing with errors of the following
> sort:
>
> fd style: Recv_send_implied
> fd style: Recv_send_implied
> Netplex Catastrophic Error: Unable to send log message - exception
> Rpc_client.Communication_error(File "uq_engines.ml", line 2576,
> characters 6-12: Assertion failed)
> Log message is: poll: Exception Rpc_client.Communication_error(File "
> uq_engines.ml", line 2576, characters 6-12: Assertion failed)
> [Wed Sep 26 21:10:39 2012] [netplex.controller] [alert] Disabling service
> after 10 startup failures
>
> (the "fd_style" line is a strategically placed print statement). I am at a
> loss in following uq_engines.ml, so I don't know where to look.
>
> I configured ocamlnet with:
>
> ./configure -with-nethttpd -enable-zip -disable-tcl -prefer-netcgi2
> -enable-pcre
>
> and the result is:
>
> able-tcl -prefer-netcgi2 -enable-pcre
> Welcome to Ocamlnet version 3.6
> Checking operating system... Generic
> Checking for findlib... found
> Checking multi-threading support... posix (ok)
> Checking word size... 32 bit
> Checking endianess... big
> Checking for GPROF... found
> Checking for PCRE... found
> Checking for win32... no
> Checking whether Ocaml has Printexc.register_printer... found
> Checking for POSIX pthread... found
> Checking whether Ocaml has fancy page tables... found
> Checking for POSIX shared memory... found
> Checking for POSIX semaphores (anonymous)... not found
> Checking for POSIX semaphores (named)... found
> Checking for POSIX spawn... not found
> Checking for POSIX fadvise... not found
> Checking for POSIX fallocate... not found
> Checking for POSIX memalign... not found
> Checking for syslog... found
> Checking for POSIX functions like openat... not found
> Checking for fchdir... found
> Checking for fdopendir... not found
> Checking for realpath... not found
> Checking for grantpt (System V style PTYs)... found
> Checking for posix_openpt (System V style PTYs)... found
> Checking for initgroups...found
> Checking for POSIX clocks...not found
> Checking for POSIX timers...not found
> Checking for eventfd (Linux)...not found
> Checking for timerfd (Linux)...not found
> Checking for epoll (Linux)...not found
> Checking for compare_and_swap (GCC)...not found
> Checking for GPROF... found
> Checking for O_SHARE_DELETE... yes
> Checking for getpeereid... found
> Checking for getpeerucred... not found
> Checking for camlzip... found
> Preprocessor for ocamlrpcgen: cpp
>
> Effective options:
>     -enable-pcre
>     -disable-gtk
>     -disable-gtk2
>     -disable-tcl
>     -disable-ssl
>     -enable-zip
>     -disable-crypto
>     -disable-apache
>     -with-nethttpd
>     -without-rpc-auth-dh
>     -bindir /opt/ocaml/bin
>     -datadir /opt/ocaml/lib/site-lib/netstring
>     -cpp cpp
>
> uname -a is "Darwin Blue.local 8.11.0 Darwin Kernel Version 8.11.0: Wed
> Oct 10 18:26:00 PDT 2007; root:xnu-792.24.17~1/RELEASE_PPC Power Macintosh
> powerpc", ocaml -version is "The OCaml toplevel, version 4.00.0" and gcc
> --version gives:
>
> powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build
> 5370)
> Copyright (C) 2005 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>
> (prehistoric, I know). If anyone has any ideas, I'd be grateful (except
> for suggestions to run Linux! ;-)
>
> Best regards,
> Matías.
>

[Ocamlnet-devel] [ocamlnet 3.6] Netplex broken on Mac OS X 10.4 PPC

[Matías G. <mgi...@gm...>]

Hi,

I send out in the hopes that, despite the ancient software and platform I'm
having problems with, someone could shed some light on this. Currently
netplex is broken on this machine, failing with errors of the following
sort:

fd style: Recv_send_implied
fd style: Recv_send_implied
Netplex Catastrophic Error: Unable to send log message - exception
Rpc_client.Communication_error(File "uq_engines.ml", line 2576, characters
6-12: Assertion failed)
Log message is: poll: Exception Rpc_client.Communication_error(File "
uq_engines.ml", line 2576, characters 6-12: Assertion failed)
[Wed Sep 26 21:10:39 2012] [netplex.controller] [alert] Disabling service
after 10 startup failures

(the "fd_style" line is a strategically placed print statement). I am at a
loss in following uq_engines.ml, so I don't know where to look.

I configured ocamlnet with:

./configure -with-nethttpd -enable-zip -disable-tcl -prefer-netcgi2
-enable-pcre

and the result is:

able-tcl -prefer-netcgi2 -enable-pcre
Welcome to Ocamlnet version 3.6
Checking operating system... Generic
Checking for findlib... found
Checking multi-threading support... posix (ok)
Checking word size... 32 bit
Checking endianess... big
Checking for GPROF... found
Checking for PCRE... found
Checking for win32... no
Checking whether Ocaml has Printexc.register_printer... found
Checking for POSIX pthread... found
Checking whether Ocaml has fancy page tables... found
Checking for POSIX shared memory... found
Checking for POSIX semaphores (anonymous)... not found
Checking for POSIX semaphores (named)... found
Checking for POSIX spawn... not found
Checking for POSIX fadvise... not found
Checking for POSIX fallocate... not found
Checking for POSIX memalign... not found
Checking for syslog... found
Checking for POSIX functions like openat... not found
Checking for fchdir... found
Checking for fdopendir... not found
Checking for realpath... not found
Checking for grantpt (System V style PTYs)... found
Checking for posix_openpt (System V style PTYs)... found
Checking for initgroups...found
Checking for POSIX clocks...not found
Checking for POSIX timers...not found
Checking for eventfd (Linux)...not found
Checking for timerfd (Linux)...not found
Checking for epoll (Linux)...not found
Checking for compare_and_swap (GCC)...not found
Checking for GPROF... found
Checking for O_SHARE_DELETE... yes
Checking for getpeereid... found
Checking for getpeerucred... not found
Checking for camlzip... found
Preprocessor for ocamlrpcgen: cpp

Effective options:
    -enable-pcre
    -disable-gtk
    -disable-gtk2
    -disable-tcl
    -disable-ssl
    -enable-zip
    -disable-crypto
    -disable-apache
    -with-nethttpd
    -without-rpc-auth-dh
    -bindir /opt/ocaml/bin
    -datadir /opt/ocaml/lib/site-lib/netstring
    -cpp cpp

uname -a is "Darwin Blue.local 8.11.0 Darwin Kernel Version 8.11.0: Wed Oct
10 18:26:00 PDT 2007; root:xnu-792.24.17~1/RELEASE_PPC Power Macintosh
powerpc", ocaml -version is "The OCaml toplevel, version 4.00.0" and gcc
--version gives:

powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build
5370)
Copyright (C) 2005 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

(prehistoric, I know). If anyone has any ideas, I'd be grateful (except for
suggestions to run Linux! ;-)

Best regards,
Matías.

[Ocamlnet-devel] PPC detected as little endian (!)

[Matías G. <mgi...@gm...>]

Hello,

after seeing ocamlnet 3.6 fail spectacularly on a Power Mac, I tried
recompiling and watching closely the configuration options. Lo and behold,
configure proudly proclaims:

Checking endianess... little

which is of course completely wrong. The problem is that pointer aliasing
via unions is undefined behavior in C (barring packing pragmas and other
compiler wizardry), and GCC/LLVM do whatever they want with that code. I've
found that I can distinguish endianness reliably with the following check:

value check(value d) {
  int i = 1;
  char *s = (char*) &i;
  return (s[0] == 0 ? Val_true : Val_false);
}

Regards,
Matías.

[Ocamlnet-devel] [ocamlnet 3.6] Error building netgzip

[Matías G. <mgi...@gm...>]

Hello,

there's a problem building ocamlnet 3.6 with OCaml 4.0 (related to some
changes to inference of optional parameters):

ocamlfind ocamlc -g  -I ../../src/netstring -package "unix camlzip"  -c
netgzip.ml
File "netgzip.ml", line 326, characters 13-33:
Error: This expression has type ?level:int -> unit -> deflating_pipe
       but an expression was expected of type
         unit -> Netchannels.io_obj_channel
make[1]: *** [netgzip.cmo] Error 2
make: *** [all] Error 2

The workaround is to edit src/netzip/netgzip.ml so that line 326 passes an
explicit ``level'' option to deflating_pipe:

let () =
  Netcompression.register
    ~iana_name:"gzip"
    ~decoder:(new inflating_pipe)
    ~encoder:(new deflating_pipe ~level:6)
    ()

Regards,
Matías.

Re: [Ocamlnet-devel] Catching 404

[Pierre-Alexandre V. <ont...@gm...>]

Sorry Gerd, I was unclear enough, I'm talking in a Netcgi context...
So : How I can catch 404 call to the server,a nd is there a way to make url
rewriting (this question is less important as I'll  probably use an inverse
proxy like nginx)

2012/9/19 Gerd Stolpmann <in...@ge...>

> Am 18.09.2012 17:31:17 schrieb(en) Pierre-Alexandre Voye:
>
>  Hello,
>>  is it possible to catch a 404 with ocamlnet ?
>> Use case example : if the page doesn't exists then, it give a way to
>> create/edit the unknown page.
>> It seems, I would have to rewrite is_error_response according to
>> http://projects.camlcity.org/**projects/dl/ocamlnet-3.2.1/**
>> doc/html-main/Http_fs.html#**VALis_error_response<http://projects.camlcity.org/projects/dl/ocamlnet-3.2.1/doc/html-main/Http_fs.html#VALis_error_response>
>> Is it the case ?
>>
>
> Hmm, Http_fs is a very thin layer on top of Http_client. HTTP codes are
> translated to Unix codes where it makes sense, so a 404 error will appear
> as Unix_error(ENOENT,_,_).
>
> If you need more control, you should use Http_client directly (the
> pipeline is accessible, so you can do things like
>
> let url = http_fs # translate "/the/path/to/the/file" in
> let call_obj = new Http_client.get url in
> http_fs # pipeline # add call_obj;
> http_fs # pipeline # run();
> let code = call_obj # response_status_code in
> ...
>
> ).
>
> So far, there is no way to override is_error_code, so you would have to
> copy http_fs.ml to your own code tree.
>
>
>  In the same way, is it possible to configure url rewriting with ocamlnet ?
>>
>
> I am not exactly sure what you mean. So far I see this is mostly a
> question when redirects are followed. Http_client is able to follow
> redirects (and this is configurable, you can have a rewriting function
> here). Http_fs normally uses the default, so redirects are followed without
> rewriting, and it is not possible to change this (this is a per-request
> setting in Http_client, and you cannot get the request object
> unfortunately).
>
> So, I'd also recommend here to use Http_client directly.
>
> Gerd
>
>
>
>> Regards,
>>
>> Pierre-Alexandre
>>
>> --
>> ---------------------
>> https://twitter.com/#!/**ontologiae/ <https://twitter.com/#!/ontologiae/>
>> http://linuxfr.org/users/**montaigne <http://linuxfr.org/users/montaigne>
>>
>>
> ------Zitierte Anlage------
>
>> ------------------------------**------------------------------**
>> ------------------
>> Live Security Virtual Conference
>> Exclusive live event will cover all the ways today's security and
>> threat landscape has changed and how IT managers can respond. Discussions
>> will include endpoint security, mobile security and the latest in malware
>> threats. http://www.accelacomm.com/jaw/**sfrnl04242012/114/50122263/<http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/>
>>
>
> ------Zitierte Anlage------
>
>> ______________________________**_________________
>> Ocamlnet-devel mailing list
>> Ocamlnet-devel@lists.**sourceforge.net<Oca...@li...>
>> https://lists.sourceforge.net/**lists/listinfo/ocamlnet-devel<https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel>
>>
>>
>
>
> --
> ------------------------------**------------------------------
> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
> Creator of GODI and camlcity.org.
> Contact details:        http://www.camlcity.org/**contact.html<http://www.camlcity.org/contact.html>
> Company homepage:       http://www.gerd-stolpmann.de
> ------------------------------**------------------------------




-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

Re: [Ocamlnet-devel] Catching 404

[Gerd S. <in...@ge...>]

Am 18.09.2012 17:31:17 schrieb(en) Pierre-Alexandre Voye:
> Hello,
>  is it possible to catch a 404 with ocamlnet ?
> Use case example : if the page doesn't exists then, it give a way to
> create/edit the unknown page.
> It seems, I would have to rewrite is_error_response according to
> http://projects.camlcity.org/projects/dl/ocamlnet-3.2.1/doc/html-main/Http_fs.html#VALis_error_response
> Is it the case ?

Hmm, Http_fs is a very thin layer on top of Http_client. HTTP codes are  
translated to Unix codes where it makes sense, so a 404 error will  
appear as Unix_error(ENOENT,_,_).

If you need more control, you should use Http_client directly (the  
pipeline is accessible, so you can do things like

let url = http_fs # translate "/the/path/to/the/file" in
let call_obj = new Http_client.get url in
http_fs # pipeline # add call_obj;
http_fs # pipeline # run();
let code = call_obj # response_status_code in
...

).

So far, there is no way to override is_error_code, so you would have to  
copy http_fs.ml to your own code tree.

> In the same way, is it possible to configure url rewriting with  
> ocamlnet ?

I am not exactly sure what you mean. So far I see this is mostly a  
question when redirects are followed. Http_client is able to follow  
redirects (and this is configurable, you can have a rewriting function  
here). Http_fs normally uses the default, so redirects are followed  
without rewriting, and it is not possible to change this (this is a  
per-request setting in Http_client, and you cannot get the request  
object unfortunately).

So, I'd also recommend here to use Http_client directly.

Gerd

> 
> Regards,
> 
> Pierre-Alexandre
> 
> --
> ---------------------
> https://twitter.com/#!/ontologiae/
> http://linuxfr.org/users/montaigne
> 

------Zitierte Anlage------
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond.  
> Discussions
> will include endpoint security, mobile security and the latest in  
> malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

------Zitierte Anlage------
> _______________________________________________
> Ocamlnet-devel mailing list
> Oca...@li...
> https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
> 



-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------

[Ocamlnet-devel] Catching 404

[Pierre-Alexandre V. <ont...@gm...>]

Hello,
 is it possible to catch a 404 with ocamlnet ?
Use case example : if the page doesn't exists then, it give a way to
create/edit the unknown page.
It seems, I would have to rewrite is_error_response according to
http://projects.camlcity.org/projects/dl/ocamlnet-3.2.1/doc/html-main/Http_fs.html#VALis_error_response
Is it the case ?

In the same way, is it possible to configure url rewriting with ocamlnet ?

Regards,

Pierre-Alexandre

-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

[Ocamlnet-devel] [ANN] Ocamlnet-3.6

[Gerd S. <in...@ge...>]

Hi,

I've just released Ocamlnet-3.6. Bigger changes (but not really big):

  - New abstraction for anonymous semaphores, improving the  
compatibility with OS X
  - PCRE is no longer needed, but it is still possible to build  
Ocamlnet with PCRE as regular expression engine

For the smaller changes, see the ChangeLog.

Get Ocamlnet, read the manual etc. from  
http://projects.camlcity.org/projects/ocamlnet.html

GODI: Because of the PCRE change, I currently don't want to update GODI  
for Ocaml-3.12. In the upcoming version for Ocaml-4.00, I've released  
Ocamlnet-3.6. There is now a new package godi-ocamlnet-pcre for those  
users who need the Netstring_pcre module. Apparently, a number of  
packages are broken because of this (at least a lot of my own), but  
it's still a few days until 4.00 is out.

Gerd
-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------

[Ocamlnet-devel] [ANN] Official OCaml-4.00 patch for ocamlnet

[Gerd S. <in...@ge...>]

Hi,

as there are some incomplete patches around, here are the official ones  
(against ocamlnet-1.5.1):

https://godirepo.camlcity.org/svn/godi-build/trunk/godi/godi-ocamlnet/patches/patch-ac-netsys-and-ocaml4
https://godirepo.camlcity.org/svn/godi-build/trunk/godi/godi-ocamlnet/patches/patch-ad-netstring-and-ocaml4

Note that type definitions like

type open_flag1 = Unix.open_flag =
      O_RDONLY | O_WRONLY | O_RDWR | O_NONBLOCK | O_APPEND | O_CREAT |  
O_TRUNC
    | O_EXCL | O_NOCTTY | O_DSYNC | O_SYNC | O_RSYNC

are not meaningless in the presence of stub libraries, and work like  
assertions for types. The addition of O_SHARE_DELETE makes this fail.  
Very good, because the corresponding list in the stub needs to be  
extended, too (which the other patches do not do, and this can lead to  
segfaults).

There will also be a release of Ocamlnet soon, but this overlaps  
currently with the final QA testing of the regular development cycle.

Gerd
-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------

Re: [Ocamlnet-devel] Problem with sendmail : invalid sender

[Pierre-Alexandre V. <ont...@gm...>]

It works !
I modified the myorigin with a domain which exists, and event the email
ont...@co... (the myorigin I seted) does not exists,
smtp.orange.fraccepts it.

Thank you

PS : Good trick, Shame on me I didn't thought to do that !

2012/5/4 Gerd Stolpmann <in...@ge...>

>
> > I did the test and it doesn't work.
> > I don't understand why it works, when I type this in the terminal :
> >
> > printf "Subject: TestnHello" | sendmail -f ont...@gm...
> > ont...@gm...
> > Is there a way to see what command line is emitted ?
>
> Create a script "fakemailer":
>
> #! /bin/sh
>
> echo "Arguments: $@"
> cat
>
> and then Netsendmail.sendmail ~mailer:"./fakemailer" msg
>
> Note that the -f switch of sendmail does not set the FROM line within the
> message but the envelope sender address. The envelope sender is the
> address that counts. Normally, sendmail extracts it from the FROM line, so
> this is another thing that could go wrong. From the log file I see that
> the envolope sender is "ontologiae@ontologiae.local" (watch out for
> "from="). This looks like a rewritten address (or simply the myorigin
> parameter from main.cf). Such an address can in deed be problematic
> because "local" is no valid domain.
>
> Gerd
>
> > 2012/5/4 Gerd Stolpmann <in...@ge...>
> >
> >> Just tried it on my local postfix installation, and did not get this
> >> error. In your case, smtp.orange.fr bounces the mail. I guess they have
> >> some additional filters in place to make spamming under wrong name
> >> impossible. Maybe these filters do not work correctly.
> >>
> >> You can experiment: If the name part of from_addr only consists of
> >> US-ASCII chars, a simpler header is created (without encoding) - maybe
> >> this is more compatible with the gateway.
> >>
> >> Gerd
> >>
> >> > Hi list,
> >> >
> >> > I'm trying to send a mail via Netsendmail.
> >> > The library is very easy to handle, so here's my code (copied from the
> >> > tutorial) :
> >> >
> >> > let msg = Netsendmail.compose
> >> >   ~in_charset:`Enc_iso885915
> >> >   ~out_charset:`Enc_iso885915
> >> >   ~from_addr:("Heinz Dräger", "mon...@fr...")
> >> >   ~to_addrs:[("Marion Schröder", "ont...@gm...")]
> >> >   ~content_type:("text/html", ["charset", Mimestring.mk_param
> >> > "ISO-8859-1"])
> >> >   ~subject:"Geschäftlich"
> >> >   "<html><body>Verkaufe Teddy-Bären für
> >> 100&euro;</body></html>";;
> >> >
> >> > Netsendmail.sendmail ~mailer:"/usr/sbin/sendmail" msg ;;
> >> >
> >> > So it actually sends an email, but it does not takes the "from" field
> >> :
> >> >
> >> > May  4 14:44:52 ontologiae postfix/master[79734]: daemon started --
> >> > version
> >> > 2.8.4, configuration /etc/postfix
> >> > May  4 14:44:52 ontologiae postfix/pickup[79735]: 0E067B55B42: uid=501
> >> > from=<ontologiae>
> >> > May  4 14:44:52 ontologiae postfix/cleanup[79737]: 0E067B55B42:
> >> > message-id=<20120504124452.0E067B55B42@ontologiae.local>
> >> > May  4 14:44:52 ontologiae postfix/qmgr[79736]: 0E067B55B42:
> >> > from=<ontologiae@ontologiae.local>, size=631, nrcpt=1 (queue active)
> >> > May  4 14:44:53 ontologiae postfix/smtp[79739]: 0E067B55B42: to=<
> >> > ont...@gm...>, relay=smtp.orange.fr[193.252.22.64]:25,
> >> delay=1.6,
> >> > delays=0.28/0.01/0.3/0.97, dsn=5.1.0, status=bounced (host
> >> > smtp.orange.fr[193.252.22.64]
> >> > said: 501 5.1.0 Emetteur invalide. Invalid Sender. OFR203_405 [405]
> >> (in
> >> > reply to MAIL FROM command))
> >> > May  4 14:44:53 ontologiae postfix/cleanup[79737]: 9DC05B55B46:
> >> > message-id=<20120504124453.9DC05B55B46@ontologiae.local>
> >> > May  4 14:44:53 ontologiae postfix/bounce[79742]: 0E067B55B42: sender
> >> > non-delivery notification: 9DC05B55B46
> >> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: from=<>,
> >> > size=2652, nrcpt=1 (queue active)
> >> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 0E067B55B42: removed
> >> > May  4 14:44:53 ontologiae postfix/local[79743]: 9DC05B55B46:
> >> > to=<ontologiae@ontologiae.local>, relay=local, delay=0.01,
> >> > delays=0/0.01/0/0, dsn=2.0.0, status=sent (delivered to mailbox)
> >> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: removed
> >> >
> >> >
> >> > Is there a solution ?
> >> >
> >> > In advance, thanks.
> >> >
> >> > Best regards,
> >> >
> >> > Pierre-Alexandre
> >> > --
> >> > ---------------------
> >> > https://twitter.com/#!/ontologiae/
> >> > http://linuxfr.org/users/montaigne
> >> >
> >>
> ------------------------------------------------------------------------------
> >> > Live Security Virtual Conference
> >> > Exclusive live event will cover all the ways today's security and
> >> > threat landscape has changed and how IT managers can respond.
> >> Discussions
> >> > will include endpoint security, mobile security and the latest in
> >> malware
> >> > threats.
> >> >
> >>
> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
> >> > Ocamlnet-devel mailing list
> >> > Oca...@li...
> >> > https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
> >> >
> >>
> >>
> >> --
> >> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
> >> Creator of GODI and camlcity.org.
> >> Contact details:        http://www.camlcity.org/contact.html
> >> Company homepage:       http://www.gerd-stolpmann.de
> >> *** Searching for new projects! Need consulting for system
> >> *** programming in Ocaml? Gerd Stolpmann can help you.
> >>
> >>
> >
> >
> > --
> > ---------------------
> > https://twitter.com/#!/ontologiae/
> > http://linuxfr.org/users/montaigne
> >
> ------------------------------------------------------------------------------
> > Live Security Virtual Conference
> > Exclusive live event will cover all the ways today's security and
> > threat landscape has changed and how IT managers can respond. Discussions
> > will include endpoint security, mobile security and the latest in malware
> > threats.
> >
> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
> > Ocamlnet-devel mailing list
> > Oca...@li...
> > https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
> >
>
>
> --
> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
> Creator of GODI and camlcity.org.
> Contact details:        http://www.camlcity.org/contact.html
> Company homepage:       http://www.gerd-stolpmann.de
> *** Searching for new projects! Need consulting for system
> *** programming in Ocaml? Gerd Stolpmann can help you.
>
>


-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

Re: [Ocamlnet-devel] Problem with sendmail : invalid sender

[Gerd S. <in...@ge...>]

> I did the test and it doesn't work.
> I don't understand why it works, when I type this in the terminal :
>
> printf "Subject: TestnHello" | sendmail -f ont...@gm...
> ont...@gm...
> Is there a way to see what command line is emitted ?

Create a script "fakemailer":

#! /bin/sh

echo "Arguments: $@"
cat

and then Netsendmail.sendmail ~mailer:"./fakemailer" msg

Note that the -f switch of sendmail does not set the FROM line within the
message but the envelope sender address. The envelope sender is the
address that counts. Normally, sendmail extracts it from the FROM line, so
this is another thing that could go wrong. From the log file I see that
the envolope sender is "ontologiae@ontologiae.local" (watch out for
"from="). This looks like a rewritten address (or simply the myorigin
parameter from main.cf). Such an address can in deed be problematic
because "local" is no valid domain.

Gerd

> 2012/5/4 Gerd Stolpmann <in...@ge...>
>
>> Just tried it on my local postfix installation, and did not get this
>> error. In your case, smtp.orange.fr bounces the mail. I guess they have
>> some additional filters in place to make spamming under wrong name
>> impossible. Maybe these filters do not work correctly.
>>
>> You can experiment: If the name part of from_addr only consists of
>> US-ASCII chars, a simpler header is created (without encoding) - maybe
>> this is more compatible with the gateway.
>>
>> Gerd
>>
>> > Hi list,
>> >
>> > I'm trying to send a mail via Netsendmail.
>> > The library is very easy to handle, so here's my code (copied from the
>> > tutorial) :
>> >
>> > let msg = Netsendmail.compose
>> >   ~in_charset:`Enc_iso885915
>> >   ~out_charset:`Enc_iso885915
>> >   ~from_addr:("Heinz Dräger", "mon...@fr...")
>> >   ~to_addrs:[("Marion Schröder", "ont...@gm...")]
>> >   ~content_type:("text/html", ["charset", Mimestring.mk_param
>> > "ISO-8859-1"])
>> >   ~subject:"Geschäftlich"
>> >   "<html><body>Verkaufe Teddy-Bären für
>> 100&euro;</body></html>";;
>> >
>> > Netsendmail.sendmail ~mailer:"/usr/sbin/sendmail" msg ;;
>> >
>> > So it actually sends an email, but it does not takes the "from" field
>> :
>> >
>> > May  4 14:44:52 ontologiae postfix/master[79734]: daemon started --
>> > version
>> > 2.8.4, configuration /etc/postfix
>> > May  4 14:44:52 ontologiae postfix/pickup[79735]: 0E067B55B42: uid=501
>> > from=<ontologiae>
>> > May  4 14:44:52 ontologiae postfix/cleanup[79737]: 0E067B55B42:
>> > message-id=<20120504124452.0E067B55B42@ontologiae.local>
>> > May  4 14:44:52 ontologiae postfix/qmgr[79736]: 0E067B55B42:
>> > from=<ontologiae@ontologiae.local>, size=631, nrcpt=1 (queue active)
>> > May  4 14:44:53 ontologiae postfix/smtp[79739]: 0E067B55B42: to=<
>> > ont...@gm...>, relay=smtp.orange.fr[193.252.22.64]:25,
>> delay=1.6,
>> > delays=0.28/0.01/0.3/0.97, dsn=5.1.0, status=bounced (host
>> > smtp.orange.fr[193.252.22.64]
>> > said: 501 5.1.0 Emetteur invalide. Invalid Sender. OFR203_405 [405]
>> (in
>> > reply to MAIL FROM command))
>> > May  4 14:44:53 ontologiae postfix/cleanup[79737]: 9DC05B55B46:
>> > message-id=<20120504124453.9DC05B55B46@ontologiae.local>
>> > May  4 14:44:53 ontologiae postfix/bounce[79742]: 0E067B55B42: sender
>> > non-delivery notification: 9DC05B55B46
>> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: from=<>,
>> > size=2652, nrcpt=1 (queue active)
>> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 0E067B55B42: removed
>> > May  4 14:44:53 ontologiae postfix/local[79743]: 9DC05B55B46:
>> > to=<ontologiae@ontologiae.local>, relay=local, delay=0.01,
>> > delays=0/0.01/0/0, dsn=2.0.0, status=sent (delivered to mailbox)
>> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: removed
>> >
>> >
>> > Is there a solution ?
>> >
>> > In advance, thanks.
>> >
>> > Best regards,
>> >
>> > Pierre-Alexandre
>> > --
>> > ---------------------
>> > https://twitter.com/#!/ontologiae/
>> > http://linuxfr.org/users/montaigne
>> >
>> ------------------------------------------------------------------------------
>> > Live Security Virtual Conference
>> > Exclusive live event will cover all the ways today's security and
>> > threat landscape has changed and how IT managers can respond.
>> Discussions
>> > will include endpoint security, mobile security and the latest in
>> malware
>> > threats.
>> >
>> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
>> > Ocamlnet-devel mailing list
>> > Oca...@li...
>> > https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
>> >
>>
>>
>> --
>> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
>> Creator of GODI and camlcity.org.
>> Contact details:        http://www.camlcity.org/contact.html
>> Company homepage:       http://www.gerd-stolpmann.de
>> *** Searching for new projects! Need consulting for system
>> *** programming in Ocaml? Gerd Stolpmann can help you.
>>
>>
>
>
> --
> ---------------------
> https://twitter.com/#!/ontologiae/
> http://linuxfr.org/users/montaigne
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats.
> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
> Ocamlnet-devel mailing list
> Oca...@li...
> https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
>


-- 
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
*** Searching for new projects! Need consulting for system
*** programming in Ocaml? Gerd Stolpmann can help you.

Re: [Ocamlnet-devel] Problem with sendmail : invalid sender

[Pierre-Alexandre V. <ont...@gm...>]

I did the test and it doesn't work.
I don't understand why it works, when I type this in the terminal :

printf "Subject: TestnHello" | sendmail -f ont...@gm...
ont...@gm...
Is there a way to see what command line is emitted ?

2012/5/4 Gerd Stolpmann <in...@ge...>

> Just tried it on my local postfix installation, and did not get this
> error. In your case, smtp.orange.fr bounces the mail. I guess they have
> some additional filters in place to make spamming under wrong name
> impossible. Maybe these filters do not work correctly.
>
> You can experiment: If the name part of from_addr only consists of
> US-ASCII chars, a simpler header is created (without encoding) - maybe
> this is more compatible with the gateway.
>
> Gerd
>
> > Hi list,
> >
> > I'm trying to send a mail via Netsendmail.
> > The library is very easy to handle, so here's my code (copied from the
> > tutorial) :
> >
> > let msg = Netsendmail.compose
> >   ~in_charset:`Enc_iso885915
> >   ~out_charset:`Enc_iso885915
> >   ~from_addr:("Heinz Dräger", "mon...@fr...")
> >   ~to_addrs:[("Marion Schröder", "ont...@gm...")]
> >   ~content_type:("text/html", ["charset", Mimestring.mk_param
> > "ISO-8859-1"])
> >   ~subject:"Geschäftlich"
> >   "<html><body>Verkaufe Teddy-Bären für 100&euro;</body></html>";;
> >
> > Netsendmail.sendmail ~mailer:"/usr/sbin/sendmail" msg ;;
> >
> > So it actually sends an email, but it does not takes the "from" field :
> >
> > May  4 14:44:52 ontologiae postfix/master[79734]: daemon started --
> > version
> > 2.8.4, configuration /etc/postfix
> > May  4 14:44:52 ontologiae postfix/pickup[79735]: 0E067B55B42: uid=501
> > from=<ontologiae>
> > May  4 14:44:52 ontologiae postfix/cleanup[79737]: 0E067B55B42:
> > message-id=<20120504124452.0E067B55B42@ontologiae.local>
> > May  4 14:44:52 ontologiae postfix/qmgr[79736]: 0E067B55B42:
> > from=<ontologiae@ontologiae.local>, size=631, nrcpt=1 (queue active)
> > May  4 14:44:53 ontologiae postfix/smtp[79739]: 0E067B55B42: to=<
> > ont...@gm...>, relay=smtp.orange.fr[193.252.22.64]:25,
> delay=1.6,
> > delays=0.28/0.01/0.3/0.97, dsn=5.1.0, status=bounced (host
> > smtp.orange.fr[193.252.22.64]
> > said: 501 5.1.0 Emetteur invalide. Invalid Sender. OFR203_405 [405] (in
> > reply to MAIL FROM command))
> > May  4 14:44:53 ontologiae postfix/cleanup[79737]: 9DC05B55B46:
> > message-id=<20120504124453.9DC05B55B46@ontologiae.local>
> > May  4 14:44:53 ontologiae postfix/bounce[79742]: 0E067B55B42: sender
> > non-delivery notification: 9DC05B55B46
> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: from=<>,
> > size=2652, nrcpt=1 (queue active)
> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 0E067B55B42: removed
> > May  4 14:44:53 ontologiae postfix/local[79743]: 9DC05B55B46:
> > to=<ontologiae@ontologiae.local>, relay=local, delay=0.01,
> > delays=0/0.01/0/0, dsn=2.0.0, status=sent (delivered to mailbox)
> > May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: removed
> >
> >
> > Is there a solution ?
> >
> > In advance, thanks.
> >
> > Best regards,
> >
> > Pierre-Alexandre
> > --
> > ---------------------
> > https://twitter.com/#!/ontologiae/
> > http://linuxfr.org/users/montaigne
> >
> ------------------------------------------------------------------------------
> > Live Security Virtual Conference
> > Exclusive live event will cover all the ways today's security and
> > threat landscape has changed and how IT managers can respond. Discussions
> > will include endpoint security, mobile security and the latest in malware
> > threats.
> >
> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
> > Ocamlnet-devel mailing list
> > Oca...@li...
> > https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
> >
>
>
> --
> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
> Creator of GODI and camlcity.org.
> Contact details:        http://www.camlcity.org/contact.html
> Company homepage:       http://www.gerd-stolpmann.de
> *** Searching for new projects! Need consulting for system
> *** programming in Ocaml? Gerd Stolpmann can help you.
>
>


-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

Re: [Ocamlnet-devel] Problem with sendmail : invalid sender

[Gerd S. <in...@ge...>]

Just tried it on my local postfix installation, and did not get this
error. In your case, smtp.orange.fr bounces the mail. I guess they have
some additional filters in place to make spamming under wrong name
impossible. Maybe these filters do not work correctly.

You can experiment: If the name part of from_addr only consists of
US-ASCII chars, a simpler header is created (without encoding) - maybe
this is more compatible with the gateway.

Gerd

> Hi list,
>
> I'm trying to send a mail via Netsendmail.
> The library is very easy to handle, so here's my code (copied from the
> tutorial) :
>
> let msg = Netsendmail.compose
>   ~in_charset:`Enc_iso885915
>   ~out_charset:`Enc_iso885915
>   ~from_addr:("Heinz Dräger", "mon...@fr...")
>   ~to_addrs:[("Marion Schröder", "ont...@gm...")]
>   ~content_type:("text/html", ["charset", Mimestring.mk_param
> "ISO-8859-1"])
>   ~subject:"Geschäftlich"
>   "<html><body>Verkaufe Teddy-Bären für 100&euro;</body></html>";;
>
> Netsendmail.sendmail ~mailer:"/usr/sbin/sendmail" msg ;;
>
> So it actually sends an email, but it does not takes the "from" field :
>
> May  4 14:44:52 ontologiae postfix/master[79734]: daemon started --
> version
> 2.8.4, configuration /etc/postfix
> May  4 14:44:52 ontologiae postfix/pickup[79735]: 0E067B55B42: uid=501
> from=<ontologiae>
> May  4 14:44:52 ontologiae postfix/cleanup[79737]: 0E067B55B42:
> message-id=<20120504124452.0E067B55B42@ontologiae.local>
> May  4 14:44:52 ontologiae postfix/qmgr[79736]: 0E067B55B42:
> from=<ontologiae@ontologiae.local>, size=631, nrcpt=1 (queue active)
> May  4 14:44:53 ontologiae postfix/smtp[79739]: 0E067B55B42: to=<
> ont...@gm...>, relay=smtp.orange.fr[193.252.22.64]:25, delay=1.6,
> delays=0.28/0.01/0.3/0.97, dsn=5.1.0, status=bounced (host
> smtp.orange.fr[193.252.22.64]
> said: 501 5.1.0 Emetteur invalide. Invalid Sender. OFR203_405 [405] (in
> reply to MAIL FROM command))
> May  4 14:44:53 ontologiae postfix/cleanup[79737]: 9DC05B55B46:
> message-id=<20120504124453.9DC05B55B46@ontologiae.local>
> May  4 14:44:53 ontologiae postfix/bounce[79742]: 0E067B55B42: sender
> non-delivery notification: 9DC05B55B46
> May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: from=<>,
> size=2652, nrcpt=1 (queue active)
> May  4 14:44:53 ontologiae postfix/qmgr[79736]: 0E067B55B42: removed
> May  4 14:44:53 ontologiae postfix/local[79743]: 9DC05B55B46:
> to=<ontologiae@ontologiae.local>, relay=local, delay=0.01,
> delays=0/0.01/0/0, dsn=2.0.0, status=sent (delivered to mailbox)
> May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: removed
>
>
> Is there a solution ?
>
> In advance, thanks.
>
> Best regards,
>
> Pierre-Alexandre
> --
> ---------------------
> https://twitter.com/#!/ontologiae/
> http://linuxfr.org/users/montaigne
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats.
> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
> Ocamlnet-devel mailing list
> Oca...@li...
> https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
>


-- 
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
*** Searching for new projects! Need consulting for system
*** programming in Ocaml? Gerd Stolpmann can help you.

[Ocamlnet-devel] Problem with sendmail : invalid sender

[Pierre-Alexandre V. <ont...@gm...>]

Hi list,

I'm trying to send a mail via Netsendmail.
The library is very easy to handle, so here's my code (copied from the
tutorial) :

let msg = Netsendmail.compose
  ~in_charset:`Enc_iso885915
  ~out_charset:`Enc_iso885915
  ~from_addr:("Heinz Dräger", "mon...@fr...")
  ~to_addrs:[("Marion Schröder", "ont...@gm...")]
  ~content_type:("text/html", ["charset", Mimestring.mk_param "ISO-8859-1"])
  ~subject:"Geschäftlich"
  "<html><body>Verkaufe Teddy-Bären für 100&euro;</body></html>";;

Netsendmail.sendmail ~mailer:"/usr/sbin/sendmail" msg ;;

So it actually sends an email, but it does not takes the "from" field :

May  4 14:44:52 ontologiae postfix/master[79734]: daemon started -- version
2.8.4, configuration /etc/postfix
May  4 14:44:52 ontologiae postfix/pickup[79735]: 0E067B55B42: uid=501
from=<ontologiae>
May  4 14:44:52 ontologiae postfix/cleanup[79737]: 0E067B55B42:
message-id=<20120504124452.0E067B55B42@ontologiae.local>
May  4 14:44:52 ontologiae postfix/qmgr[79736]: 0E067B55B42:
from=<ontologiae@ontologiae.local>, size=631, nrcpt=1 (queue active)
May  4 14:44:53 ontologiae postfix/smtp[79739]: 0E067B55B42: to=<
ont...@gm...>, relay=smtp.orange.fr[193.252.22.64]:25, delay=1.6,
delays=0.28/0.01/0.3/0.97, dsn=5.1.0, status=bounced (host
smtp.orange.fr[193.252.22.64]
said: 501 5.1.0 Emetteur invalide. Invalid Sender. OFR203_405 [405] (in
reply to MAIL FROM command))
May  4 14:44:53 ontologiae postfix/cleanup[79737]: 9DC05B55B46:
message-id=<20120504124453.9DC05B55B46@ontologiae.local>
May  4 14:44:53 ontologiae postfix/bounce[79742]: 0E067B55B42: sender
non-delivery notification: 9DC05B55B46
May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: from=<>,
size=2652, nrcpt=1 (queue active)
May  4 14:44:53 ontologiae postfix/qmgr[79736]: 0E067B55B42: removed
May  4 14:44:53 ontologiae postfix/local[79743]: 9DC05B55B46:
to=<ontologiae@ontologiae.local>, relay=local, delay=0.01,
delays=0/0.01/0/0, dsn=2.0.0, status=sent (delivered to mailbox)
May  4 14:44:53 ontologiae postfix/qmgr[79736]: 9DC05B55B46: removed


Is there a solution ?

In advance, thanks.

Best regards,

Pierre-Alexandre
-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

Re: [Ocamlnet-devel] Strange error with HTTP_Client

[Pierre-Alexandre V. <ont...@gm...>]

I found a fix :
I seted parallelizer with
 Netplex_mp.mp ~keep_fd_open:true ()

It was a lot intuitive (as the way I found it in the doc) and i don't
understand completely why it works now. Seems that Netplex closes
descriptor too early.

Le 6 avril 2012 15:51, Pierre-Alexandre Voye <ont...@gm...> a écrit
:

> Hi,
>
> my app failed systematically when it try to connect to my Alfresco server.
>
>
> Context :
> my program is a cgi http server which make himself certain http call.
> It is runned in a multiprocess manner, and it is interesting to note that
> my problem disappear if you set to multi-thread instead of multiprocess.
>
> In the main I have :
>
>  let nethttpd_factory =  Nethttpd_plex.nethttpd_factory
> ~handlers:handlers () in
>   Netplex_main.startup  parallelizer
>     Netplex_log.logger_factories   (* allow all built-in logging styles *)
>     Netplex_workload.workload_manager_factories (* ... all ways of
> workload management *)
>     [ nethttpd_factory ]           (* make this nethttpd available *)
>     cmdline_cfg
> ;;
>
>
>
> Now, the error
>
>
> Apparently (according to stack trace), it fails on
> Http_client.Convenience.http_get
>
> It doesn't throw an error, but (captured in the error of the CGI which is
> called) :
> Unix.Unix_error(EINVAL, "read (result invalid)", "")
>
> Here's the code where it failed :
>
> let requete_get url          =    Utils.info ("GET "^url) ; try
>
> Http_client.Convenience.http_get    url
> <---------------------------------------
>                                                         with
>
> Http_client.Http_error (nerr,msg) ->
>
> Utils.log2  ("requete_get : "^(string_of_int nerr))
> (Printexc.get_backtrace ()) msg; ""
>
> The associated log with "-debug Http_client" parameter is show in [1]
>
>
> More suprising, I tried to modify requete_get by
>
> let requete_get url          =    Utils.info ("GET "^url) ;
>                                   Utils.execute_command ("curl -s
> '"^url^"'")   (* execute_command : string -> string       execute an unix
> command and return  the result in a string*)
>
> It failed too, but works sometimes and the logs give :
>
> [Fri Apr  6 15:34:35 2012] [debug] [52085:0] Netlog: WARNING track_fd:
> descriptor already tracked as 4(Netsys_posix - create_event)
> [Fri Apr  6 15:34:35 2012] [debug] [52085:0] Netlog: track_fd: tracking
> descriptor 4(Netsys_posix - create_event)
> [Fri Apr  6 15:34:35 2012] [nethttpd] [crit] run: Exception
> Unix.Unix_error(EINVAL, "read (result invalid)", "")
>
> and
> Sys_error("Bad file descriptor");;Raised by primitive operation at file "
> pervasives.ml", line 325, characters 12-32
> Called from file "Utils.ml", line 97, characters 24-41
> <---------------------- in the Utils.execute_command function
>
>
> In advance, thank you for your help
>
> Regards,
>
> Pierre-Alexandre
>
>
> [1] Logs of the first version of function requete_get which use Http_client
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: Got Call 2121!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: pipelining=true persistency=false close_connection=false->false
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2121 -
> postprocessing
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: Shutdown!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: Closing socket!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
> checking remaining pipeline requests
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
> creating direct connection to alfrescodev:8080
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> direct connection to alfrescodev:8080: Connected!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP Connection:
> adding call 2353
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2353:
> initialize transmitter
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - Call
> 2353 - HTTP request: GET
> /alfresco/service/docasu/ui/folder/docs?nodeId=8eb7be3d-0790-493d-a173-0140f7b91df1&alf_ticket=TICKET_9fecb90892eb5703a31ff18793709955bbf550d9
> HTTP/1.1
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: Got Call 2353!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: pipelining=true persistency=false close_connection=false->false
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2353 -
> postprocessing
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: Shutdown!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
> connection: Closing socket!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
> checking remaining pipeline requests
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - Shutdown
> error: Unix.Unix_error(EBADF, "shutdown", "")
> [Fri Apr  6 15:13:07 2012] [nethttpd] [crit] Unix.shutdown (fd 11):
> Exception Unix.Unix_error(EBADF, "shutdown", "")
> [Fri Apr  6 15:13:07 2012] [nethttpd] [crit] Unix.close (fd 11): Exception
> Unix.Unix_error(EBADF, "close", "")
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
> creating direct connection to alfrescodev:8080
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - HTTP
> direct connection to alfrescodev:8080: Connected!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Netlog: WARNING track_fd:
> descriptor already tracked as 4(Netsys_posix - create_event)
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Netlog: track_fd: tracking
> descriptor 4(Http_client - HTTP direct connection to alfrescodev:8080)
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP Connection:
> adding call 2614
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2614:
> initialize transmitter
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - Call 2614
> - HTTP request: GET
> /alfresco/service/docasu/ui/folder/docs?nodeId=f82b2906-9f56-463b-8ff8-f2e0545ed2f8&alf_ticket=TICKET_9fecb90892eb5703a31ff18793709955bbf550d9
> HTTP/1.1
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
> reset
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - HTTP
> connection: Shutdown!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - HTTP
> connection: Closing socket!
> [Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2614 -
> error Http_client.No_reply
>
>
>
>
>
> --
> ---------------------
> https://twitter.com/#!/ontologiae/
> http://linuxfr.org/users/montaigne
>
>


-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

[Ocamlnet-devel] Strange error with HTTP_Client

[Pierre-Alexandre V. <ont...@gm...>]

Hi,

my app failed systematically when it try to connect to my Alfresco server.


Context :
my program is a cgi http server which make himself certain http call.
It is runned in a multiprocess manner, and it is interesting to note that
my problem disappear if you set to multi-thread instead of multiprocess.

In the main I have :

 let nethttpd_factory =  Nethttpd_plex.nethttpd_factory  ~handlers:handlers
() in
  Netplex_main.startup  parallelizer
    Netplex_log.logger_factories   (* allow all built-in logging styles *)
    Netplex_workload.workload_manager_factories (* ... all ways of workload
management *)
    [ nethttpd_factory ]           (* make this nethttpd available *)
    cmdline_cfg
;;



Now, the error


Apparently (according to stack trace), it fails on
Http_client.Convenience.http_get

It doesn't throw an error, but (captured in the error of the CGI which is
called) :
Unix.Unix_error(EINVAL, "read (result invalid)", "")

Here's the code where it failed :

let requete_get url          =    Utils.info ("GET "^url) ; try

Http_client.Convenience.http_get    url
<---------------------------------------
                                                        with

Http_client.Http_error (nerr,msg) ->
                                                                Utils.log2
("requete_get : "^(string_of_int nerr))  (Printexc.get_backtrace ()) msg; ""

The associated log with "-debug Http_client" parameter is show in [1]


More suprising, I tried to modify requete_get by

let requete_get url          =    Utils.info ("GET "^url) ;
                                  Utils.execute_command ("curl -s
'"^url^"'")   (* execute_command : string -> string       execute an unix
command and return  the result in a string*)

It failed too, but works sometimes and the logs give :

[Fri Apr  6 15:34:35 2012] [debug] [52085:0] Netlog: WARNING track_fd:
descriptor already tracked as 4(Netsys_posix - create_event)
[Fri Apr  6 15:34:35 2012] [debug] [52085:0] Netlog: track_fd: tracking
descriptor 4(Netsys_posix - create_event)
[Fri Apr  6 15:34:35 2012] [nethttpd] [crit] run: Exception
Unix.Unix_error(EINVAL, "read (result invalid)", "")

and
Sys_error("Bad file descriptor");;Raised by primitive operation at file "
pervasives.ml", line 325, characters 12-32
Called from file "Utils.ml", line 97, characters 24-41
<---------------------- in the Utils.execute_command function


In advance, thank you for your help

Regards,

Pierre-Alexandre


[1] Logs of the first version of function requete_get which use Http_client
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: Got Call 2121!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: pipelining=true persistency=false close_connection=false->false
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2121 -
postprocessing
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: Shutdown!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: Closing socket!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
checking remaining pipeline requests
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
creating direct connection to alfrescodev:8080
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
direct connection to alfrescodev:8080: Connected!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP Connection:
adding call 2353
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2353:
initialize transmitter
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - Call 2353
- HTTP request: GET
/alfresco/service/docasu/ui/folder/docs?nodeId=8eb7be3d-0790-493d-a173-0140f7b91df1&alf_ticket=TICKET_9fecb90892eb5703a31ff18793709955bbf550d9
HTTP/1.1
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: Got Call 2353!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: pipelining=true persistency=false close_connection=false->false
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2353 -
postprocessing
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: Shutdown!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - HTTP
connection: Closing socket!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
checking remaining pipeline requests
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 11 - Shutdown
error: Unix.Unix_error(EBADF, "shutdown", "")
[Fri Apr  6 15:13:07 2012] [nethttpd] [crit] Unix.shutdown (fd 11):
Exception Unix.Unix_error(EBADF, "shutdown", "")
[Fri Apr  6 15:13:07 2012] [nethttpd] [crit] Unix.close (fd 11): Exception
Unix.Unix_error(EBADF, "close", "")
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
creating direct connection to alfrescodev:8080
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - HTTP
direct connection to alfrescodev:8080: Connected!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Netlog: WARNING track_fd:
descriptor already tracked as 4(Netsys_posix - create_event)
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Netlog: track_fd: tracking
descriptor 4(Http_client - HTTP direct connection to alfrescodev:8080)
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP Connection:
adding call 2614
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2614:
initialize transmitter
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - Call 2614
- HTTP request: GET
/alfresco/service/docasu/ui/folder/docs?nodeId=f82b2906-9f56-463b-8ff8-f2e0545ed2f8&alf_ticket=TICKET_9fecb90892eb5703a31ff18793709955bbf550d9
HTTP/1.1
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: HTTP connection:
reset
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - HTTP
connection: Shutdown!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: FD 4 - HTTP
connection: Closing socket!
[Fri Apr  6 15:13:07 2012] [debug] [49549:0] Http_client: Call 2614 - error
Http_client.No_reply





-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

Re: [Ocamlnet-devel] Unix error 17 on nqueens example

[Gerd S. <in...@ge...>]

Am Donnerstag, den 01.03.2012, 14:57 +0100 schrieb Pierre-Alexandre
Voye:
> Hello, after installing new version 3.5.1, I have new issues :
> /private/tmp/multicore$ ./nqueens -n 13 -mp
> Problem size: 13
> [Thu Mar  1 14:52:48 2012] [netmcore_0] [err] Exception in worker
> process 0: Invalid_argument("format_camlbox (no POSIX semaphores)")

OSX does not support unnamed POSIX semaphores. See e.g.
http://macosx.com/forums/software-programming-web-scripting/49458-semaphores-alternatives.html

What might be possible here is to provide a scheme to replace them with
named semaphores (which are available). The name would just be derived
from the address of the semaphore (e.g. "<prefix>_<address>"). The
not-so-simple part is to ensure that all processes open the semaphores
before their first use (that's the reason why unnamed semaphores work
better here). There could be a process-wide hash table for this, but
this would mean every semaphore operation needs a hash table lookup.

Destruction of semaphores is even worse. In order to do it cleanly, one
needs a special management table (in shared memory or as file), and keep
track which semaphores are still in use.

> Time: 0.025
> |/private/tmp/multicore$ ./nqueens -n 13 -mp2
> Problem size: 13
> Fatal error: exception Unix.Unix_error(34, "lseek", "")

This could be an lseek of shared memory. I've no idea why OSX complains
about this.

Gerd

> |/private/tmp/multicore$ ./nqueens -n 13 -sht
> Problem size: 13
> Fatal error: exception Unix.Unix_error(34, "lseek", "")
> |/private/tmp/multicore$ ./nqueens -n 13 -sht2
> Problem size: 13
> Fatal error: exception Unix.Unix_error(34, "lseek", "")
> |/private/tmp/multicore$ d
> 
> Doc says error 34 (ESPIPE) is "Invalid seek e.g. on a pipe"
> 
> 2012/2/29 Pierre-Alexandre Voye <ont...@gm...>
>         
>         
>         2012/2/29 Gerd Stolpmann <in...@ge...>
>                 Am Sonntag, den 26.02.2012, 23:45 +0100 schrieb
>                 Pierre-Alexandre Voye:
>                 > Hi list,
>                 >
>                 > In my mac (Lion, 10.7) I'm unable to make nqueens
>                 example work with
>                 > -mp or even -sht flags
>                 > Here my attempts :
>                 >
>                 > $ time ./nqueens -n 3 -mp > /dev/null
>                 > Fatal error: exception Unix.Unix_error(17, "",
>                 >
>                 "/Users/ontologiae/Documents/Projets/COWEBO/CoweboDev/Etudes/ocamlnet-3.5/examples/multicore/run_nqueens/netplex.controller/admin")
>                 
>                 
>                 Unix error 17 = ENAMETOOLONG. You can find this out by
>                 running in the
>                 toploop:
>                 
>                 # (Obj.magic 17 : Unix.error);;
>                 - : Unix.error = Unix.ENAMETOOLONG
>         
>         Thank you ! I didn't knew that, it will help me to investigate
>         more by myself. 
>         
>                 
>                 This refers to the path name of the Unix domain
>                 socket. Because of
>                 historical reasons, such paths are limited to 108
>                 bytes. You can work
>                 around by e.g. creating a symlink to your working
>                 directory in /tmp.
>                 
>                 > time ./nqueens -n 3 -sht > /dev/null
>                 > Fatal error: exception Unix.Unix_error(12, "fchmod",
>                 "")
>                 >
>                 >
>                 > $ time ./nqueens -n 3 -mp2 > /dev/null
>                 > Fatal error: exception Unix.Unix_error(12, "fchmod",
>                 "")
>                 >
>                 > Is it a mac/BSD linked problem ?
>                 
>                 
>                 The fchmod problem probably is. fchmod is here called
>                 on a file
>                 descriptor referring to shared memory. This is strange
>                 because fchmod is
>                 part of the POSIX API for shared memory.
>                 
>                 Anyway, I'll just catch EINVAL and ignore it.
>         
>         It's probably a good idea :
>         /tmp/multicore$ ./nqueens -mp
>         Problem size: 8
>         [Wed Feb 29 15:48:46 2012] [netmcore_0] [err] Exception in
>         worker process 0: Unix.Unix_error(EINVAL, "fchmod", "")
>         Time: 0.023 
>         
>                 
>                 Gerd
>                 
>                 >
>                 > Regards
>                 >
>                 > Pierre-Alexandre
>                 >
>                 > --
>                 > ---------------------
>                 > https://twitter.com/#!/ontologiae/
>                 > http://linuxfr.org/users/montaigne
>                 >
>                 
>                 
>                 
>                 
> 
> 
> 
> -- 
> ---------------------
> https://twitter.com/#!/ontologiae/
> http://linuxfr.org/users/montaigne
> 

-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
*** Searching for new projects! Need consulting for system
*** programming in Ocaml? Gerd Stolpmann can help you.
------------------------------------------------------------

Re: [Ocamlnet-devel] Unix error 17 on nqueens example

[Pierre-Alexandre V. <ont...@gm...>]

Hello, after installing new version 3.5.1, I have new issues :
/private/tmp/multicore$ ./nqueens -n 13 -mp
Problem size: 13
[Thu Mar  1 14:52:48 2012] [netmcore_0] [err] Exception in worker process
0: Invalid_argument("format_camlbox (no POSIX semaphores)")
Time: 0.025
|/private/tmp/multicore$ ./nqueens -n 13 -mp2
Problem size: 13
Fatal error: exception Unix.Unix_error(34, "lseek", "")
|/private/tmp/multicore$ ./nqueens -n 13 -sht
Problem size: 13
Fatal error: exception Unix.Unix_error(34, "lseek", "")
|/private/tmp/multicore$ ./nqueens -n 13 -sht2
Problem size: 13
Fatal error: exception Unix.Unix_error(34, "lseek", "")
|/private/tmp/multicore$ d

Doc says error 34 (ESPIPE) is "Invalid seek e.g. on a pipe"

2012/2/29 Pierre-Alexandre Voye <ont...@gm...>

>
>
> 2012/2/29 Gerd Stolpmann <in...@ge...>
>
>> Am Sonntag, den 26.02.2012, 23:45 +0100 schrieb Pierre-Alexandre Voye:
>> > Hi list,
>> >
>> > In my mac (Lion, 10.7) I'm unable to make nqueens example work with
>> > -mp or even -sht flags
>> > Here my attempts :
>> >
>> > $ time ./nqueens -n 3 -mp > /dev/null
>> > Fatal error: exception Unix.Unix_error(17, "",
>> >
>> "/Users/ontologiae/Documents/Projets/COWEBO/CoweboDev/Etudes/ocamlnet-3.5/examples/multicore/run_nqueens/netplex.controller/admin")
>>
>> Unix error 17 = ENAMETOOLONG. You can find this out by running in the
>> toploop:
>>
>> # (Obj.magic 17 : Unix.error);;
>> - : Unix.error = Unix.ENAMETOOLONG
>>
>
> Thank you ! I didn't knew that, it will help me to investigate more by
> myself.
>
>>
>> This refers to the path name of the Unix domain socket. Because of
>> historical reasons, such paths are limited to 108 bytes. You can work
>> around by e.g. creating a symlink to your working directory in /tmp.
>>
>> > time ./nqueens -n 3 -sht > /dev/null
>> > Fatal error: exception Unix.Unix_error(12, "fchmod", "")
>> >
>> >
>> > $ time ./nqueens -n 3 -mp2 > /dev/null
>> > Fatal error: exception Unix.Unix_error(12, "fchmod", "")
>> >
>> > Is it a mac/BSD linked problem ?
>>
>> The fchmod problem probably is. fchmod is here called on a file
>> descriptor referring to shared memory. This is strange because fchmod is
>> part of the POSIX API for shared memory.
>>
>> Anyway, I'll just catch EINVAL and ignore it.
>>
>
> It's probably a good idea :
> /tmp/multicore$ ./nqueens -mp
> Problem size: 8
> [Wed Feb 29 15:48:46 2012] [netmcore_0] [err] Exception in worker process
> 0: Unix.Unix_error(EINVAL, "fchmod", "")
> Time: 0.023
>
>>
>> Gerd
>>
>> >
>> > Regards
>> >
>> > Pierre-Alexandre
>> >
>> > --
>> > ---------------------
>> > https://twitter.com/#!/ontologiae/
>> > http://linuxfr.org/users/montaigne
>> >
>>
>>
>>


-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne

[Ocamlnet-devel] [ANN] OCamlnet-3.5.1

[Gerd S. <in...@ge...>]

There is a patch release available, 3.5.1, fixing a few errors, mostly
build-related:

- Missing symbols on FreeBSD-9
- Missing symbols on Debian Wheezy (probably because of stricter 
  linking)
- posix_spawn is disabled on Mac (for the time being until the problem
  can be further tracked down)
- Mac OS does not like fchmod on shared memory

Gerd


Am Donnerstag, den 23.02.2012, 00:04 +0100 schrieb Gerd Stolpmann:
> Hi,
> 
> it is time for another version of Ocamlnet. The new release 3.5 focuses on
> the system interface, and includes a long list of smaller improvements.
> 
> The system interface, Netsys_posix, is extended by:
>  - Support for POSIX clocks and POSIX timers (with nanosecond resolution)
>  - Netsys_posix.spawn usses now the posix_spawn call if present on the OS
>  - Adding support for pollable events (as e.g. provided by Linux via
>    eventfd). For other OS an emulation is available.
>  - Support for epoll on Linux
> 
> Note that Netsys_posix covers now large parts of POSIX realtime.
> 
> Other improvements:
> 
>  - The code generator for XDR has been improved. A new switch -direct for
>    ocamlrpcgen can be used to generate direct mappings between OCaml values
>    and binary representation (in many cases). Speedups up to 50% are
>    possible for large XDR values.
>  - The new module Uq_mt allows it to access an event-driven resource from
>    several kernel threads (e.g. use an RPC client commonly from several
>    threads).
>  - The thread-safety of Netplex container functions has been improved.
>  - Netmulticore condition variables can now be polled, for better
>    integration into event-based programs.
>  - Option greedy_accepts for Netplex to support servers that accept
>    many connections per second. With this improvement, Netplex can
>    now accept more than 5000 connections/s, and assign them to worker
>    processes.
> 
> Last but not least there is now a new tutorial for Equeue (event systems
> and engines). In particular, the section about combining Ocamlnet with Lwt
> might be interesting.
> 
> For the full list of changes (especially bug fixes), see:
> 
> https://godirepo.camlcity.org/svn/lib-ocamlnet2/trunk/code/ChangeLog
> 
> The download, manual, and other resources:
> 
> http://projects.camlcity.org/projects/ocamlnet.html
> 
> GODI has been updated.
> 
> Gerd

-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
Creator of GODI and camlcity.org.
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
*** Searching for new projects! Need consulting for system
*** programming in Ocaml? Gerd Stolpmann can help you.
------------------------------------------------------------

Re: [Ocamlnet-devel] Unix error 17 on nqueens example

[Pierre-Alexandre V. <ont...@gm...>]

2012/2/29 Gerd Stolpmann <in...@ge...>

> Am Sonntag, den 26.02.2012, 23:45 +0100 schrieb Pierre-Alexandre Voye:
> > Hi list,
> >
> > In my mac (Lion, 10.7) I'm unable to make nqueens example work with
> > -mp or even -sht flags
> > Here my attempts :
> >
> > $ time ./nqueens -n 3 -mp > /dev/null
> > Fatal error: exception Unix.Unix_error(17, "",
> >
> "/Users/ontologiae/Documents/Projets/COWEBO/CoweboDev/Etudes/ocamlnet-3.5/examples/multicore/run_nqueens/netplex.controller/admin")
>
> Unix error 17 = ENAMETOOLONG. You can find this out by running in the
> toploop:
>
> # (Obj.magic 17 : Unix.error);;
> - : Unix.error = Unix.ENAMETOOLONG
>

Thank you ! I didn't knew that, it will help me to investigate more by
myself.

>
> This refers to the path name of the Unix domain socket. Because of
> historical reasons, such paths are limited to 108 bytes. You can work
> around by e.g. creating a symlink to your working directory in /tmp.
>
> > time ./nqueens -n 3 -sht > /dev/null
> > Fatal error: exception Unix.Unix_error(12, "fchmod", "")
> >
> >
> > $ time ./nqueens -n 3 -mp2 > /dev/null
> > Fatal error: exception Unix.Unix_error(12, "fchmod", "")
> >
> > Is it a mac/BSD linked problem ?
>
> The fchmod problem probably is. fchmod is here called on a file
> descriptor referring to shared memory. This is strange because fchmod is
> part of the POSIX API for shared memory.
>
> Anyway, I'll just catch EINVAL and ignore it.
>

It's probably a good idea :
/tmp/multicore$ ./nqueens -mp
Problem size: 8
[Wed Feb 29 15:48:46 2012] [netmcore_0] [err] Exception in worker process
0: Unix.Unix_error(EINVAL, "fchmod", "")
Time: 0.023

>
> Gerd
>
> >
> > Regards
> >
> > Pierre-Alexandre
> >
> > --
> > ---------------------
> > https://twitter.com/#!/ontologiae/
> > http://linuxfr.org/users/montaigne
> >
> >
> ------------------------------------------------------------------------------
> > Virtualization & Cloud Management Using Capacity Planning
> > Cloud computing makes use of virtualization - but cloud computing
> > also focuses on allowing computing to be delivered as a service.
> > http://www.accelacomm.com/jaw/sfnl/114/51521223/
> > _______________________________________________ Ocamlnet-devel mailing
> list Oca...@li...
> https://lists.sourceforge.net/lists/listinfo/ocamlnet-devel
>
> --
> ------------------------------------------------------------
> Gerd Stolpmann, Darmstadt, Germany    ge...@ge...
> Creator of GODI and camlcity.org.
> Contact details:        http://www.camlcity.org/contact.html
> Company homepage:       http://www.gerd-stolpmann.de
> *** Searching for new projects! Need consulting for system
> *** programming in Ocaml? Gerd Stolpmann can help you.
> ------------------------------------------------------------
>
>


-- 
---------------------
https://twitter.com/#!/ontologiae/
http://linuxfr.org/users/montaigne