From sur.pignard at wanadoo.fr  Sat Aug  1 08:45:59 2009
From: sur.pignard at wanadoo.fr (Pascal)
Date: Sat, 1 Aug 2009 08:45:59 +0200
Subject: [Scilab-Dev] Porting Scilab 5.1.1 on Mac OS X 10.4 PPC.
In-Reply-To: <1248882326.28150.2037.camel@zlarin>
References: <9AAE3A2B-4598-4AB9-953F-1B081DF8B794@wanadoo.fr> <1248882326.28150.2037.camel@zlarin>
Message-ID: <8B2F1E58-0822-4A21-871E-46EA6BE02C84@wanadoo.fr>

Hello Sylvestre.

Many thanks for your answer.

I've tried first proposal. I've added "-Xlinker -m" to many makefiles.
I had to add it on several makefiles because LDFLAGS is redefined  
inside makefiles without taken in account possible previous definition:
LDFLAGS =   -L/sw/lib/
instead of:
LDFLAGS =   -L/sw/lib/  $(LDFLAGS)
(maybe same way for other xxFLAGS)
I can send appart the whole multiply defined symbols.

All or almost has been build following:
$ ./configure --prefix=/Users/pascal/Documents/Programmation/ 
scilab-5.1.1-alpha-1/build --without-ocaml --without-fftw --without- 
pvm --without-umfpack --without-matio
$ make
...
$ make install
...
$ export SCI=`pwd`
$ ./build/bin/scilab

Then, I've got a exec error:
./build/bin/scilab: line 592: 12872 Bus error                
"$SCILABBIN" "$@"
Thread: 2

Exception:  EXC_BAD_ACCESS (0x0001)
Codes:      KERN_PROTECTION_FAILURE (0x0002) at 0x00000058

Thread 0:
0   libSystem.B.dylib        	0x9000af48 mach_msg_trap + 8
1   libSystem.B.dylib        	0x9000ae9c mach_msg + 60
2   com.apple.CoreFoundation 	0x907de9ac __CFRunLoopRun + 832
3   com.apple.CoreFoundation 	0x907de2b0 CFRunLoopRunSpecific + 268
4   com.apple.HIToolbox      	0x932beb20 RunCurrentEventLoopInMode + 264
5   com.apple.HIToolbox      	0x932be1b4 ReceiveNextEventCommon + 380
6   com.apple.HIToolbox      	0x932be020  
BlockUntilNextEventMatchingListInMode + 96
7   com.apple.AppKit         	0x937a3734 _DPSNextEvent + 384
8   com.apple.AppKit         	0x937a33f8 -[NSApplication  
nextEventMatchingMask:untilDate:inMode:dequeue:] + 116
9   com.apple.AppKit         	0x9379f93c -[NSApplication run] + 472
10  libawt.jnilib            	0x9fe4d9e8 +[AWTStarter startAWT:] + 1848
11  com.apple.Foundation     	0x92c04d40 __NSFireMainThreadPerform + 276
12  com.apple.CoreFoundation 	0x9080f928 __CFRunLoopPerformPerform + 104
13  com.apple.CoreFoundation 	0x907df300 __CFRunLoopDoSources0 + 384
14  com.apple.CoreFoundation 	0x907de830 __CFRunLoopRun + 452
15  com.apple.CoreFoundation 	0x907de2b0 CFRunLoopRunSpecific + 268
16  com.apple.CoreFoundation 	0x907ed710 CFRunLoopRun + 52
17  libscishell.5.0.1.dylib  	0x00436a1c initMacOSXEnv + 380  
(initMacOSXEnv.c:244)
18  scilab-bin               	0x00004ac0 main + 404 (mainscic.c:173)
19  scilab-bin               	0x000048b0 _start + 760
20  scilab-bin               	0x000045b4 start + 48

Any idea?
Should I try second proposal?

Thanks for your help, Pascal.
http://blady.pagesperso-orange.fr




Le 29 juil. 09 ? 17:45, Sylvestre Ledru a ?crit :

>
>> ...
>> /usr/bin/ld: multiple definitions of symbol _xerbla_
>> .libs/xerbla.o definition of _xerbla_ in section (__TEXT,__text)
>> ../../libs/blas/.libs/libsciblas.dylib(single module) definition of
>> _xerbla_
>> /usr/bin/ld: multiple definitions of symbol _lsame_
>> .libs/lsame.o definition of _lsame_ in section (__TEXT,__text)
>> ../../libs/blas/.libs/libsciblas.dylib(single module) definition of
>> _lsame_
>> collect2: ld returned 1 exit status
>> make[2]: *** [libscilapack.la] Error 1
>> make[1]: *** [all-recursive] Error 1
>> make: *** [all-recursive] Error 1
> Hmm, unusual problem ;)
>
> I would say that you can try three things here:
> * disable the error in the linker (LDFLAGS) when having a multiple
> definition.
> * echo "" > libs/lapack/xerbla.f
> * comment the compilation of xerbla.f in libs/lapack/Makefile.am
> and relaunch the autotools (automake && ./configure <your options>)
>
> Hope this helps,
> Sylvestre
>
>
>> Thanks for giving some help.
>>
>> Regards, Pascal Pignard.
>> http://blady.pagesperso-orange.fr
>>
>>
>>
>>
>>
>>
>
>





From sur.pignard at wanadoo.fr  Sat Aug  1 08:49:16 2009
From: sur.pignard at wanadoo.fr (Pascal)
Date: Sat, 1 Aug 2009 08:49:16 +0200
Subject: [Scilab-Dev] Object orientation in Scilab 6
In-Reply-To: <20090731153726.4c21e6bb@caprera.inria.fr>
References: <20090731153726.4c21e6bb@caprera.inria.fr>
Message-ID: <F18309F4-99D6-41E2-9A9C-C6299806BB16@wanadoo.fr>

Hello Simon.
I'm very new to Scilab, I was quite surprise to discover Fortran  
source codes.
I believed that from the Ada had replaced Fortran even for  
mathematical computing for long time ago.
Moreover Ada has now pretty object oriented features with 2005 standard.
Why not using Ada source codes?

Sorry if this post is off-topic, was just a question ;-)

Regards, Pascal.
http://blady.pagesperso-orange.fr




Le 31 juil. 09 ? 15:37, Simon Lipp a ?crit :

> Hi,
>
> For those ignoring it, I?m in charge of introducing an object
> orientation to Scilab for the version 6.0. This is a first public
> proposal, for discussion.
>
> There?s nothing really fancy, since we agreed on the very big picture
> (dynamic langage with introspection, for the big picture). There?s
> however some points I would like to drag your attention onto :
>  * Are the unusual visibility rules okay, or too unusal ?
>  * "class" is a keyword used for declarating a class, but is also a
> property of instances, and an attribute of slots. Seems reasonble for
> you, or not ?
>  * Visibility of the %remove_* and %install_* methods (public, in this
> proposal)
>  * Should we propose an acces to the current level in methods and
> accessors, like we propose an access to parente level (with the  
> special
> variable super) ? I didn?t made so because similar langages in this
> regard (Objective-C) doesn?t do it, an it doesn?t seems te be a real
> issue to its users.
>  * Should we propose an acces to the sender in methods ? And to sender
> level ? It would allow to emulate visibility in magic methods
> (%get_slot and %set_slot). It seems not very useful to me (and really,
> KISS ;)), but it can be discussed.
>  * Should we propose a way to dynamically bind a function to an  
> object,
> that is, making it a method but not a slot ? It would be like this:
>    function m = a_method_which_transforms_a_function_into_method(f)
> 	m = this.%bind(f)
>    endfunction
>    ...
>    meth = obj.a_method_which_transforms_a_function_into_method(MyFunc)
>    meth()
>    // meth will be treated as a obj method ; it will has access to obj
> with this, and to private and proteced members.
>
>    Same here: seems not very useful, KISS, but let?s discuss.
>  * If a matrix of objects contains several objects, what is allowed,
>    what is forbidden ? In this SEP, only extracting elements is
>    allowed, and methods and properties can be manipulated only with
>    single objects. Good idea at your opinion, or not ?
>
> For whose having read the proposal on the private wiki, this SEP is
> intended to be the same, but simpler. I hope. At least, I tried to  
> make
> it so?
>
> That?s all ;)<objet-SEP.odt>





From sylvestre.ledru at scilab.org  Tue Aug  4 13:49:06 2009
From: sylvestre.ledru at scilab.org (sylvestre.ledru at scilab.org)
Date: Tue, 04 Aug 2009 13:49:06 +0200
Subject: [Scilab-Dev] Porting Scilab 5.1.1 on Mac OS X 10.4 PPC.
Message-ID: <1a6eefc29069cc516744707353012114@localhost>

On Sat, 1 Aug 2009 08:45:59 +0200, Pascal <sur.pignard at wanadoo.fr> wrote:
> Hello Sylvestre.
> 
> Many thanks for your answer.
> 
> I've tried first proposal. I've added "-Xlinker -m" to many makefiles.
> I had to add it on several makefiles because LDFLAGS is redefined  
> inside makefiles without taken in account possible previous definition:
> LDFLAGS =   -L/sw/lib/
> instead of:
> LDFLAGS =   -L/sw/lib/  $(LDFLAGS)
> (maybe same way for other xxFLAGS)
> I can send appart the whole multiply defined symbols.
You should try to add them into the configure.ac and relaunch autoconf.

> All or almost has been build following:
> $ ./configure --prefix=/Users/pascal/Documents/Programmation/ 
> scilab-5.1.1-alpha-1/build --without-ocaml --without-fftw --without- 
> pvm --without-umfpack --without-matio
> $ make
> ...
> $ make install
> ...
> $ export SCI=`pwd`
> $ ./build/bin/scilab
> 
> Then, I've got a exec error:
> ./build/bin/scilab: line 592: 12872 Bus error                
> "$SCILABBIN" "$@"
> Thread: 2
It seems that it is the Java / TckTk sync issue. Adding --without-tk in the
configure should fix your problem.

Sylvestre


From calixte.denizet at ac-rennes.fr  Thu Aug  6 22:26:41 2009
From: calixte.denizet at ac-rennes.fr (Calixte Denizet)
Date: Thu, 06 Aug 2009 22:26:41 +0200
Subject: Toolbox to display xml in console
Message-ID: <1249590401.2447.11.camel@debian>

Hi,

I wrote a new toolbox, SciMathML, to display MathML in Scilab's console
and to convert LaTeX formulas in MathML.
I'm working now to display MathML in graphics windows.

I made new releases of SciMax (few bugs fixed and Maxima compiled with
Clisp managed) and of Overload (one bug fixed).

I put these tbx at sourceforge.net to manage the tracking (impossible at
scilab.org) :
http://sourceforge.net/projects/scilabtbxset/

If someone is interested on making Windows or Mac OS versions please
contact me.

Calixte




From sur.pignard at wanadoo.fr  Fri Aug  7 22:02:11 2009
From: sur.pignard at wanadoo.fr (Pascal)
Date: Fri, 7 Aug 2009 22:02:11 +0200
Subject: [Scilab-Dev] Porting Scilab 5.1.1 on Mac OS X 10.4 PPC.
In-Reply-To: <1a6eefc29069cc516744707353012114@localhost>
References: <1a6eefc29069cc516744707353012114@localhost>
Message-ID: <34EBD299-37FD-49EE-A4E4-7DA40AA3F55F@wanadoo.fr>

Hello Sylvestre.

Following your advice, I've made some modification on configure:
--- configure.0 2009-04-14 09:22:35.000000000 +0200
+++ configure   2009-08-06 23:00:41.000000000 +0200
@@ -10906,7 +10906,7 @@
    *-*-darwin*)
          CFLAGS=" $CFLAGS -I/sw/include "
          CPPFLAGS=" $CPPFLAGS -I/sw/include "
-        LDFLAGS=" $LDFLAGS -L/sw/lib/ "
+        LDFLAGS=" $LDFLAGS -Xlinker -m -L/sw/lib/ "
         ;;
  esac
  ;; # end of the gcc case on the $CC
@@ -11743,6 +11743,10 @@
         # Sun
            machine=sparc
            ;;
+               *Macintosh*)
+       # Mac OS X
+          machine=`uname -p`
+          ;;
      esac

      # Check for known JDK installation layouts
@@ -11868,7 +11872,8 @@
                  ac_java_jvm_jni_lib_runtime_path=$D
                  ac_java_jvm_jni_lib_flags="-L$D -ljvm"

-                D=$ac_java_jvm_dir/jre/lib/i386/server
+#                D=$ac_java_jvm_dir/jre/lib/i386/server
+                D=$ac_java_jvm_dir/../Libraries
                  ac_java_jvm_jni_lib_runtime_path="$ 
{ac_java_jvm_jni_lib_runtime_path}:$D"
                   
ac_java_jvm_jni_lib_flags="$ac_java_jvm_jni_lib_flags -L$D -ljvm"
              fi

and call:
$ ./configure --prefix=/Users/pascal/Documents/Programmation/ 
scilab-5.1.1-alpha-1/build --without-ocaml --without-fftw --without- 
pvm --without-umfpack --without-matio --without-tk
All build.

When I launch Scilab, I've got the error:
$ ./scilab
The native library sciconsole does not exist or cannot be found :
no sciconsole in java.library.path
Current java.library.path is : :/Users/pascal/Documents/Programmation/ 
scilab-5.1.1-alpha-1/build/share/scilab//../../lib/scilab/
Though, the library existe in this directory:
$ ls
...
libsciconsole.5.0.1.dylib  libsciconsole.dylib
libsciconsole.5.dylib      libsciconsole.la
...
Any idea ?

Then I've tried:
$ ./scilab -nwni
Unfortunately, the help is not available in this mode:
-->help
  !--error 10000
help : Le navigateur d'aide n'est pas accessible dans le mode NWNI.
at line      33 of function help called by :
help

Regards, Pascal.
http://blady.pagesperso-orange.fr




Le 4 ao?t 09 ? 13:49, <sylvestre.ledru at scilab.org>  
<sylvestre.ledru at scilab.org> a ?crit :

>
> On Sat, 1 Aug 2009 08:45:59 +0200, Pascal <sur.pignard at wanadoo.fr>  
> wrote:
>> Hello Sylvestre.
>>
>> Many thanks for your answer.
>>
>> I've tried first proposal. I've added "-Xlinker -m" to many  
>> makefiles.
>> I had to add it on several makefiles because LDFLAGS is redefined
>> inside makefiles without taken in account possible previous  
>> definition:
>> LDFLAGS =   -L/sw/lib/
>> instead of:
>> LDFLAGS =   -L/sw/lib/  $(LDFLAGS)
>> (maybe same way for other xxFLAGS)
>> I can send appart the whole multiply defined symbols.
> You should try to add them into the configure.ac and relaunch  
> autoconf.
>
>> All or almost has been build following:
>> $ ./configure --prefix=/Users/pascal/Documents/Programmation/
>> scilab-5.1.1-alpha-1/build --without-ocaml --without-fftw --without-
>> pvm --without-umfpack --without-matio
>> $ make
>> ...
>> $ make install
>> ...
>> $ export SCI=`pwd`
>> $ ./build/bin/scilab
>>
>> Then, I've got a exec error:
>> ./build/bin/scilab: line 592: 12872 Bus error
>> "$SCILABBIN" "$@"
>> Thread: 2
> It seems that it is the Java / TckTk sync issue. Adding --without- 
> tk in the
> configure should fix your problem.
>
> Sylvestre
>





From arsenm2 at rpi.edu  Mon Aug 10 20:29:35 2009
From: arsenm2 at rpi.edu (Matt Arsenault)
Date: Mon, 10 Aug 2009 14:29:35 -0400
Subject: Building scilab from git problem
Message-ID: <1249928975.10175.2.camel@tribble>

Hi, I'm having trouble building scilab from git on 64 bit Ubuntu Jaunty.
After compiling a fresh checkout for a while, it errors with:


[javac] /home/matt/src/scilab/scilab/modules/javasci/examples/others/ExampleEvents2.java

[javac] /home/matt/src/scilab/scilab/modules/javasci/examples/others/ExampleEvents2.java:27: warning: [deprecation] stop() in java.lang.Thread has been deprecated
    [javac]   		this.stop();
    [javac]   		    ^
    [javac] 1 warning

BUILD SUCCESSFUL
Total time: 2 seconds
make[2]: Leaving directory
`/home/matt/src/scilab/scilab/modules/javasci'
make[1]: Leaving directory `/home/matt/src/scilab/scilab/modules'
Making all in contrib
make[1]: Entering directory `/home/matt/src/scilab/scilab/contrib'
make[1]: Nothing to be done for `all'.
make[1]: Leaving directory `/home/matt/src/scilab/scilab/contrib'
make[1]: Entering directory `/home/matt/src/scilab/scilab'
gcc -DHAVE_CONFIG_H -I. -I./modules/core/includes     -g -O2  -DNDEBUG
-fno-stack-protector -DNARROWPROTO -m64 -I./modules/core/includes/
-I./libs/MALLOC/includes/ -I./modules/localization/includes/ -MT
scilab_bin-mainscic.o -MD -MP -MF .deps/scilab_bin-mainscic.Tpo -c -o
scilab_bin-mainscic.o `test -f './modules/shell/src/c/others/mainscic.c'
|| echo './'`./modules/shell/src/c/others/mainscic.c
mv -f .deps/scilab_bin-mainscic.Tpo .deps/scilab_bin-mainscic.Po
/bin/bash ./libtool --tag=CC   --mode=link gcc  -g -O2  -DNDEBUG
-fno-stack-protector -DNARROWPROTO -m64 -I./modules/core/includes/
-I./libs/MALLOC/includes/ -I./modules/localization/includes/   -o
scilab-bin
scilab_bin-mainscic.o ./modules/shell/libscishell.la ./modules/core/libscicore.la ./modules/completion/libscicompletion.la ./modules/localization/libscilocalization.la ./modules/io/libsciio.la  ./modules/libscilab.la -lpthread -ldl -lcurses 
mkdir .libs
gcc -g -O2 -DNDEBUG -fno-stack-protector -DNARROWPROTO -m64
-I./modules/core/includes/ -I./libs/MALLOC/includes/
-I./modules/localization/includes/ -o .libs/scilab-bin
scilab_bin-mainscic.o  ./modules/shell/.libs/libscishell.so ./modules/core/.libs/libscicore.so ./modules/completion/.libs/libscicompletion.so ./modules/localization/.libs/libscilocalization.so ./modules/io/.libs/libsciio.so ./modules/.libs/libscilab.so -lpthread -ldl -lcurses  -Wl,--rpath -Wl,/usr/local/lib/scilab
./modules/core/.libs/libscicore.so: undefined reference to
`gw_output_stream'
./modules/core/.libs/libscicore.so: undefined reference to `print_'
./modules/io/.libs/libsciio.so: undefined reference to `fmttyp_'
./modules/core/.libs/libscicore.so: undefined reference to `prntid_'
/usr/local/lib/scilab/libscipolynomials.so.5: undefined reference to
`formatnumber_'
/usr/local/lib/scilab/libscicacsd.so.5: undefined reference to `dmdspf_'
collect2: ld returned 1 exit status
make[1]: *** [scilab-bin] Error 1
make[1]: Leaving directory `/home/matt/src/scilab/scilab'
make: *** [all-recursive] Error 1




From vincent.lejeune at institutoptique.fr  Wed Aug 12 16:53:51 2009
From: vincent.lejeune at institutoptique.fr (Vincent)
Date: Wed, 12 Aug 2009 16:53:51 +0200
Subject: Howto use default installed maths lib under windows + ACML
Message-ID: <200908121653.51662.vincent.lejeune@institutoptique.fr>

Hi,

I use some routine from Lapack and Blas for my project ; however, these 
routine have several "version", the raw ones from netlib, the one from ATLAS, 
the one from the Intel MKL.

Until now, I have mostly worked under Linux, so I only had to link my project 
against libatlas with a single command flag. However, the window version of 
scilab is shipped with netlib lapack, and the MKL can be installed later.
I'd like my project to call functions in the library installed in the user 
system, without knowing if it is MKL or netlib version. In addition, I don't 
know if I have the right to link against MKL, as it is a proprietary library 
and I don't have the licence...But I think I have the right to link against a 
"agnostic" library that can wrapp call to MKL function.
Is there any way to do so ?

On a different topic, I've found that AMD provide their own version of MKL, 
it's called ACML. Contrary to MKL, it's free and redistribuable,as long as 
their "redistribution agreement" is signed. Is there any plan  to use ACML as 
a "backend" for matrix computation under SciLab ? I don't know how well it 
does perform against MKL or ATLAS, but it can be easier to use than ATLAS(ie 
no config/compile/tune step).

Regards,
Vincent.




From allan.cornet at scilab.org  Mon Aug 17 09:16:08 2009
From: allan.cornet at scilab.org (Allan CORNET)
Date: Mon, 17 Aug 2009 09:16:08 +0200
Subject: [Scilab-Dev] Howto use default installed maths lib under windows + ACML
In-Reply-To: <200908121653.51662.vincent.lejeune@institutoptique.fr>
References: <200908121653.51662.vincent.lejeune@institutoptique.fr>
Message-ID: <001e01ca1f0a$979c0900$c6d41b00$@cornet@scilab.org>

Hi,

If you use blasplus.lib to link to your project , you will have any problems

(licence, choice of the library between mkl or atlas or others).

ACML seems to have some troubles to work correctly
(http://bugzilla.scilab.org/show_bug.cgi?id=4523)

Allan

N.B: please notice that blas (atlas, mkl) library for Scilab it is a subset
of the full blas library.
(we did not embedded all blas functions)

To know exported symbols:  dumpbin /exports blasplus.dll


-----Message d'origine-----
De?: Vincent [mailto:vincent.lejeune at institutoptique.fr] 
Envoy??: mercredi 12 ao?t 2009 16:54
??: dev at lists.scilab.org
Objet?: [Scilab-Dev] Howto use default installed maths lib under windows +
ACML

Hi,

I use some routine from Lapack and Blas for my project ; however, these 
routine have several "version", the raw ones from netlib, the one from
ATLAS, 
the one from the Intel MKL.

Until now, I have mostly worked under Linux, so I only had to link my
project 
against libatlas with a single command flag. However, the window version of 
scilab is shipped with netlib lapack, and the MKL can be installed later.
I'd like my project to call functions in the library installed in the user 
system, without knowing if it is MKL or netlib version. In addition, I don't

know if I have the right to link against MKL, as it is a proprietary library

and I don't have the licence...But I think I have the right to link against
a 
"agnostic" library that can wrapp call to MKL function.
Is there any way to do so ?

On a different topic, I've found that AMD provide their own version of MKL, 
it's called ACML. Contrary to MKL, it's free and redistribuable,as long as 
their "redistribution agreement" is signed. Is there any plan  to use ACML
as 
a "backend" for matrix computation under SciLab ? I don't know how well it 
does perform against MKL or ATLAS, but it can be easier to use than ATLAS(ie

no config/compile/tune step).

Regards,
Vincent.





From vincent.lejeune at institutoptique.fr  Wed Aug 19 00:20:08 2009
From: vincent.lejeune at institutoptique.fr (Vincent)
Date: Wed, 19 Aug 2009 00:20:08 +0200
Subject: How to create a Scilab structure ?
Message-ID: <200908190020.08665.vincent.lejeune@institutoptique.fr>

Hi,

I would like to create a SciLab structure from C code. By scilab structure I 
mean a variable that has some "children", for instance, as outputed by scilab 
:

 a  =
 
   a1: 1
   a2: 2
   a3: 3
   a4: 4

also with structure member instancied by a.a1=1, a.a2=2...
It appears that such object are neither list nor tlist. So I have no idea how 
to produce structure like that...

Regards,
Vincent.




From yann.collette at scilab.org  Wed Aug 19 16:05:31 2009
From: yann.collette at scilab.org (Yann Collette)
Date: Wed, 19 Aug 2009 16:05:31 +0200
Subject: [Scilab-Dev] How to create a Scilab structure ?
In-Reply-To: <200908190020.08665.vincent.lejeune@institutoptique.fr>
References: <200908190020.08665.vincent.lejeune@institutoptique.fr>
Message-ID: <4A8C06AB.1030602@scilab.org>

Vincent a ?crit :
> Hi,
>
> I would like to create a SciLab structure from C code. By scilab structure I 
> mean a variable that has some "children", for instance, as outputed by scilab 
> :
>
>  a  =
>  
>    a1: 1
>    a2: 2
>    a3: 3
>    a4: 4
>
> also with structure member instancied by a.a1=1, a.a2=2...
> It appears that such object are neither list nor tlist. So I have no idea how 
> to produce structure like that...
>
> Regards,
> Vincent.
>
>
>   
A structure is like a mlist:

a = struct();
a.a1 = 1;
a.a2 = 2;
a.a3 = [1 2 3 4];

getfield(1,a) =  !st  dims  a1  a2  a3  !

'st' is the type of the "struct" mlist and corresponds to the values 
returned by 'typeof(a)'
'dims' stores the dimension of a and is the value returned by 'size(a)'
and the remaining things are the labels.

You can access the values like a matlab struct:
disp(a.a1)
or like a Scilab mlist
disp(a('a1'))

To create such a structure using the new api_scilab:

int sci_new_api_mlist_out(char * fname)
{
  int res, i;
  int * p_out_address;
  char  *pstLabels[]   = {"st","dims","a1","a2","a3"};
  double pdblDims[]  = {1,1};
  double pdblA1[] = {1};
  double pdblA2[] = {1}; 
  double pdblA3[] = {1,2,3,4};

  res = createMList(1, 5, &p_out_address);

  // Add the labels at position 1 (in the first field).
  // We can get the labels using getfield(1,mymlist) in Scilab
  createMatrixOfStringInList(1, p_out_address, 1, 1, 5, pstLabels);

  // Add a vector to the 'dims' field
  createMatrixOfDoubleInList(1, p_out_address, 2, 1, 2, pdblDims);

  // Add the 'Ai' fields
  createMatrixOfDoubleInList(1, p_out_address, 3, 1, 1, pdblA1);
  createMatrixOfDoubleInList(1, p_out_address, 3, 1, 1, pdblA2);
  createMatrixOfDoubleInList(1, p_out_address, 3, 1, 4, pdblA3);

  // Return the created mlist
  LhsVar(1) = 1;

  return 0;
}



YC


From sylvestre.ledru at scilab.org  Wed Aug 26 11:36:05 2009
From: sylvestre.ledru at scilab.org (Sylvestre Ledru)
Date: Wed, 26 Aug 2009 11:36:05 +0200
Subject: SEP # : 30 =?UTF-8?Q?=E2=80=93?= Change of header files directory
 in the Scilab distribution
Message-ID: <1251279365.12951.542.camel@zlarin>

Hello,

Here is a SEP about moving Scilab header files used by the dynamic links
from a "per modules hierarchy" to "all in one directory".
Please, comments/critics.

Cheers,
Sylvestre

PS: Serge and Michael, I think this SEP will interest you since we had
many chats about this issue.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: SEP_30_dynamic_link_includes.pdf
Type: application/pdf
Size: 286062 bytes
Desc: not available
URL: <https://lists.scilab.org/pipermail/dev/attachments/20090826/ea284bfd/attachment.pdf>

From BlanchardJ at ieee.org  Wed Aug 26 20:24:40 2009
From: BlanchardJ at ieee.org (Jonathan Blanchard)
Date: Wed, 26 Aug 2009 15:24:40 -0300
Subject: =?UTF-8?Q?Re=3A_=5BScilab=2DDev=5D_SEP_=23_=3A_30_=E2=80=93_Change_of_header_fil?=
	=?UTF-8?Q?es_directory_in_the_Scilab_distribution?=
In-Reply-To: <1251279365.12951.542.camel@zlarin>
References: <1251279365.12951.542.camel@zlarin>
Message-ID: <f3a3df490908261124n711fbe53seafd9b2f9f62e164@mail.gmail.com>

On Wed, Aug 26, 2009 at 6:36 AM, Sylvestre
Ledru<sylvestre.ledru at scilab.org> wrote:
> Hello,
>
> Here is a SEP about moving Scilab header files used by the dynamic links
> from a "per modules hierarchy" to "all in one directory".
> Please, comments/critics.
>
> Cheers,
> Sylvestre
>
> PS: Serge and Michael, I think this SEP will interest you since we had
> many chats about this issue.
>
>

That would be pretty awesome...

Jonathan Blanchard