Nabble - octave-dev

Re: Problem testing nnet's subset()

2009-12-09T13:24:58Z

tir, 08 12 2009 kl. 23:22 +0100, skrev Thomas Weber:

> Hi,
>
> with SVN rev 6619:
> octave:1> test subset
> testing subset
> error: empty error text, probably Ctrl-C --- aborting
> error: called from:
> error: /usr/share/octave/3.2.3/m/testfun/test.m at line 479, column 4
>
>
> I'm using Octave 3.2.3. Maybe something changed in Octave's error handling?

I am not able to reproduce this with the development version. However,
from time to time the tests do fail for me. I guess this is an algorithm
that somehow depends on randomisation?

Søren

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Request of admittance for a Developer of Italian translation for Octave

2009-12-09T13:16:25Z

tir, 08 12 2009 kl. 13:35 -0800, skrev Vortigher Last:
> I plan to translate completly in Italian the Octave manual, man pages, and info of the functions;
> keeping the translation update.

Great! In general we ask for an initial submission before we grant
access to the subversion repository. Could you please send some initial
work to this list?

I don't know much about the translation infrastructure, so I can't
provide much assistance. However, other people on this list are working
active on translations into different languages, so just ask if you have
questions.

Søren

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Registration request (decay_lumi)

2009-12-09T13:12:38Z

Hi

Sorry about the late reply.

man, 07 12 2009 kl. 09:09 +0100, skrev Filippo Piffaretti:
> I would like to submit to the GNU Octave Repository a piece of my work
> that I'm wrote for the analysis of luminescence exponential decay
> signal.

It's always great when new people want to share :-)

> In the present script, the function leasqr and expfit are used to
> analyze coma separated data. Plots and results files are created and
> stored in a "Results" folder created in the working directory.

I have no experience what-so-ever with these types of problems, so take
my comments with a grain of salt; there is a pretty good chance that I'm
mistaken.

I get the impression that your function is more like a final application
rather than a general purpose tool. Do you agree with this? If you do,
how much would need to be changed to get a general-purpose tool?

Personally, I don't like the idea of distributing functions that write
results to the disk. Results should IMHO be returned as variables such
that the user can do whatever suits him or her. Writing to the disk is
just fine for your own code, but I don't believe it is suitable for
general purpose code.

Søren

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Advice needed on contribution of .xls (Excel) access scripts

2009-12-09T11:01:04Z

Philip Nienhuis wrote:
>
> Hopefully the next windows binary has an up-to-date octave-forge. That's
> the binary used mainly by folks who don't do svn...

Do you have a particular package or specific patch in mind?
I can include a post-release version of an octave-forge
package in the binary installer.
I haven't done this as standard, only when the last available
released package did not build cleanly.

benjamin

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Gnuplot Drid Data File

2009-12-09T02:43:41Z

Hello Developers,

here is a routine that produces a Gnuplot Grid Data File,
that you need, if you want to plod 3D-Files directly
via Gnuplot.

Regards,

Stefan

--
Tel.: 0731-3805149
Ochsensteige 48
89075 Ulm

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

mylib_gnuplot_grid_data.m (13K) Download Attachment

Problem testing nnet's subset()

2009-12-08T14:22:25Z

Hi,

with SVN rev 6619:
octave:1> test subset
testing subset
error: empty error text, probably Ctrl-C --- aborting
error: called from:
error: /usr/share/octave/3.2.3/m/testfun/test.m at line 479, column 4

I'm using Octave 3.2.3. Maybe something changed in Octave's error handling?

Thanks
Thomas

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Request of admittance for a Developer of Italian translation for Octave

2009-12-08T13:35:37Z

Hi, my name is Antonio Guidetti
SourceForge username : antonioguidetti

I plan to translate completly in Italian the Octave manual, man pages, and info of the functions;
keeping the translation update.
http://www.gnu.org/software/octave/docs.html
http://octave.sourceforge.net/translation.html
http://octave.sourceforge.net/packages.html#language

My previus work :
Italian Language file for DragMath
https://sourceforge.net/search/?type_of_search=soft&words=DragMath
: it.xml

Antonio Guidetti
vortigher0@...

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Registration request (decay_lumi)

2009-12-07T00:09:43Z

Dear -

My user name is: pyff

I would like to submit to the GNU Octave Repository a piece of my work that I'm wrote for the analysis of luminescence exponential decay signal.

In the present script, the function leasqr and expfit are used to analyze coma separated data. Plots and results files are created and stored in a "Results" folder created in the working directory.

Please let me know if you need more informations or if there is some problem

With my best regards

Filippo Piffaretti

_____________________________________
Filippo Piffaretti
PhD Student, EPFL-LPAS

Station 6
CH H5 534 (Bâtiment CH)
1015 Lausanne
SWITZERLAND

**************************************
Work: +4121/693 76 23
Labo: +4121/693 36 12
Work Fax: +4121/693 51 45
Mobile: +4176/399 37 59
E-mail: filippo.piffaretti@...
______________________________________

[decay_lumi.m]

## Filippo Piffaretti, EPFL 2009 ## ## This program is free software; you can redistribute it and/or ## modify it under the terms of the GNU General Public ## License as published by the Free Software Foundation; ## either version 2, or (at your option) any later version. ## ## Octave is distributed in the hope that it will be useful, ## but WITHOUT ANY WARRANTY; without even the implied ## warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR ## PURPOSE. See the GNU General Public License for more ## details. ## ## You should have received a copy of the GNU General Public ## License along with Octave; see the file COPYING. If not, ## write to the Free Software Foundation, 59 Temple Place - ## Suite 330, Boston, MA 02111-1307, USA. ## usage: decay_lumi (param, indata, deg, pin) ## ## Small script that will help to analyse exponential decay ## phenomena. This script will coordinate diffenret type of ## approved program (Data retrivering, prony Method, partial ## differentiation, Levenberg-Marquart non-linear fitting ## algorithm. ## The results of the calculation will be displayed with plots ## and the results will be saved in different text documents in ## "Results" folder. ## Author: Filippo Piffaretti <filippo.piffaretti@...> ## Maintainer: Filippo Piffaretti <filippo.piffaretti@...> ## Created: November 2009 ## Version: 0.6 ## Keywords: Exponential fit, luminescence function decay_lumi (param, indata, deg, pin) ds = 0.001; %differential step stol= 0.0001; niter=50; minstep = [0; 0]; maxstep = [0.8; 0.8]; offset = 0; options = [minstep, maxstep]; %################################################## %### Argument processing ### %################################################## if (nargin < 3) % Check the number of arguments give by the user %----------------------------------------------- printf("\t Read the 'ReadMe.txt' file to set correctely the function parameters\n"); else %------------------------------------------------- %************************************************ %* Start to treat the parameters and the data * %************************************************ for i = 1:nargin %Handling the execution mode chose if (i==1) %--------------------------------- switch(param) case{"Silent" "silent" "S" "s"} param = 0; %silent case{"Verbose" "verbose" "V" "v"} param = 1; %verbose mode case{"Print" "print" "P" "p"} param = 2; %verbose mode otherwise printf("\t Incorrect setting of the first parameter %s ", param); param = 0; endswitch endif %--------------------------------------------- if (i==2) if(isdir(indata) == true) %Check if "indata" is a directoy or a single file %----------------------------------------- files=readdir(indata); %Create files array confirm_recursive_rmdir (0); %Do not ask for confirmation to cancel folder recursively if (indata(length(indata)) == "/") %Test the folder parameter entred, is it ending with "/"? rmdir (strcat (indata, "Results"),'s'); FileFolder = indata; %Set the file folder else rmdir (strcat (indata, "/Results"),'s'); FileFolder = strcat (indata, "/"); endif mkdir (strcat (FileFolder, "Results")); %Create the directory for the results else %If "indata" is a file %--------------------- if(fopen(indata,'r') == -1) %Check if the input is an existing data printf("\t !!ERROR: The file name entred does not exist!! \n"); else if (rindex (indata, "/") != 0) FileFolder = substr (indata, 1, rindex (indata, "/")); files = cellstr(indata(rindex (indata, "/")+1:end)); else FileFolder = "./"; files= cellstr(indata); %Put the file name to be treated in an array endif if (!isdir(strcat (FileFolder, "Results"))) %Create the folder only if it has not yet been created mkdir (strcat (FileFolder, "Results")); endif endif endif endif %--------------------------------------------- if (i == 4) % Transform the units of the entred parameters for w = 1:(length(pin)) if (mod(w,2) == 0) % (Starting point entred in: Volt; Seconds) ... pin(w)=1/pin(w); % to the parameters needed for the algorithm ... endif % (Volt; 1/Second). endfor pin = [offset; pin] if (param==1) %If verbose send to stdout %>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> % Print the informations to the SdtOut printf("\t User estimated parameters: \n\n\t Offset %d Volt\n", pin(1)); for w = 2:(length(pin)) if (mod(w,2)) printf("\t Lifetime: %.2f Âµs\n\n", 1/(pin(w)*1e6)); else printf("\t Pre-exp: %.2f Volt\n", pin(w)); endif endfor %<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< endif endif %-------------------------------------------------------- endfor %******************************************************** %########################################################## %### Loop to analyse the file or the folder files ### %########################################################## for file_n = 1:rows(files) lastwarn ("ok"); %Reset last warning message to zero (Singular Matrix) FileName = cellstr(split(files{file_n,1},".")); %File name of number "file_n" in the folder if ((isdir(files{file_n,1}) == false) && strcmp(FileName{length(FileName),1},"txt")) %Check if the FileName is not a directory and if it and with *.txt %----------------------------------------------------------------- if (rindex (FileName{1,1}, "/") != 0) %Define the file name that is going to be treated FileName = FileName{1,1}(rindex(FileName{1,1}, "/")+1:end); else FileName = FileName{1,1}; endif %Define the output filenames savePlot = strcat (FileFolder, "Results/", FileName, "_Plot.ps"); saveRes = strcat (FileFolder, "Results/", FileName, "_Res.txt"); saveSum = strcat (FileFolder, "Results/", "summary.txt"); %Retrieve the data values form the data file %------------------------------------------- matrix = data_coma (strcat(FileFolder,files{file_n,1})); h = (matrix(2,1) - matrix(1,1)); %Characterize the data N = length(matrix); t = (0:h:(N-1)*h)'; data = abs(matrix(:,2)); %******************************************** %*** Prony Method will Guess the *** %*** the starting point needed *** %*** for the L-M algorithm *** %******************************************** if (nargin == 3) [alpha,c,rms] = expfit(deg,0,h,data); alpha =(abs(alpha(:)))'; % Evaluate the module of the complex number c = (abs(c(:)))'; % Evaluate the module of the complex number pin = [c;alpha]; pin = vec(pin); % Rearrenge the results of the Prony pin = [offset; pin]; % method to feed the L-M algorithm if (param==1) %If verbose send to stdout %>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> % Print the informations to the SdtOut printf("\t Prony fitted parameters: Florescence fit %.2f Volt\n",c); printf("\t Prony fitted parameters: Life time %.2f us\n",(1./alpha)*1e6); %<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< endif endif %-------------------------------------------------------- %_________________________________________________________ %************************************ %*** Levenberg-Marquart *** %*** itterative - algorithm *** %************************************ switch (deg) %Define the number of terms needed for the mathematical case {1} %model dp = [ds; ds; ds]; %------------------------------------------------------ F = @exprf1; case {2} dp = [ds; ds; ds; ds; ds]; F = @exprf2; case {3} dp = [ds; ds; ds; ds; ds; ds; ds]; F = @exprf3; case {4} dp = [ds; ds; ds; ds; ds; ds; ... ds; ds; ds]; F = @exprf4; endswitch %-------------------------------------------------------- wt1 = ones(length(data),1)./sqrt(var(data)); %The moust satisfaying scheme to have a weighted fit, %wt1 = ones(length(data),1); %this will give more importance the the first value than %wt1 = ones(length(data),1)./sqrt(data); %the last one, normally subjected higher noise dFdp = @dfdp; %Calculate the estimated first derivative of the mathematical model %------------------------------------------------------------------ [f1, p1, kvg1, iter1, corp1, covp1, covr1, stdresid1, Z1, r21] = ... leasqr (t, data, pin, F, stol, niter, wt1, dp, dFdp);%, options); % Levenberg-Marquart algorithm %_________________________________________________________ %------------------------------ % Send to stdout is matrix singular %>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> file_warn = cellstr(split(lastwarn,",")); %Check for sigular matrix %------------------------ if (strcmp(file_warn{1,1},"warning: inverse: matrix singular to machine precision")) printf("!!! ERROR for %s, The results can not be interpreted, the fitting exponential order my not be adequate !!!\n",FileName); endif %<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< %*************************** %** Auto correlation plot ** %*************************** resid = (f1-data); autocRes = autocor(resid); ResMean=mean (resid); ResSdt=sqrt(var(resid)); cl95 = 1.96/sqrt(N); %Confidential Limit P(95%) cl99 = 2.576/sqrt(N); %Confidential Limit P(99%) stdev=sqrt(diag(covp1)); %Standard Deviation of the parameters estimated stdev=stdev*2.91; p2 = ones (length(p1)/2,2); stdev1= ones (length(p1)/2,2); %-------------------------------------------------------- %******************************************* %** Rearranging the solution vector ** %** to ease the handling of the results ** %******************************************* offset = p1(1); % From a coloumn A1;B1;A2;B2; vector...to a more confortable A1,B1;A2,B2; for i = 1:(length(p1)-1) % ---------------------------------------------------------------------- if (mod(i+1,2)) p2(int8(i/2),2)=p1(i+1,:); stdev1(int8(i/2),2)=stdev(i+1,:); pNL(int8(i/2),2)=pin(i+1,:); else p2(int8(i/2),1)=p1(i+1,:); stdev1(int8(i/2),1)=stdev(i+1,:); pNL(int8(i/2),1)=pin(i+1,:); endif endfor [s,i] = sort (p2(:,2)); %Sort the results form the longest lifetime to the shortest one p2 = (p2(i,:)); %-------------------------------------------------------------- % -------------------------------------------------------------- %*********************************** %** Calculate the rute chosed for ** %** the desexcitation ** %*********************************** rute = ((p2(:,1)'*diag(1./p2(:,2))))./((p2(:,1)'*(1./p2(:,2)))); %How many molecules choose the desexcitate through rute_x % -------------------------------------------------------------- %************************************* %** Write the results to differents ** %** output: stdout, files, plot ** %************************************* if (param==1) %If verbose send to stdout %>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> % Print the informations to the SdtOut printf("\n\t Lebenberg-Marquart fit\n"); printf("\t Offset %.2f Volt\n",offset); printf("\t Florescence fit %.2f Volt\n",p2(:,1)); printf("\t Life time %.2f us\n",(1./p2(:,2))*1e6); printf("\t Imortanza %.2f\n",rute(:)); printf("\t Number of iteration needed: %i; \n\t Max iteration %i \n", iter1, niter); %<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< endif % -------------------------------------------------------------- %******************* %** Generate Plot ** %******************* t=t*1000000; %Change the time scale for sec to us to have more readable graphs %>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> % Enable to show the graphs,or disable it fig = gcf (); if (param == 0 || param == 2) set (fig, "visible", "off"); else set (fig, "visible", "on"); endif %<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< subplot (221); %First subplot to rappresent the mathematical fit to the data (log scale) %----------------------------------------------------------------------- semilogy (t, data, "-1; Decay data;", t, f1,"-2 ;Decay exponential Model;"); title ("Experimental data fitting",'FontSize', 13); %axis([0,t(length(t)),0.2, 1]) xlabel ("Time, usec"); ylabel ("Log luminescece decay, a.u."); %---------------------------------------------------- %subplot(222); %Second subplot to rappresent the mathematical fit to the data (normal scale) %--------------------------------------------------------------------------- %plot(t, data, "-1; Decay data;", t, f1,"-2 ;Decay exponential Model;"); %axis([0,t(length(t)),0.2, 1]) %xlabel ("Time, usec"); %ylabel ("Luminescece decay, a.u."); %---------------------------------------------------- subplot(223); %Third subplot to rappresent the residials of the fit x=[0,t(length(t))]; %---------------------------------------------------- zero=[0,0]; plot (t,resid,"-3",x,zero,"-0 ;;"); title ("Residuals",'FontSize', 13); xlabel ("Time, usec"); ylabel ("Fit residuals, a.u."); axis ([0,t(length(t)),-.08,.08]); %---------------------------------------------------- subplot(224); %Forth subplot to rappresent the autocorrelation of the residials lag = 0:1:length(t)-1; %---------------------------------------------------------------- x=[0,lag(length(lag))+1]; l=[cl95,cl95]; %Setup the confidential bans limits l1=-l; b=[cl99,cl99]; b1=-b; plot(lag,autocRes,"-3;Residuals Autocorrelation;",x,l,"-4;;",x,l1,"-4;;", x,b, "-5;;",x,b1,"-5;;"); title ("Residuals autocorrelation",'FontSize', 13); xlabel ("Lags [a.u]"); axis ([0,x(2),-.2,1]); %Plot the confidential bands text (x(2)*2/4, 0.22, "Confidential Bands:"); text (x(2)*6/7, 0.08, "99%"); text (x(2)*6/7,-0.13, "95%"); %---------------------------------------------------- %Write the numerical results of the ritting at the place of subplot 222 %---------------------------------------------------------------------- text (x(2)*1/2, 2.8, "Summary of the Results",'HorizontalAlignment', 'center',... 'FontName','Times-Roman', 'FontSize', 22, 'Interpreter', 'none'); %Write the title text (x(2)*1/2, 2.55, strcat("Data file: ", FileName), ... 'HorizontalAlignment', 'center','FontName','Times-Roman', 'FontSize', 17, 'Interpreter', 'none'); offset_text = sprintf ("Offset: %0.2f Volt", offset); text (x(2)*1/2, 2.30, offset_text, 'HorizontalAlignment', ... 'center','FontName','Times-Roman', 'FontSize', 17); for i = 1:deg % Write the exponential parameters fit_A = sprintf("%0.2f",p2(i,1)); fit_tau = sprintf("%0.2f",(1./p2(i,2))*1e6); fit_deg = sprintf("%d",i); pos_text = 2.30 - i/5; text (x(2)*1/2, pos_text, strcat("A_",fit_deg,": ", fit_A, " Volt", ' \ldots {\fontsize{16}\tau}_', fit_deg, ": ", fit_tau, ' {\fontsize{12}\mu}s'),... 'HorizontalAlignment','center','FontName','Times-Roman', 'FontSize', 17); endfor %Check for sigular matrix %------------------------ if (strcmp(file_warn{1,1},"warning: inverse: matrix singular to machine precision")) text (x(2)*1/2, 1.6, strcat("! - Warning: Singular Matrix - !"),... 'HorizontalAlignment','center','FontName','Times-Roman', 'FontSize', 20); endif %---------------------------------------------------- print (savePlot, "-dpsc2", "-S1000,800", "-landscape"); %Save the plot to a file %>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> % Print the the plots to the default priter if (param==2) print ("-landscape"); endif %<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< %********************************** %** Write a complete text report ** %********************************** % Write all the informations about the data fitting in the following text file % Summary of the results of the specific data set ResFit_id = fopen(saveRes, 'w',"ieee-le"); fprintf(ResFit_id, "The function has been fitted with an exponential function of degree \r\n\t%i \r\n", deg); if (nargin == 2) fprintf(ResFit_id, "\r\n The initial parameters feeded to NLLS were found by fittig the data by Prony Method:\r\n"); else fprintf(ResFit_id, "\r\n The initial parameters has been given by the user:\r\n"); endif fprintf(ResFit_id, "\t Florescence fit: %.2f \r\n",pNL(:,1)); fprintf(ResFit_id, "\t Life time: %.2f us \r\n",(1./pNL(:,2))*1e6); fprintf(ResFit_id, "\r\n Number of iteration needed to reach a minima \r\n\t%i \r\n\r\n", iter1); fprintf(ResFit_id, "\r\n The Expenential parameters found are Sum(A_i*(expB_i)) + Offset:\r\n"); fprintf(ResFit_id, "\t Offset %.2f \r\n",offset); fprintf(ResFit_id, "\t Florescence fit: %.2f \r\n",p2(:,1)); fprintf(ResFit_id, "\t Life time: %.2f us \r\n",(1./p2(:,2))*1e6); fprintf(ResFit_id, "\r\n The relevance of the desexitation rute 'x' is expressed by [(A_x*B_x)/(Sum(A_i*B_i)]:\r\n"); fprintf(ResFit_id, "\t Rute %.1f%%\r\n",rute(:)*100); fprintf(ResFit_id, "\r\n Autocorrelation Plot, First Lag \r\n\t%0.4f [Confidence 95%%, %0.4f]\r\n", autocRes(2), cl95); fprintf(ResFit_id, "\r\n Correlation coefficient RÂ² \r\n\t%0.4f\r\n", r21); fprintf(ResFit_id, "\r\n The proceded files and the saved ones are located here:\r\n\t%s \r\n", pwd()); fclose(ResFit_id); % ######################################################################### % Write the fitting results in a single comma separeted summary *.txt file % ######################################################################### if(fopen(saveSum,'r') == -1) %Check if the file has already been created SumFit_id = fopen (saveSum, 'a', "ieee-le"); % If not -> write the file title and the columns identificator fprintf(SumFit_id, "Results: Exponential fit\n"); for i = 1:deg % The columns identificator fprintf(SumFit_id, "Pre-exponential coefficient %d [Volt], Lifetime %d [Âµs], ", i, i); endfor fprintf(SumFit_id, "Data file name, Fitting Offset [Volt], Degree of the exponential model \n\n"); else SumFit_id = fopen (saveSum, 'a', "ieee-le"); endif for i = 1:deg fprintf(SumFit_id, "%.4f, %.4f, ", p2(i,1),(1./p2(i,2))*1e6); endfor fprintf(SumFit_id, "%s, %.3f, %d \r\n", FileName, offset, deg); fclose(SumFit_id); % ############################################################################################# endif endfor endif endfunction function y = exprf1(x,p) y = p(1)+p(2)*exp(-p(3)*x); endfunction function y = exprf2(x,p) y = p(1) + p(2)*exp(-p(3)*x) + p(4)*exp(-p(5)*x); endfunction function y = exprf3(x,p) y = p(1) + p(2)*exp(-p(3)*x) + p(4)*exp(-p(5)*x) + ... p(6)*exp(-p(7)*x); endfunction function y = exprf4(x,p) y = p(1) + p(2)*exp(-p(3)*x) + p(4)*exp(-p(5)*x) + ... p(6)*exp(-p(7)*x) + p(8)*exp(-p(9)*x); endfunction

## Function: dacay_lumy.m
##
## usage: decay_lumi (param, indata, deg, [pin])
##
## Small script that will help to analyse exponential decay
## phenomena. This script will coordinate different type of
## approved program (Data retrieving, prony Method, partial
## differentiation, Levenberg-Marquart non-linear fitting
## algorithm.
## The results of the calculation will be displayed with plots
## and the results will be saved in different text documents in
## "Results" folder.

The function was written to ease the analyse the of exponential
luminescence decay data acquired with a home made time resolved
spectrophotometer.

You need to enter at least 3 parameter to the function.

param
Function parameter that will define the way the output will be print

- "Silent", "silent", "S", "s": Silent mode, no text and no plot, the fitting results are just saved in the Results folder

- "Verbose", "verbose", "V", "v": Verbose mode, everything is shown

- "Print", "print", "P", "p": Printout mode, similar to the silent mode, but the graphical results are sent to the default printer.

%%
"Print" or "S" or "v"
%%
--

indata
The experimental data, recorded with a Lecroy LT342, need to be
in a coma separated format to be analysed.

%%
Example of a coma separate data file "coma_separated.txt":

3.25e-006,-0.855649
3.35e-006,-0.775937
3.45e-006,-0.797849
3.55e-006,-0.740049
3.65e-006,-0.854062
3.75e-006,-0.761899
3.85e-006,-0.715024
3.95e-006,-0.700986
...
%%
--

deg
Exponential fitting degree, the data are fitted with a 1 to 4 terms
exponential mathematical model. By setting this parameter the
exponential model can be set. For this parameter, set the
smallest value possible to have stable residual autocorrelation
plot.

%%
1 or 2 or 3 or 4
%%
--

[pin]
Optional, first guess of the characterizing parameters of the
exponential function, needed for the Levenberg-Marquart algorithm.
The first guess will be used to find the minimum of the
residual function. The more the initial guess are close to the final
solution the more the iterative algorithm will be stable (less
sensible to possible local minima in the residual function).

The initial parameter has to be entered in the form:
[Volt_1; Lifetime_1 (in seconds); Volt_2; Lifetime_2; ...; Lifetime_2]

%%
[.1;5e-7;3e-2;1e-7]
%%

If the user do not provide any initial guess parameter the script
will use the Prony methods to estimate the fitting solution to
feed the iterative algorithm with acceptable initial solution.

Despite this possibility to find "automaticaly" the initial guess,
it's highly recommended that the user fix this parameter. It's also
very important the use the same initial guess for all the measurement
of the same series. The mathematical fit to a data set is
a very sensible process. The iterative Levenberg-Marquart
algorithm demonstrated very good stability but it is always
very wise to fix all possible parameter.
--
------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Changing tests of integration package

2009-12-06T11:24:21Z

Hi,

the integration package has a file inst/test/run2dtests.m which tests
the package, but is extremely verbose and expects the user to check the
results manually.

Unless someone objects, I'll remove the script and distribute the test
in the corresponding files.

Additionnaly, is the following output from the test script expected?

======================================================================
Integral did not converge--singularity likely
f = 0
correct_ans = 2
Integral did not converge--singularity likely
f = 0
correct_ans = 2
Integral did not converge--singularity likely
No protocol specified
warning: unable to open X11 DISPLAY
f = 0
correct_ans = 2
Integral did not converge--singularity likely
f = 0
correct_ans = 2.66666666666667
Integral did not converge--singularity likely
f = 0
correct_ans = 6.40000000000000
Integral did not converge--singularity likely
f = 0
correct_ans = 10.6666666666667
Integral did not converge--singularity likely
f = 0
correct_ans = 18.2857142857143
Integral did not converge--singularity likely
f = 0
correct_ans = 32
Integral did not converge--singularity likely
f = 0
correct_ans = 56.8888888888889
Integral did not converge--singularity likely
f = 0
correct_ans = 102.400000000000
Integral did not converge--singularity likely
f = 0
correct_ans = 186.181818181818
======================================================================

Is this expected, is this a bug in the test suite, is it a bug in the
package or something else?

Thomas

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Problem with testing octave-image's qtdecomp()

2009-12-06T10:24:44Z

Hi,

I get an error in qtdecomp's testsuite when using Octave 3.2.3; it works
with 3.0.5, though. SVN rev is 6599.

========================================================================
octave:1> test qtdecomp
PASSES 14 out of 14 tests
octave:2> version
ans = 3.0.1

========================================================================

octave:1> test qtdecomp
***** assert(full(qtdecomp(A,@f)),[ones(4),B4;ones(4,8)]);
!!!!! test failed
`A' undefined near line 11 column 7shared variables error: `A' undefined
near line 11 column 7
error: evaluating argument list element number 1
error: called from:
error: f at line 11, column 4
error: evaluating argument list element number 4
error: evaluating argument list element number 2
error: evaluating argument list element number 1
error: /usr/share/octave/3.2.3/m/testfun/test.m at line 498, column 2
octave:1> version
ans = 3.2.3
========================================================================

Did anything in Octave's test harness change?

Thanks
Thomas

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Advice needed on contribution of .xls (Excel) access scripts

2009-12-05T02:07:26Z

Michael Goffioul wrote:
> On Sat, Dec 5, 2009 at 12:39 AM, Philip Nienhuis <pr.nienhuis@...> wrote:
>> I'll try to ask there.
>> My scripts don't actually "link" to COM/ActiveX, they just invoke it.
>> I think if any trespassing has ever occurred, it's probably in the
>> (compiled) Windows package itself. But... AINAL
>
> The COM interface is only linked to Win32 system libraries,
> afaik this is allowed by GPL.

That's more or less what I figured.

BTW I could have added "at all" after "occurred" in the above sentence,
that would have made it more clear that I do not really think there has
ever been a license issue with the Windows package. Sorry.

Thanks,

Philip

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Advice needed on contribution of .xls (Excel) access scripts

2009-12-05T00:15:01Z

Philip Nienhuis wrote:

>> I tend to use texinfo so that help pages appear correctly and I can
>> include all of the documentation in a *.info in the directory with
>> the functions and "doc <func>" can give more complete documentation
>> including examples and context than the help file can.
>
> Hmmm, texinfo proficiency is not my greatest virtue :-)
>
> I did add texinfo-like stanzas in the m-file themselves, those 4 or 5
> codes I found in other m-files were at least comprehensible and it
> looks OK in octave when using "help <script>".

Take a look at the comms and fixed packages documentation for ideas you
can basically just use the mkdoc and mktexi scripts in octave-forge to
allow you to include your existing help strings in the docs with the
DOCSTRING macro and add some extra explanation around it..
>
> Could WinTexmacs help here?
Don't know...

>
>>> 3. Place to live in issue
>>> -------------------------
>>>
>>> My current idea is to put all the files in the IO package (licenses
>>> permitting?). It doesn't really hurt if neither COM nor Java/ POI are
>>> installed; xls stuff won't read .xls files then but at least xlsread.m
>>> will still function as it did before (= read .csv files).
>>>
>> Not sure that is the best idea, as you'll be adding a dependency on
>> the java package (that has been troublesome in the past).
>
> Well, dependency... the way I made it work is this:
>
> 1. <My xls stuff> first checks for ActiveX/COM. If not found, too bad,
> try next .xls interface. So having the Windows package is helpful, but
> it doesn't really depend on it.
> 2. If no ActiveX/COM is found it tries Java.
> If Java doesn't work it reverts to csvread stuff, so IMO there's no
> real dependency on Java either.
> 3. If Java is found to be working (by a try/catch on javaclasspath)
> the classpath is checked for the Apache/POI classes. If not all of
> them found, again csvread looms.
>
> And so on.
>
> Some interface method will eventually work, but the more options you
> got installed the better the Excel file interface you might get. Al
> sorted out by the scripts w/o much user intervention.
>
> All in all, that's why IMO it won't hurt to put it in the IO package.
> But I'll happily accept better suggestions.

Ok, then I suppose that means that the java and windows packages are
"suggested", and you should use a keyword Suggested (that will be
ignored by octave-forge) in the DESCRIPTION file of the io package if
you include your files there.

>
> I think it is prnienhuis -a-t- users.sf.net
> BTW When I reactivated it last summer, I was immediately drowned in
> spam with that very sourceforge email address forged as sender, so I'm
> a bit wary here. Apparently the sourceforge security system is a bit
> leaky.
>
> Only thing I need to sort out is how ssl (putty?) works on Windows (I
> already got the octave-forge svn on harddisk). I rarely use linux
> these days but I might give svn/ssl a try there.
>
> I'll await more concrete directions on where to put all the xls scripts.

Ok I added you as a developer.. Its not sourceforge that is sending the
spam (read the headers) someone if using the e-mail address to spam you
and trying to get around spam filters by using the destination e-mail
address for the sender address (as this will be a valid address).
Sourceforge e-mail addresses are fairly easy for spammers to get as they
can basically get them from the sourceforge website.

David

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Advice needed on contribution of .xls (Excel) access scripts

2009-12-04T18:17:28Z

On Sat, Dec 5, 2009 at 12:39 AM, Philip Nienhuis <pr.nienhuis@...> wrote:
> I'll try to ask there.
> My scripts don't actually "link" to COM/ActiveX, they just invoke it.
> I think if any trespassing has ever occurred, it's probably in the
> (compiled) Windows package itself. But... AINAL

The COM interface is only linked to Win32 system libraries,
afaik this is allowed by GPL.

Michael.

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Advice needed on contribution of .xls (Excel) access scripts

2009-12-04T16:39:36Z

David Bateman wrote:

> Philip Nienhuis wrote:
>> Hi,
>>
>> At my employer's office we use Matlab intensively and many, many data
>> are kept & handled in Excel spreadsheet files. So to be able to use
>> octave there I've made xlsread / xlswrite / xlsfinfo scripts as
>> octave drop-in replacements for use in procedures otherwise done with
>> Matlab. Presently it consists of a collection of a dozen scripts.
>>
>> My xlsread.m & xlswrite.m (written from scratch) do mainly pre- and
>> post-processing and for the rest are mere wrappers for COM (i.e. Excel
>> itself) & Java /Apache POI interfaces and the existing csvread stuff,
>> in that order of preference, to do the actual reading/writing.
>> I've taken care to make sure that adding other interfaces should be
>> fairly easy.
>>
>> I've made it all a bit more versatile than Matlab's counterparts.
>> E.g., using my stuff, for each and every call Excel /Java don't need
>> to be started and .xls files need not be read/written again; this may
>> speed up processing significantly.
>>
>> It all seems to work OK, both under WinXP and Mandriva Linux 2009.0,
>> and I'd like to contribute the scripts collection very soon now (just
>> some more testing needed on non-Windows platforms + little buglet
>> fixing).

... Especially reliable Excel formula evaluation through Java is a
problem :-(

I'm just trying JExcelAPI too, hopefully that's more robust with Excel
formulas, but it's a little less versatile then Apache POI.

> Cool...
>
>> But I might first need a bit of advice on contribution:
>>
>> 1. License issues: (IANAL)
>> ------------------
>>
>> - Is it actually allowed to interface directly to Excel through COM in
>> Octave's (octave-forge's) license terms? If not, is there an alternative
>> way to contribute the relevant scripts? LGPL?
>> (I somewhat suppose that just running octave on Windows already might
>> boil down to more or less the same issue)
>>
> It depends... From what I see COM can be considered as a system library
> that then calls excel, word etc if they are installed. If so then the
> section on "System Libraries" in the GPL probably holds and linking a
> GPL program to the COM interface should be ok.. Maybe ask on
> licensing@... would get a clearer answer though

I'll try to ask there.
My scripts don't actually "link" to COM/ActiveX, they just invoke it.
I think if any trespassing has ever occurred, it's probably in the
(compiled) Windows package itself. But... AINAL

<snip>

>
>> 2. Technical issues:
>> --------------------
>>
>> - I'd like to add some documentation too (README file or so). Are
>> there any standards for that on octave-forge?
>>
>> - Apache POI only supports .xls files from Excel 97 and up (BIFF8 and
>> OOXML). Matlab in "basic" mode only supports .xls files type Excel '95
>> (BIFF5). So Matlab compatibility (in the sense of smooth file
>> interchange) is not complete on systems w/o Excel. (But OpenOffice.org
>> can be used for xls conversion).
>>
> I tend to use texinfo so that help pages appear correctly and I can
> include all of the documentation in a *.info in the directory with the
> functions and "doc <func>" can give more complete documentation
> including examples and context than the help file can.

Hmmm, texinfo proficiency is not my greatest virtue :-)

I did add texinfo-like stanzas in the m-file themselves, those 4 or 5
codes I found in other m-files were at least comprehensible and it looks
OK in octave when using "help <script>".

Could WinTexmacs help here?

>> - Octave (& octave-forge) itself has small issues:
>> * javaclasspath still uses split.m instead of strsplit.m leading to a
>> classpath split up into cell array with wrong dimensions (e.g.,
>> (1,1), (2,1), ... rather than (1,1), (1,2), ...
>> A workaround is probably easy but I'd rather see the java stuff
>> fixed. I can send a fixed javaclasspath.m separately.
>>
> Michael should probably respond to that, the java code is his..
>
>> * The windows package (1.0.8) as distributed is not up-to-date;
>> the latest fixes in svn should have been applied to have the
>> COM interfaces work properly.
>>
> Michael again, though I'm sure he'd welcome help maintaining the package..

He already committed a patch from me.

Hopefully the next windows binary has an up-to-date octave-forge. That's
the binary used mainly by folks who don't do svn...

>> 3. Place to live in issue
>> -------------------------
>>
>> My current idea is to put all the files in the IO package (licenses
>> permitting?). It doesn't really hurt if neither COM nor Java/ POI are
>> installed; xls stuff won't read .xls files then but at least xlsread.m
>> will still function as it did before (= read .csv files).
>>
> Not sure that is the best idea, as you'll be adding a dependency on the
> java package (that has been troublesome in the past).

Well, dependency... the way I made it work is this:

1. <My xls stuff> first checks for ActiveX/COM. If not found, too bad,
try next .xls interface. So having the Windows package is helpful, but
it doesn't really depend on it.
2. If no ActiveX/COM is found it tries Java.
If Java doesn't work it reverts to csvread stuff, so IMO there's no real
dependency on Java either.
3. If Java is found to be working (by a try/catch on javaclasspath) the
classpath is checked for the Apache/POI classes. If not all of them
found, again csvread looms.

And so on.

Some interface method will eventually work, but the more options you
got installed the better the Excel file interface you might get. Al
sorted out by the scripts w/o much user intervention.

All in all, that's why IMO it won't hurt to put it in the IO package.
But I'll happily accept better suggestions.

I know java can be a bit troublesome (see the various threads on java
installation in this forum), I had my share as well. But in most linux
distros that has all been sorted out for users.
E.g. on Mandriva 2009.0, octave-3.2.3 can be installed easily by .rpm,
including octave-forge (an older release with java-1.2.5 but that *does*
work with Apache POI). I only had to add a symbolic link in /usr/lib/jvm
(as at Mandriva they seemingly like to complicate their system layout
beyond conventional comprehension of mere mortal users like me).

>> Q:
>>
>> 1. Any objections to this?
>>
>> 2. Would a maintainer be so kind to copy files into place, or should I
>> really do it myself? I know package index scripts have to be updated
>> too, I'll prepare them happily.
>>
> Better if you send your sourceforge username and get you added to the
> maintainers and you commit the code yourself.. That's pretty much the
> octave-forge way, with the only bar being some code sent to the list so
> as to keep the maintainers and package owners to those that actually
> commit code. Though as you've already committed to octave-forge perhaps
> that condition should be dropped in your case.

I think it is prnienhuis -a-t- users.sf.net
BTW When I reactivated it last summer, I was immediately drowned in spam
with that very sourceforge email address forged as sender, so I'm a bit
wary here. Apparently the sourceforge security system is a bit leaky.

Only thing I need to sort out is how ssl (putty?) works on Windows (I
already got the octave-forge svn on harddisk). I rarely use linux these
days but I might give svn/ssl a try there.

I'll await more concrete directions on where to put all the xls scripts.

Thanks,

Philip

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Advice needed on contribution of .xls (Excel) access scripts

2009-12-04T13:38:32Z

Philip Nienhuis wrote:

> Hi,
>
> At my employer's office we use Matlab intensively and many, many data
> are kept & handled in Excel spreadsheet files. So to be able to use
> octave there I've made xlsread / xlswrite / xlsfinfo scripts as
> octave drop-in replacements for use in procedures otherwise done with
> Matlab. Presently it consists of a collection of a dozen scripts.
>
> My xlsread.m & xlswrite.m (written from scratch) do mainly pre- and
> post-processing and for the rest are mere wrappers for COM (i.e. Excel
> itself) & Java /Apache POI interfaces and the existing csvread stuff, in
> that order of preference, to do the actual reading/writing.
> I've taken care to make sure that adding other interfaces should be
> fairly easy.
>
> I've made it all a bit more versatile than Matlab's counterparts. E.g.,
> using my stuff, for each and every call Excel /Java don't need to be
> started and .xls files need not be read/written again; this may speed up
> processing significantly.
>
> It all seems to work OK, both under WinXP and Mandriva Linux 2009.0, and
> I'd like to contribute the scripts collection very soon now (just some
> more testing needed on non-Windows platforms + little buglet fixing).
>

Cool...

> But I might first need a bit of advice on contribution:
>
> 1. License issues: (IANAL)
> ------------------
>
> - Is it actually allowed to interface directly to Excel through COM in
> Octave's (octave-forge's) license terms? If not, is there an alternative
> way to contribute the relevant scripts? LGPL?
> (I somewhat suppose that just running octave on Windows already might
> boil down to more or less the same issue)
>

It depends... From what I see COM can be considered as a system library
that then calls excel, word etc if they are installed. If so then the
section on "System Libraries" in the GPL probably holds and linking a
GPL program to the COM interface should be ok.. Maybe ask on
licensing@... would get a clearer answer though

> - The Apache POI Java libs have an Apache license. Is that sufficiently
> compatible with octave's GPL license?
> (The jars do not have to be distributed, users can d/l & install them)
>
Here is what

http://www.gnu.org/licenses/license-list.html

<quote>

Apache License, Version 2.0 <http://www.apache.org/licenses/LICENSE-2.0>

This is a free software license, compatible with version 3 of the GPL.

Please note that this license is not compatible with GPL version 2,
because it has some requirements that are not in the older version.
These include certain patent termination and indemnification provisions.

</quote>

so probably no problems..

> 2. Technical issues:
> --------------------
>
> - I'd like to add some documentation too (README file or so). Are there
> any standards for that on octave-forge?
>
> - Apache POI only supports .xls files from Excel 97 and up (BIFF8 and
> OOXML). Matlab in "basic" mode only supports .xls files type Excel '95
> (BIFF5). So Matlab compatibility (in the sense of smooth file
> interchange) is not complete on systems w/o Excel. (But OpenOffice.org
> can be used for xls conversion).
>

I tend to use texinfo so that help pages appear correctly and I can
include all of the documentation in a *.info in the directory with the
functions and "doc <func>" can give more complete documentation
including examples and context than the help file can.

> - Octave (& octave-forge) itself has small issues:
> * javaclasspath still uses split.m instead of strsplit.m leading to a
> classpath split up into cell array with wrong dimensions (e.g.,
> (1,1), (2,1), ... rather than (1,1), (1,2), ...
> A workaround is probably easy but I'd rather see the java stuff
> fixed. I can send a fixed javaclasspath.m separately.
>
Michael should probably respond to that, the java code is his..

> * The windows package (1.0.8) as distributed is not up-to-date;
> the latest fixes in svn should have been applied to have the
> COM interfaces work properly.
>
Michael again, though I'm sure he'd welcome help maintaining the package..

> 3. Place to live in issue
> -------------------------
>
> My current idea is to put all the files in the IO package (licenses
> permitting?). It doesn't really hurt if neither COM nor Java/ POI are
> installed; xls stuff won't read .xls files then but at least xlsread.m
> will still function as it did before (= read .csv files).
>
Not sure that is the best idea, as you'll be adding a dependency on the
java package (that has been troublesome in the past).

> Q:
>
> 1. Any objections to this?
>
> 2. Would a maintainer be so kind to copy files into place, or should I
> really do it myself? I know package index scripts have to be updated
> too, I'll prepare them happily.
>
Better if you send your sourceforge username and get you added to the
maintainers and you commit the code yourself.. That's pretty much the
octave-forge way, with the only bar being some code sent to the list so
as to keep the maintainers and package owners to those that actually
commit code. Though as you've already committed to octave-forge perhaps
that condition should be dropped in your case.

> Many years ago I had access as octave-forge developer when I made the
> Windows FAQ, but that access is probably outdated now.
>
In fact the developer list hasn't been cleaned up for at least a year or
probably 18 months, with the criteria that a commit in the last 12
months kept the developer status.... Though yes it appears that you
aren't listed as a developer

D.

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

[Solved] Re: Cannot install miscellaneous package on Snow Leopard (OSX 10.6)

2009-12-04T12:00:34Z

From the link you supplied, I eventually found this fix:

http://www.supermegaultragroovy.com/blog/2009/09/17/installing-octave-packages-in-snow-leopard/

Installing miscellaneous works fine now. Thanks!

-Dave
On Dec 4, 2009, at 11:30 AM, Thomas Treichl wrote:

> David Hirsch schrieb:
>> This is my first time on this list. I've searched for reports of
>> this problem on this list's archive and the other (nabble-based)
>> list, without finding any. I would like to know if there is a fix
>> for this - will the package install if I try to build Octave from
>> source? It would appear not. (I tried using MacPorts, but it
>> wouldn't even install the base Octave system) Here's the problem:
>> I've downloaded the current OSX Octave Binary, which runs fine.
>> When I try to install the miscellaneous package, I get these errors:
>
> Hi David,
>
> this is for the 3.2.3 binary of Octave.app that you have downloaded
> from Sourceforge. I expect you find the problem and the solution here:
>
> http://old.nabble.com/Octave.app-3.2.3-for-Mac-OS-X-10.4.x-and-10.5.x-to25727982.html#a25774716
>
> I try to fix that for Octave.app 3.2.4...
>
> Regards,
>
> Thomas

============================
Dave Hirsch
Associate Professor
Department of Geology
Western Washington University
persistent email: dhirsch@...
http://www.davehirsch.com
voice: (360) 389-3583
aim: dhirsch@...
vCard: http://almandine.geol.wwu.edu/~dave/personal/DaveHirsch.vcf
============================

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Cannot install miscellaneous package on Snow Leopard (OSX 10.6)

2009-12-04T11:30:31Z

David Hirsch schrieb:
> This is my first time on this list. I've searched for reports of this
> problem on this list's archive and the other (nabble-based) list,
> without finding any. I would like to know if there is a fix for this -
> will the package install if I try to build Octave from source? It
> would appear not. (I tried using MacPorts, but it wouldn't even install
> the base Octave system) Here's the problem:
> I've downloaded the current OSX Octave Binary, which runs fine. When I
> try to install the miscellaneous package, I get these errors:

Hi David,

this is for the 3.2.3 binary of Octave.app that you have downloaded from
Sourceforge. I expect you find the problem and the solution here:

http://old.nabble.com/Octave.app-3.2.3-for-Mac-OS-X-10.4.x-and-10.5.x-to25727982.html#a25774716

I try to fix that for Octave.app 3.2.4...

Regards,

Thomas

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Cannot install miscellaneous package on Snow Leopard (OSX 10.6)

2009-12-04T10:57:13Z

This is my first time on this list. I've searched for reports of this problem on this list's archive and the other (nabble-based) list, without finding any. I would like to know if there is a fix for this - will the package install if I try to build Octave from source? It would appear not. (I tried using MacPorts, but it wouldn't even install the base Octave system) Here's the problem:

I've downloaded the current OSX Octave Binary, which runs fine. When I try to install the miscellaneous package, I get these errors:

ld: warning: in /Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/bin/octave-3.2.3, file is not of required architecture

ld: warning: in /Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib/octave-3.2.3/liboctinterp.dylib, file is not of required architecture

(plus about 25 more similar errors)

waitbar.cc: In function ‘octave_value_list Fwaitbar(const octave_value_list&, int)’:

waitbar.cc:126: warning: deprecated conversion from string constant to ‘char*’

waitbar.cc:147: warning: deprecated conversion from string constant to ‘char*’

waitbar.cc:148: warning: deprecated conversion from string constant to ‘char*’

Undefined symbols:

"Array<octave_value>::resize_fill_value()", referenced from:

Array<octave_value>::resize(int) in waitbar.o

"Array<char>::~Array()", referenced from:

vtable for Array<char>in waitbar.o

"Array<double>::~Array()", referenced from:

vtable for Array<double>in waitbar.o

(plus about 30 more similar errors)

ld: symbol(s) not found

collect2: ld returned 1 exit status

make: *** [waitbar.oct] Error 1

'make' returned the following error: mkoctfile -DHAVE_OCTAVE_32 -v -DUSE_TERM -DHAVE_TERMCAP_H waitbar.cc -ltermcap

g++ -c -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/curl -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/readline -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/curl -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/readline -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/curl -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/readline -fPIC -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/octave-3.2.3 -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/octave-3.2.3/octave -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -mieee-fp -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/freetype2 -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/curl -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/readline -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/curl -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/readline -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/curl -I/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/include/readline -DHAVE_OCTAVE_32 -DUSE_TERM -DHAVE_TERMCAP_H waitbar.cc -o waitbar.o

g++ -bundle -bundle_loader /Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/bin/octave-3.2.3 -o waitbar.oct waitbar.o -ltermcap -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib/octave-3.2.3 -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib -loctinterp -loctave -lcruft -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib/octave-3.2.3 -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib/octave-3.2.3 -Wl,-framework -Wl,vecLib -lfftw3 -lfftw3f -lreadline -lncurses -lhdf5 -lz -lm -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/lib -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/bin/../lib/gcc-lib/i386-apple-darwin8.11.1/4.0.3/ -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/bin/../lib/gcc-lib/i386-apple-darwin8.11.1/4.0.3 -L/usr/lib/gcc// -L/Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/bin/../lib/gcc-lib/i386-apple-darwin8.11.1/4.0.3/// -L/usr/lib// -lhdf5 -lz -lf95 -lm -lSystemStubs -lmx

error: called from `pkg>configure_make' in file /Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/share/octave/3.2.3/m/pkg/pkg.m near line 1253, column 2

error: called from:

error: /Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/share/octave/3.2.3/m/pkg/pkg.m at line 714, column 5

error: /Users/dave/Applications/Geology/Octave/Octave.app/Contents/Resources/share/octave/3.2.3/m/pkg/pkg.m at line 287, column 7

============================

Dave Hirsch

Associate Professor

Department of Geology

Western Washington University

persistent email: dhirsch@...

http://www.davehirsch.com

voice: (360) 389-3583

aim: dhirsch@...

vCard: http://almandine.geol.wwu.edu/~dave/personal/DaveHirsch.vcf

============================

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Wiki password please!

2009-12-03T16:05:18Z

Done!

Quoting Benjamin Christensen <bchristensen@...> (Thu, 3
Dec 2009 10:09:35 -0700):

> Benj Christensen
>
> Product Design Lead
>
> Willowstick Technologies
>
> (801) 984-9867 11814 Election Rd. Ste. 100
>
> Draper, UT 84020

--
http://www.isr.ist.utl.pt/~etienne

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Advice needed on contribution of .xls (Excel) access scripts

2009-12-03T15:35:16Z

Hi,

At my employer's office we use Matlab intensively and many, many data
are kept & handled in Excel spreadsheet files. So to be able to use
octave there I've made xlsread / xlswrite / xlsfinfo scripts as
octave drop-in replacements for use in procedures otherwise done with
Matlab. Presently it consists of a collection of a dozen scripts.

My xlsread.m & xlswrite.m (written from scratch) do mainly pre- and
post-processing and for the rest are mere wrappers for COM (i.e. Excel
itself) & Java /Apache POI interfaces and the existing csvread stuff, in
that order of preference, to do the actual reading/writing.
I've taken care to make sure that adding other interfaces should be
fairly easy.

I've made it all a bit more versatile than Matlab's counterparts. E.g.,
using my stuff, for each and every call Excel /Java don't need to be
started and .xls files need not be read/written again; this may speed up
processing significantly.

It all seems to work OK, both under WinXP and Mandriva Linux 2009.0, and
I'd like to contribute the scripts collection very soon now (just some
more testing needed on non-Windows platforms + little buglet fixing).

But I might first need a bit of advice on contribution:

1. License issues: (IANAL)
------------------

- Is it actually allowed to interface directly to Excel through COM in
Octave's (octave-forge's) license terms? If not, is there an alternative
way to contribute the relevant scripts? LGPL?
(I somewhat suppose that just running octave on Windows already might
boil down to more or less the same issue)

- The Apache POI Java libs have an Apache license. Is that sufficiently
compatible with octave's GPL license?
(The jars do not have to be distributed, users can d/l & install them)

2. Technical issues:
--------------------

- I'd like to add some documentation too (README file or so). Are there
any standards for that on octave-forge?

- Apache POI only supports .xls files from Excel 97 and up (BIFF8 and
OOXML). Matlab in "basic" mode only supports .xls files type Excel '95
(BIFF5). So Matlab compatibility (in the sense of smooth file
interchange) is not complete on systems w/o Excel. (But OpenOffice.org
can be used for xls conversion).

- Octave (& octave-forge) itself has small issues:
* javaclasspath still uses split.m instead of strsplit.m leading to a
classpath split up into cell array with wrong dimensions (e.g.,
(1,1), (2,1), ... rather than (1,1), (1,2), ...
A workaround is probably easy but I'd rather see the java stuff
fixed. I can send a fixed javaclasspath.m separately.
* The windows package (1.0.8) as distributed is not up-to-date;
the latest fixes in svn should have been applied to have the
COM interfaces work properly.

3. Place to live in issue
-------------------------

My current idea is to put all the files in the IO package (licenses
permitting?). It doesn't really hurt if neither COM nor Java/ POI are
installed; xls stuff won't read .xls files then but at least xlsread.m
will still function as it did before (= read .csv files).

Q:

1. Any objections to this?

2. Would a maintainer be so kind to copy files into place, or should I
really do it myself? I know package index scripts have to be updated
too, I'll prepare them happily.
Many years ago I had access as octave-forge developer when I made the
Windows FAQ, but that access is probably outdated now.

(Oh and before you ask: yes I know, .xls and friends ARE a proprietary
format. But so is Matlab's .mat, for which octave has an interface - in
core octave, for that matter.
And xls files are indeed widely used out there....)

Philip

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Wiki password please!

2009-12-03T09:09:35Z

Benj Christensen

Product Design Lead

Willowstick Technologies

(801) 984-9867

11814 Election Rd. Ste. 100

Draper, UT 84020

Re: File 'jni.h' missing in java-1.2.6.tar.gz

2009-12-03T00:14:45Z

> Hi Alex,
>
> That's a request for the Windows binary packager (Benjamin).
> I'm not providing binary packages anymore. Note that this is a
> time-consuming job. I guess/hope that Benjamin will eventually
> include the java package within the binary release.
>

Yes, I gathered from frequent posts on the mailing list that having
the java package available is an item on the wishlist :)

In other words, I plan to include it for the 3.2.4 release.

benjamin

--
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/chbrowser

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Simple change on openmpi_ext

2009-12-02T04:27:05Z

Sorry, the result is the same.
Michael

On Wed, Dec 2, 2009 at 12:19 PM, Riccardo Corradini <riccardocorradini@...> wrote:

I've uploaded a small change on all cc files
Please let me know if it improves.
Bests
Riccardo

--- Mer 2/12/09, Michael Creel <michael.creel@...> ha scritto:

Da: Michael Creel <michael.creel@...>
Oggetto: Re: [OctDev] A simple C++ example of a class with MPI_Comunicatorproperty
A: "Riccardo Corradini" <riccardocorradini@...>
Cc: "octave-forge list" <octave-dev@...>
Data: Mercoledì 2 dicembre 2009, 11:41

Hi Riccardo,
I'm still seeing problems with this, both for helloworld.m and for mc_example2.m. This is with Octave 3.2.2 and Open MPI 1.3.2 from Ubuntu 9.10, AMD64.

Below is the output I get from helloworld. Please let me know if I can do anything to generate useful information about this.

Thanks,
Michael

michael@yosemite:~/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/inst$ mpirun -np 3 octave -q --eval helloworld
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 0
p = 3
We are at rank 0 that is master etc..
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 1
p = 3
info for sending is is
info = 0
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 2
p = 3
info for sending is is
info = 0

Re: Simple change on openmpi_ext

2009-12-02T03:19:43Z

I've uploaded a small change on all cc files
Please let me know if it improves.
Bests
Riccardo

--- Mer 2/12/09, Michael Creel <michael.creel@...> ha scritto:

Da: Michael Creel <michael.creel@...>
Oggetto: Re: [OctDev] A simple C++ example of a class with MPI_Comunicatorproperty
A: "Riccardo Corradini" <riccardocorradini@...>
Cc: "octave-forge list" <octave-dev@...>
Data: Mercoledì 2 dicembre 2009, 11:41

Hi Riccardo,
I'm still seeing problems with this, both for helloworld.m and for mc_example2.m. This is with Octave 3.2.2 and Open MPI 1.3.2 from Ubuntu 9.10, AMD64.

Below is the output I get from helloworld. Please let me know if I can do anything to generate useful information about this.

Thanks,
Michael

michael@yosemite:~/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/inst$ mpirun -np 3 octave -q --eval helloworld
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 0
p = 3
We are at rank 0 that is master etc..
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 1
p = 3
info for sending is is
info = 0
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 2
p = 3
info for sending is is
info = 0

Re: A simple C++ example of a class with MPI_Comunicatorproperty

2009-12-02T03:04:44Z

On my computer everything work, but I uploaded a small change
have a look at it on svn
Please let me know if it improves.
Bests
Riccardo

--- Mer 2/12/09, Michael Creel <michael.creel@...> ha scritto:

Da: Michael Creel <michael.creel@...>
Oggetto: Re: [OctDev] A simple C++ example of a class with MPI_Comunicatorproperty
A: "Riccardo Corradini" <riccardocorradini@...>
Cc: "octave-forge list" <octave-dev@...>
Data: Mercoledì 2 dicembre 2009, 11:41

Hi Riccardo,
I'm still seeing problems with this, both for helloworld.m and for mc_example2.m. This is with Octave 3.2.2 and Open MPI 1.3.2 from Ubuntu 9.10, AMD64.

Below is the output I get from helloworld. Please let me know if I can do anything to generate useful information about this.

Thanks,
Michael

michael@yosemite:~/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/inst$ mpirun -np 3 octave -q --eval helloworld
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 0
p = 3
We are at rank 0 that is master etc..
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 1
p = 3
info for sending is is
info = 0
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 2
p = 3
info for sending is is
info = 0
Greetings from process: 1!
We are at rank 0 that is master etc..
Greetings from process: 2!
*** glibc detected *** octave: double free or corruption (fasttop): 0x00000000017b18c0 ***
======= Backtrace: =========
/lib/libc.so.6[0x7f78ee49fdd6]
/lib/libc.so.6(cfree+0x6c)[0x7f78ee4a470c]
/usr/lib/libstdc++.so.6(_ZNSsD1Ev+0x39)[0x7f78eecdb0c9]
/lib/libc.so.6(exit+0xe2)[0x7f78ee462c12]
/usr/lib/octave-3.2.2/liboctinterp.so(octave_main+0xe1c)[0x7f78f4dec9ac]
/lib/libc.so.6(__libc_start_main+0xfd)[0x7f78ee448abd]
octave[0x400879]
======= Memory map: ========
00400000-00401000 r-xp 00000000 08:04 142133                             /usr/bin/octave-3.2.2
00600000-00601000 r--p 00000000 08:04 142133                             /usr/bin/octave-3.2.2
00601000-00602000 rw-p 00001000 08:04 142133                             /usr/bin/octave-3.2.2
00edb000-017be000 rw-p 00000000 00:00 0                                  [heap]
410ba000-410bc000 rwxp 00000000 00:0f 1652                               /dev/zero
7f78d8000000-7f78d8021000 rw-p 00000000 00:00 0
7f78d8021000-7f78dc000000 ---p 00000000 00:00 0
7f78df7d0000-7f78df7e2000 r-xp 00000000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df7e2000-7f78df9e2000 ---p 00012000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df9e2000-7f78df9e4000 r--p 00012000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df9e4000-7f78df9e5000 rw-p 00014000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df9e5000-7f78dfa18000 r-xp 00000000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfa18000-7f78dfc18000 ---p 00033000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfc18000-7f78dfc1b000 r--p 00033000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfc1b000-7f78dfc1c000 rw-p 00036000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfc1c000-7f78dfc37000 r-xp 00000000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78dfc37000-7f78dfe37000 ---p 0001b000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78dfe37000-7f78dfe39000 r--p 0001b000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78dfe39000-7f78dfe3a000 rw-p 0001d000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78e592d000-7f78e5948000 r-xp 00000000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5948000-7f78e5b48000 ---p 0001b000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5b48000-7f78e5b4a000 r--p 0001b000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5b4a000-7f78e5b4b000 rw-p 0001d000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5b4b000-7f78e5b65000 r-xp 00000000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e5b65000-7f78e5d65000 ---p 0001a000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e5d65000-7f78e5d67000 r--p 0001a000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e5d67000-7f78e5d68000 rw-p 0001c000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e83df000-7f78e83e1000 r-xp 00000000 08:04 548                        /lib/libutil-2.10.1.so
7f78e83e1000-7f78e85e0000 ---p 00002000 08:04 548                        /lib/libutil-2.10.1.so
7f78e85e0000-7f78e85e1000 r--p 00001000 08:04 548                        /lib/libutil-2.10.1.so
7f78e85e1000-7f78e85e2000 rw-p 00002000 08:04 548                        /lib/libutil-2.10.1.so
7f78e85e2000-7f78e862f000 r-xp 00000000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e862f000-7f78e882f000 ---p 0004d000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e882f000-7f78e8830000 r--p 0004d000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e8830000-7f78e8832000 rw-p 0004e000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e8832000-7f78e8855000 rw-p 00000000 00:00 0
7f78e8855000-7f78e8898000 r-xp 00000000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8898000-7f78e8a98000 ---p 00043000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8a98000-7f78e8a99000 r--p 00043000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8a99000-7f78e8a9b000 rw-p 00044000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8a9b000-7f78e8a9d000 rw-p 00000000 00:00 0
7f78e8a9d000-7f78e8b2a000 r-xp 00000000 08:04 140189                     /usr/lib/openmpi/lib/libmpi.so.0.0.0--------------------------------------------------------------------------
mpirun noticed that process rank 2 with PID 29356 on node yosemite exited on signal 6 (Aborted).
--------------------------------------------------------------------------
2 total processes killed (some possibly by mpirun during cleanup)
panic: Aborted -- stopping myself...
attempting to save variables to `octave-core'...
panic: attempted clean up apparently failed -- aborting...
panic: Aborted -- stopping myself...
attempting to save variables to `octave-core'...
panic: attempted clean up apparently failed -- aborting...
panic: Segmentation fault -- stopping myself...
attempting to save variables to `octave-core'...
michael@yosemite:~/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/inst$

On Wed, Dec 2, 2009 at 10:35 AM, Riccardo Corradini <riccardocorradini@...> wrote:

Hello Michael,
in the svn there is a simpler but improved version. In my amd 64 machine dual core linux ubuntu jaunty no problem. No segmentation .. no errors with the new version.
I tested it with the new function MPI_Comm_Load. All the examples with montercarlo work as well.
Bests
Riccardo

--- Mar 1/12/09, Michael Creel <michael.creel@...> ha scritto:

Da: Michael Creel <michael.creel@...>
Oggetto: Re: [OctDev] A simple C++ example of a class with MPI_Comunicator property
A: "Riccardo Corradini" <riccardocorradini@...>
Data: Martedì 1 dicembre 2009, 13:49

Hi Riccardo,
You might find this screenshot amusing. This is a PelicanHPC made using Debian Sid, with Octave 3.2.3, and with your MPI bindings. I'm looking forward to trying out your fix for the comm stuff.
Cheers, MIchael

On Tue, Dec 1, 2009 at 12:43 PM, Riccardo Corradini <riccardocorradini@...> wrote:

Dear Jaroslav
after goggling a lot I found a nice example that I converted to fix the MPI_Comunicator problem

Here there is the code adjusted from http://www.codeguru.com/cpp/cpp/cpp_mfc/article.php/c4031/

there are just two files Comm.h and Comexample.cc

#define FILL_COMM(RNAME,LNAME)        \
MPI_Comm RNAME##comm = MPI_LNAME;\

#include "mpi.h"

#define READ_ONLY 1
#define WRITE_ONLY 2
#define READ_WRITE 3

template <typename Container, typename ValueType, int nPropType>
class property
{
public:
property()
{
m_cObject = NULL;
Set = NULL;
Get = NULL;
}
//-- This to set a pointer to the class that contain the
//   property --
void setContainer(Container* cObject)
{
m_cObject = cObject;
}
//-- Set the set member function that will change the value --
void setter(void (Container::*pSet)(ValueType value))
{
if((nPropType == WRITE_ONLY) || (nPropType == READ_WRITE))
    Set = pSet;
else
    Set = NULL;
}
//-- Set the get member function that will retrieve the value --
void getter(ValueType (Container::*pGet)())
{
if((nPropType == READ_ONLY) || (nPropType == READ_WRITE))
    Get = pGet;
else
    Get = NULL;
}
//-- Overload the '=' sign to set the value using the set
//   member --
ValueType operator =(const ValueType& value)
{
assert(m_cObject != NULL);
assert(Set != NULL);
(m_cObject->*Set)(value);
return value;
}
//-- To make possible to cast the property class to the
//   internal type --
operator ValueType()
{
assert(m_cObject != NULL);
assert(Get != NULL);
return (m_cObject->*Get)();
}
private:
Container* m_cObject; //-- Pointer to the module that
                         //   contains the property --
void (Container::*Set)(ValueType value);
                         //-- Pointer to set member function --
ValueType (Container::*Get)();
                         //-- Pointer to get member function --
};

class PropTest
{
public:
PropTest()
{
    Comm.setContainer(this);
    Comm.setter(&PropTest::setComm);
    Comm.getter(&PropTest::getComm);
}
MPI_Comm getComm()
{
    return m_nComm;
}
void setComm(MPI_Comm nComm)
{
    m_nComm = nComm;
}
property<PropTest,MPI_Comm,READ_WRITE> Comm;

private:
MPI_Comm m_nComm;
};

And the DLD function
#include <octave/oct.h>
#include "Comm.h"
DEFUN_DLD(CommExample,args,,"A simple example using a C++ class holding Comunicator as property")
{

// Example setting
PropTest SetTest;
SetTest.Comm = MPI_COMM_WORLD;
// Example getting
MPI_Comm ex_comm;
ex_comm = SetTest.Comm;
// the problem is how could SetTest comunicate with GNU Octave ?
}

I think the situation here is simpler because we have a simple class holding all pieces of info and perhaps the pointers to this class are easier to handle.. and more universal among different MPI implementations ...
What do you think ?
Thanks a lot in advance for all your patience and help.
Bests
Riccardo

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev

_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: File 'jni.h' missing in java-1.2.6.tar.gz

2009-12-02T02:48:02Z

Hi Alex,

That's a request for the Windows binary packager (Benjamin).
I'm not providing binary packages anymore. Note that this is a
time-consuming job. I guess/hope that Benjamin will eventually
include the java package within the binary release.

Michael.

On Wed, Dec 2, 2009 at 9:11 AM, AlexG1 <alxgel@...> wrote:

>
> Hi Michael,
>
> I'm also having this problem with Octave 3.2.3 (on Windows Vista platform)
> and couldn't make it work.
> If you remember about a month ago we exchanged a few mails about trying to
> make the package work with Octave 3.2.2,with no success.
>
> I'm wondering if it's possible to make the java package part of the Octave
> installation (like it was with Octave 3.0 where it worked great), as I've
> seen a few other posts of people struggling to install the package.
> --
> View this message in context: http://old.nabble.com/File-%27jni.h%27-missing-in-java-1.2.6.tar.gz-tp26583091p26605723.html
> Sent from the octave-dev mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> Join us December 9, 2009 for the Red Hat Virtual Experience,
> a free event focused on virtualization and cloud computing.
> Attend in-depth sessions from your desk. Your couch. Anywhere.
> http://p.sf.net/sfu/redhat-sfdev2dev
> _______________________________________________
> Octave-dev mailing list
> Octave-dev@...
> https://lists.sourceforge.net/lists/listinfo/octave-dev
>

Re: A simple C++ example of a class with MPI_Comunicatorproperty

2009-12-02T02:41:16Z

Hi Riccardo,
I'm still seeing problems with this, both for helloworld.m and for mc_example2.m. This is with Octave 3.2.2 and Open MPI 1.3.2 from Ubuntu 9.10, AMD64.

Below is the output I get from helloworld. Please let me know if I can do anything to generate useful information about this.

Thanks,
Michael

michael@yosemite:~/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/inst$ mpirun -np 3 octave -q --eval helloworld
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 0
p = 3
We are at rank 0 that is master etc..
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 1
p = 3
info for sending is is
info = 0
Variables in the current scope:

Attr Name        Size                     Bytes Class
==== ====        ====                     ===== =====
       CW         -1x-1                         0 simple

Total is 1 element using 0 bytes

my_rank = 2
p = 3
info for sending is is
info = 0
Greetings from process: 1!
We are at rank 0 that is master etc..
Greetings from process: 2!
*** glibc detected *** octave: double free or corruption (fasttop): 0x00000000017b18c0 ***
======= Backtrace: =========
/lib/libc.so.6[0x7f78ee49fdd6]
/lib/libc.so.6(cfree+0x6c)[0x7f78ee4a470c]
/usr/lib/libstdc++.so.6(_ZNSsD1Ev+0x39)[0x7f78eecdb0c9]
/lib/libc.so.6(exit+0xe2)[0x7f78ee462c12]
/usr/lib/octave-3.2.2/liboctinterp.so(octave_main+0xe1c)[0x7f78f4dec9ac]
/lib/libc.so.6(__libc_start_main+0xfd)[0x7f78ee448abd]
octave[0x400879]
======= Memory map: ========
00400000-00401000 r-xp 00000000 08:04 142133                             /usr/bin/octave-3.2.2
00600000-00601000 r--p 00000000 08:04 142133                             /usr/bin/octave-3.2.2
00601000-00602000 rw-p 00001000 08:04 142133                             /usr/bin/octave-3.2.2
00edb000-017be000 rw-p 00000000 00:00 0                                  [heap]
410ba000-410bc000 rwxp 00000000 00:0f 1652                               /dev/zero
7f78d8000000-7f78d8021000 rw-p 00000000 00:00 0
7f78d8021000-7f78dc000000 ---p 00000000 00:00 0
7f78df7d0000-7f78df7e2000 r-xp 00000000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df7e2000-7f78df9e2000 ---p 00012000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df9e2000-7f78df9e4000 r--p 00012000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df9e4000-7f78df9e5000 rw-p 00014000 08:04 574357                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Finalize.oct
7f78df9e5000-7f78dfa18000 r-xp 00000000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfa18000-7f78dfc18000 ---p 00033000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfc18000-7f78dfc1b000 r--p 00033000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfc1b000-7f78dfc1c000 rw-p 00036000 08:04 574361                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Send.oct
7f78dfc1c000-7f78dfc37000 r-xp 00000000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78dfc37000-7f78dfe37000 ---p 0001b000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78dfe37000-7f78dfe39000 r--p 0001b000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78dfe39000-7f78dfe3a000 rw-p 0001d000 08:04 574353                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_size.oct
7f78e592d000-7f78e5948000 r-xp 00000000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5948000-7f78e5b48000 ---p 0001b000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5b48000-7f78e5b4a000 r--p 0001b000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5b4a000-7f78e5b4b000 rw-p 0001d000 08:04 574351                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_rank.oct
7f78e5b4b000-7f78e5b65000 r-xp 00000000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e5b65000-7f78e5d65000 ---p 0001a000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e5d65000-7f78e5d67000 r--p 0001a000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e5d67000-7f78e5d68000 rw-p 0001c000 08:04 574356                     /home/michael/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/src/MPI_Comm_Load.oct
7f78e83df000-7f78e83e1000 r-xp 00000000 08:04 548                        /lib/libutil-2.10.1.so
7f78e83e1000-7f78e85e0000 ---p 00002000 08:04 548                        /lib/libutil-2.10.1.so
7f78e85e0000-7f78e85e1000 r--p 00001000 08:04 548                        /lib/libutil-2.10.1.so
7f78e85e1000-7f78e85e2000 rw-p 00002000 08:04 548                        /lib/libutil-2.10.1.so
7f78e85e2000-7f78e862f000 r-xp 00000000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e862f000-7f78e882f000 ---p 0004d000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e882f000-7f78e8830000 r--p 0004d000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e8830000-7f78e8832000 rw-p 0004e000 08:04 140193                     /usr/lib/openmpi/lib/libopen-pal.so.0.0.0
7f78e8832000-7f78e8855000 rw-p 00000000 00:00 0
7f78e8855000-7f78e8898000 r-xp 00000000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8898000-7f78e8a98000 ---p 00043000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8a98000-7f78e8a99000 r--p 00043000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8a99000-7f78e8a9b000 rw-p 00044000 08:04 140194                     /usr/lib/openmpi/lib/libopen-rte.so.0.0.0
7f78e8a9b000-7f78e8a9d000 rw-p 00000000 00:00 0
7f78e8a9d000-7f78e8b2a000 r-xp 00000000 08:04 140189                     /usr/lib/openmpi/lib/libmpi.so.0.0.0--------------------------------------------------------------------------
mpirun noticed that process rank 2 with PID 29356 on node yosemite exited on signal 6 (Aborted).
--------------------------------------------------------------------------
2 total processes killed (some possibly by mpirun during cleanup)
panic: Aborted -- stopping myself...
attempting to save variables to `octave-core'...
panic: attempted clean up apparently failed -- aborting...
panic: Aborted -- stopping myself...
attempting to save variables to `octave-core'...
panic: attempted clean up apparently failed -- aborting...
panic: Segmentation fault -- stopping myself...
attempting to save variables to `octave-core'...
michael@yosemite:~/Desktop/of/octave/trunk/octave-forge/extra/openmpi_ext/inst$

On Wed, Dec 2, 2009 at 10:35 AM, Riccardo Corradini <riccardocorradini@...> wrote:

Hello Michael,
in the svn there is a simpler but improved version. In my amd 64 machine dual core linux ubuntu jaunty no problem. No segmentation .. no errors with the new version.
I tested it with the new function MPI_Comm_Load. All the examples with montercarlo work as well.
Bests
Riccardo

--- Mar 1/12/09, Michael Creel <michael.creel@...> ha scritto:

Da: Michael Creel <michael.creel@...>
Oggetto: Re: [OctDev] A simple C++ example of a class with MPI_Comunicator property
A: "Riccardo Corradini" <riccardocorradini@...>
Data: Martedì 1 dicembre 2009, 13:49

Hi Riccardo,
You might find this screenshot amusing. This is a PelicanHPC made using Debian Sid, with Octave 3.2.3, and with your MPI bindings. I'm looking forward to trying out your fix for the comm stuff.
Cheers, MIchael

On Tue, Dec 1, 2009 at 12:43 PM, Riccardo Corradini <riccardocorradini@...> wrote:

Dear Jaroslav
after goggling a lot I found a nice example that I converted to fix the MPI_Comunicator problem

Here there is the code adjusted from http://www.codeguru.com/cpp/cpp/cpp_mfc/article.php/c4031/

there are just two files Comm.h and Comexample.cc

#define FILL_COMM(RNAME,LNAME)        \
MPI_Comm RNAME##comm = MPI_LNAME;\

#include "mpi.h"

#define READ_ONLY 1
#define WRITE_ONLY 2
#define READ_WRITE 3

template <typename Container, typename ValueType, int nPropType>
class property
{
public:
property()
{
m_cObject = NULL;
Set = NULL;
Get = NULL;
}
//-- This to set a pointer to the class that contain the
//   property --
void setContainer(Container* cObject)
{
m_cObject = cObject;
}
//-- Set the set member function that will change the value --
void setter(void (Container::*pSet)(ValueType value))
{
if((nPropType == WRITE_ONLY) || (nPropType == READ_WRITE))
    Set = pSet;
else
    Set = NULL;
}
//-- Set the get member function that will retrieve the value --
void getter(ValueType (Container::*pGet)())
{
if((nPropType == READ_ONLY) || (nPropType == READ_WRITE))
    Get = pGet;
else
    Get = NULL;
}
//-- Overload the '=' sign to set the value using the set
//   member --
ValueType operator =(const ValueType& value)
{
assert(m_cObject != NULL);
assert(Set != NULL);
(m_cObject->*Set)(value);
return value;
}
//-- To make possible to cast the property class to the
//   internal type --
operator ValueType()
{
assert(m_cObject != NULL);
assert(Get != NULL);
return (m_cObject->*Get)();
}
private:
Container* m_cObject; //-- Pointer to the module that
                         //   contains the property --
void (Container::*Set)(ValueType value);
                         //-- Pointer to set member function --
ValueType (Container::*Get)();
                         //-- Pointer to get member function --
};

class PropTest
{
public:
PropTest()
{
    Comm.setContainer(this);
    Comm.setter(&PropTest::setComm);
    Comm.getter(&PropTest::getComm);
}
MPI_Comm getComm()
{
    return m_nComm;
}
void setComm(MPI_Comm nComm)
{
    m_nComm = nComm;
}
property<PropTest,MPI_Comm,READ_WRITE> Comm;

private:
MPI_Comm m_nComm;
};

And the DLD function
#include <octave/oct.h>
#include "Comm.h"
DEFUN_DLD(CommExample,args,,"A simple example using a C++ class holding Comunicator as property")
{

// Example setting
PropTest SetTest;
SetTest.Comm = MPI_COMM_WORLD;
// Example getting
MPI_Comm ex_comm;
ex_comm = SetTest.Comm;
// the problem is how could SetTest comunicate with GNU Octave ?
}

I think the situation here is simpler because we have a simple class holding all pieces of info and perhaps the pointers to this class are easier to handle.. and more universal among different MPI implementations ...
What do you think ?
Thanks a lot in advance for all your patience and help.
Bests
Riccardo

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev

_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: A simple C++ example of a class with MPI_Comunicator property

2009-12-02T01:35:10Z

Hello Michael,
in the svn there is a simpler but improved version. In my amd 64 machine dual core linux ubuntu jaunty no problem. No segmentation .. no errors with the new version.
I tested it with the new function MPI_Comm_Load. All the examples with montercarlo work as well.
Bests
Riccardo

--- Mar 1/12/09, Michael Creel <michael.creel@...> ha scritto:

Da: Michael Creel <michael.creel@...>
Oggetto: Re: [OctDev] A simple C++ example of a class with MPI_Comunicator property
A: "Riccardo Corradini" <riccardocorradini@...>
Data: Martedì 1 dicembre 2009, 13:49

Hi Riccardo,
You might find this screenshot amusing. This is a PelicanHPC made using Debian Sid, with Octave 3.2.3, and with your MPI bindings. I'm looking forward to trying out your fix for the comm stuff.
Cheers, MIchael

On Tue, Dec 1, 2009 at 12:43 PM, Riccardo Corradini <riccardocorradini@...> wrote:

Dear Jaroslav
after goggling a lot I found a nice example that I converted to fix the MPI_Comunicator problem

Here there is the code adjusted from http://www.codeguru.com/cpp/cpp/cpp_mfc/article.php/c4031/

there are just two files Comm.h and Comexample.cc

#define FILL_COMM(RNAME,LNAME)        \
MPI_Comm RNAME##comm = MPI_LNAME;\

#include "mpi.h"

#define READ_ONLY 1
#define WRITE_ONLY 2
#define READ_WRITE 3

template <typename Container, typename ValueType, int nPropType>
class property
{
public:
property()
{
m_cObject = NULL;
Set = NULL;
Get = NULL;
}
//-- This to set a pointer to the class that contain the
//   property --
void setContainer(Container* cObject)
{
m_cObject = cObject;
}
//-- Set the set member function that will change the value --
void setter(void (Container::*pSet)(ValueType value))
{
if((nPropType == WRITE_ONLY) || (nPropType == READ_WRITE))
    Set = pSet;
else
    Set = NULL;
}
//-- Set the get member function that will retrieve the value --
void getter(ValueType (Container::*pGet)())
{
if((nPropType == READ_ONLY) || (nPropType == READ_WRITE))
    Get = pGet;
else
    Get = NULL;
}
//-- Overload the '=' sign to set the value using the set
//   member --
ValueType operator =(const ValueType& value)
{
assert(m_cObject != NULL);
assert(Set != NULL);
(m_cObject->*Set)(value);
return value;
}
//-- To make possible to cast the property class to the
//   internal type --
operator ValueType()
{
assert(m_cObject != NULL);
assert(Get != NULL);
return (m_cObject->*Get)();
}
private:
Container* m_cObject; //-- Pointer to the module that
                         //   contains the property --
void (Container::*Set)(ValueType value);
                         //-- Pointer to set member function --
ValueType (Container::*Get)();
                         //-- Pointer to get member function --
};

class PropTest
{
public:
PropTest()
{
    Comm.setContainer(this);
    Comm.setter(&PropTest::setComm);
    Comm.getter(&PropTest::getComm);
}
MPI_Comm getComm()
{
    return m_nComm;
}
void setComm(MPI_Comm nComm)
{
    m_nComm = nComm;
}
property<PropTest,MPI_Comm,READ_WRITE> Comm;

private:
MPI_Comm m_nComm;
};

And the DLD function
#include <octave/oct.h>
#include "Comm.h"
DEFUN_DLD(CommExample,args,,"A simple example using a C++ class holding Comunicator as property")
{

// Example setting
PropTest SetTest;
SetTest.Comm = MPI_COMM_WORLD;
// Example getting
MPI_Comm ex_comm;
ex_comm = SetTest.Comm;
// the problem is how could SetTest comunicate with GNU Octave ?
}

I think the situation here is simpler because we have a simple class holding all pieces of info and perhaps the pointers to this class are easier to handle.. and more universal among different MPI implementations ...
What do you think ?
Thanks a lot in advance for all your patience and help.
Bests
Riccardo

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev

_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: File 'jni.h' missing in java-1.2.6.tar.gz

2009-12-02T01:11:44Z

Hi Michael,

I'm also having this problem with Octave 3.2.3 (on Windows Vista platform) and couldn't make it work.
If you remember about a month ago we exchanged a few mails about trying to make the package work with Octave 3.2.2,with no success.

I'm wondering if it's possible to make the java package part of the Octave installation (like it was with Octave 3.0 where it worked great), as I've seen a few other posts of people struggling to install the package.

Re: File 'jni.h' missing in java-1.2.6.tar.gz

2009-12-01T12:42:44Z

jni.h is part of Java Development Kit (JDK). You need it
installed in order to compile the java package. If it's installed,
you may need to add the JDK include dirs to CPPFLAGS
with the -I compilation flag.

Michael.

On Mon, Nov 30, 2009 at 10:19 PM, william.zink@...
<william.zink@...> wrote:

> All,
>
> My latest attempt to install the java package (using the 3.2.3 binary on Vista) fails at line 1253 in file 'pkg.m'. Apparently, the file 'jni.h' is not in the tar file (see error snippet).
>
>
> In file included from __java__.cc:17:
> __java__.h:22:17: error: jni.h: No such file or directory
> In file included from __java__.cc:17:
> __java__.h:34: error: expected ')' before '*' token
> __java__.h:38: error: expected ')' before '*' token
> __java__.h:75: error: ISO C++ forbids declaration of 'JNIEnv' with no type
>
> Experts please advise. Do I need a corrected 'java-1.2.5.tar.gz' file?
>
> Thanks,
>
> BZ
>
> ____________________________________________________________
> Doctorate Degrees Online
> Boost your career with an online doctoral degree. Enroll today!
> http://thirdpartyoffers.juno.com/TGL2131/c?cp=247Bvh_Ms4pnbNdVn2NW1AAAJz3uVYjR4RyDRi3M6ulKIXhDAAQAAAAFAAAAALpclT4AAAMlAAAAAAAAAAAAAAAAAAAyOQAAAAA=
>
> ------------------------------------------------------------------------------
> Join us December 9, 2009 for the Red Hat Virtual Experience,
> a free event focused on virtualization and cloud computing.
> Attend in-depth sessions from your desk. Your couch. Anywhere.
> http://p.sf.net/sfu/redhat-sfdev2dev
> _______________________________________________
> Octave-dev mailing list
> Octave-dev@...
> https://lists.sourceforge.net/lists/listinfo/octave-dev
>

Re: A simple C++ example of a class with MPI_Comunicator property

2009-12-01T04:06:50Z

On Tue, Dec 1, 2009 at 12:43 PM, Riccardo Corradini
<riccardocorradini@...> wrote:

>
> Dear Jaroslav
> after goggling a lot I found a nice example that I converted to fix the MPI_Comunicator problem
>
> Here there is the code adjusted from http://www.codeguru.com/cpp/cpp/cpp_mfc/article.php/c4031/
>
> there are just two files Comm.h and Comexample.cc
>
> #define FILL_COMM(RNAME,LNAME)        \
> MPI_Comm RNAME##comm = MPI_LNAME;\
>
> #include "mpi.h"
>
>
>
> #define READ_ONLY 1
> #define WRITE_ONLY 2
> #define READ_WRITE 3
>
> template <typename Container, typename ValueType, int nPropType>
> class property
> {
> public:
> property()
> {
> m_cObject = NULL;
> Set = NULL;
> Get = NULL;
> }
> //-- This to set a pointer to the class that contain the
> //   property --
> void setContainer(Container* cObject)
> {
> m_cObject = cObject;
> }
> //-- Set the set member function that will change the value --
> void setter(void (Container::*pSet)(ValueType value))
> {
> if((nPropType == WRITE_ONLY) || (nPropType == READ_WRITE))
>     Set = pSet;
> else
>     Set = NULL;
> }
> //-- Set the get member function that will retrieve the value --
> void getter(ValueType (Container::*pGet)())
> {
> if((nPropType == READ_ONLY) || (nPropType == READ_WRITE))
>     Get = pGet;
> else
>     Get = NULL;
> }
> //-- Overload the '=' sign to set the value using the set
> //   member --
> ValueType operator =(const ValueType& value)
> {
> assert(m_cObject != NULL);
> assert(Set != NULL);
> (m_cObject->*Set)(value);
> return value;
> }
> //-- To make possible to cast the property class to the
> //   internal type --
> operator ValueType()
> {
> assert(m_cObject != NULL);
> assert(Get != NULL);
> return (m_cObject->*Get)();
> }
> private:
> Container* m_cObject; //-- Pointer to the module that
>                          //   contains the property --
> void (Container::*Set)(ValueType value);
>                          //-- Pointer to set member function --
> ValueType (Container::*Get)();
>                          //-- Pointer to get member function --
> };
>
>
> class PropTest
> {
> public:
> PropTest()
> {
>     Comm.setContainer(this);
>     Comm.setter(&PropTest::setComm);
>     Comm.getter(&PropTest::getComm);
> }
> MPI_Comm getComm()
> {
>     return m_nComm;
> }
> void setComm(MPI_Comm nComm)
> {
>     m_nComm = nComm;
> }
> property<PropTest,MPI_Comm,READ_WRITE> Comm;
>
>
> private:
> MPI_Comm m_nComm;
> };
>
>
> And the DLD function
> #include <octave/oct.h>
> #include "Comm.h"
> DEFUN_DLD(CommExample,args,,"A simple example using a C++ class holding Comunicator as property")
> {
>
> // Example setting
> PropTest SetTest;
> SetTest.Comm = MPI_COMM_WORLD;
> // Example getting
> MPI_Comm ex_comm;
> ex_comm = SetTest.Comm;
> // the problem is how could SetTest comunicate with GNU Octave ?
> }
>
> I think the situation here is simpler because we have a simple class holding all pieces of info and perhaps the pointers to this class are easier to handle.. and more universal among different MPI implementations ...
> What do you think ?
> Thanks a lot in advance for all your patience and help.
> Bests
> Riccardo
>
>

Hi Riccardo,

unfortunately I still don't have time for MPI, so do whatever you want
now, I'll throw my ideas in later.

--
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

A simple C++ example of a class with MPI_Comunicator property

2009-12-01T03:43:31Z

Dear Jaroslav
after goggling a lot I found a nice example that I converted to fix the MPI_Comunicator problem

Here there is the code adjusted from http://www.codeguru.com/cpp/cpp/cpp_mfc/article.php/c4031/

there are just two files Comm.h and Comexample.cc

#define FILL_COMM(RNAME,LNAME)        \
MPI_Comm RNAME##comm = MPI_LNAME;\

#include "mpi.h"

#define READ_ONLY 1
#define WRITE_ONLY 2
#define READ_WRITE 3

template <typename Container, typename ValueType, int nPropType>
class property
{
public:
property()
{
m_cObject = NULL;
Set = NULL;
Get = NULL;
}
//-- This to set a pointer to the class that contain the
//   property --
void setContainer(Container* cObject)
{
m_cObject = cObject;
}
//-- Set the set member function that will change the value --
void setter(void (Container::*pSet)(ValueType value))
{
if((nPropType == WRITE_ONLY) || (nPropType == READ_WRITE))
    Set = pSet;
else
    Set = NULL;
}
//-- Set the get member function that will retrieve the value --
void getter(ValueType (Container::*pGet)())
{
if((nPropType == READ_ONLY) || (nPropType == READ_WRITE))
    Get = pGet;
else
    Get = NULL;
}
//-- Overload the '=' sign to set the value using the set
//   member --
ValueType operator =(const ValueType& value)
{
assert(m_cObject != NULL);
assert(Set != NULL);
(m_cObject->*Set)(value);
return value;
}
//-- To make possible to cast the property class to the
//   internal type --
operator ValueType()
{
assert(m_cObject != NULL);
assert(Get != NULL);
return (m_cObject->*Get)();
}
private:
Container* m_cObject; //-- Pointer to the module that
                         //   contains the property --
void (Container::*Set)(ValueType value);
                         //-- Pointer to set member function --
ValueType (Container::*Get)();
                         //-- Pointer to get member function --
};

class PropTest
{
public:
PropTest()
{
    Comm.setContainer(this);
    Comm.setter(&PropTest::setComm);
    Comm.getter(&PropTest::getComm);
}
MPI_Comm getComm()
{
    return m_nComm;
}
void setComm(MPI_Comm nComm)
{
    m_nComm = nComm;
}
property<PropTest,MPI_Comm,READ_WRITE> Comm;

private:
MPI_Comm m_nComm;
};

And the DLD function
#include <octave/oct.h>
#include "Comm.h"
DEFUN_DLD(CommExample,args,,"A simple example using a C++ class holding Comunicator as property")
{

// Example setting
PropTest SetTest;
SetTest.Comm = MPI_COMM_WORLD;
// Example getting
MPI_Comm ex_comm;
ex_comm = SetTest.Comm;
// the problem is how could SetTest comunicate with GNU Octave ?
}

I think the situation here is simpler because we have a simple class holding all pieces of info and perhaps the pointers to this class are easier to handle.. and more universal among different MPI implementations ...
What do you think ?
Thanks a lot in advance for all your patience and help.
Bests
Riccardo

File 'jni.h' missing in java-1.2.6.tar.gz

2009-11-30T14:19:32Z

All,

My latest attempt to install the java package (using the 3.2.3 binary on Vista) fails at line 1253 in file 'pkg.m'. Apparently, the file 'jni.h' is not in the tar file (see error snippet).

In file included from __java__.cc:17:
__java__.h:22:17: error: jni.h: No such file or directory
In file included from __java__.cc:17:
__java__.h:34: error: expected ')' before '*' token
__java__.h:38: error: expected ')' before '*' token
__java__.h:75: error: ISO C++ forbids declaration of 'JNIEnv' with no type

Experts please advise. Do I need a corrected 'java-1.2.5.tar.gz' file?

Thanks,

BZ

____________________________________________________________
Doctorate Degrees Online
Boost your career with an online doctoral degree. Enroll today!
http://thirdpartyoffers.juno.com/TGL2131/c?cp=247Bvh_Ms4pnbNdVn2NW1AAAJz3uVYjR4RyDRi3M6ulKIXhDAAQAAAAFAAAAALpclT4AAAMlAAAAAAAAAAAAAAAAAAAyOQAAAAA=

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Gnuplot bug - mingw-octave

2009-11-29T21:07:01Z

Hello

--- Tatsuro MATSUOKA
> It is better to ask the problem at the gnuplot ML.
> Before you ask the problem, please try the same script at the latest gnuplot-4.5 (cvs)
> (2009-11-26)
> at
> http://www.tatsuromatsuoka.com/gnuplot/Eng/winbin/
>

Release candidate 1 for gnuplot 4.4 has been uploaded.
Please also test this newest release.

http://sourceforge.net/projects/gnuplot/files/
gp440rc1_win.zip

Regards

Tatsuro

--------------------------------------
Learn more about HIV/AIDS - Red Ribbon Campaign 2009
http://pr.mail.yahoo.co.jp/redribbon/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev

Re: Gnuplot bug - mingw-octave

2009-11-29T15:10:30Z

Hello

--- Bernd Heinzewrote:
>
> I tried using gnuplot directly and it doesn't work there as well, so it's apparently a gnuplot
> issue, annoying nontheless.
>
It is better to ask the problem at the gnuplot ML.
Before you ask the problem, please try the same script at the latest gnuplot-4.5 (cvs) (2009-11-26)
at
http://www.tatsuromatsuoka.com/gnuplot/Eng/winbin/

The gnuplot bundled with octave is rather bit old (4.3.0-2009-07-08) but modified for octave.

Regards

Tatsuro

--------------------------------------
What is your No.1 Entertainment of 2009? -Yahoo! JAPAN Net BANZUKE 2009
http://pr.mail.yahoo.co.jp/banzuke/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Octave-dev mailing list
Octave-dev@...
https://lists.sourceforge.net/lists/listinfo/octave-dev