plplot-general — PLplot user discussion group

Re: [Plplot-general] [Plplot-devel] Odd problem with major tickmarks/lines

[Arjen M. <arj...@wl...>]

Arjen Markus wrote:
> 
> Hello,
> 
> I am drawing an axis around a plot using plbox.
> 
> I want a spacing of the major tickmarks of 0.2 and the range is from 0.0
> to 1.6.
> So I expect tickmarks at 0.0, 0.2, ..., 1.6. I do get these.
> 
> But to my surprise I also get marks at 0.5, 1.0 and 1.5! (These occur
> also
> when I set the interval to 0 - automatic determination of the tickmark
> spacing)
> 
> I know they are major tickmarks, because when I request ticlines, all
> are drawn as lines.
> 
> What should I do about this?
> 
> Regards,
> 
> Arjen
> 

I apologize: a stupid mistake on my part. I use plbox twice to get 
all the settings right and one of the calls uses a tickmark spacing 
of zero ...

Regards,

Arjen

[Plplot-general] [Plplot-devel] Odd problem with major tickmarks/lines

[Arjen M. <arj...@wl...>]

Hello,

I am drawing an axis around a plot using plbox.

I want a spacing of the major tickmarks of 0.2 and the range is from 0.0
to 1.6.
So I expect tickmarks at 0.0, 0.2, ..., 1.6. I do get these.

But to my surprise I also get marks at 0.5, 1.0 and 1.5! (These occur
also
when I set the interval to 0 - automatic determination of the tickmark
spacing)

I know they are major tickmarks, because when I request ticlines, all
are drawn as lines.

What should I do about this?

Regards,

Arjen

[Plplot-general] Re: Second candidate tarball (RC2) for PLplot 5.3.1 is released

[Rafael L. <rla...@us...>]

* Rafael Laboissiere <rla...@us...> [2004-06-18 08:15]:

> The second release candidate tarball for the upcoming 5.3.1 version of
> PLplot is out.

I forgot to attach the CVS ChangeLog.  You will find it below.

-- 
Rafael

[Plplot-general] Second candidate tarball (RC2) for PLplot 5.3.1 is released

[Rafael L. <rla...@us...>]

The second release candidate tarball for the upcoming 5.3.1 version of
PLplot is out.  It is available at:

    http://plplot.sourceforge.net/cvs-tarball/

The tarball is called plplot-5.3.0.rc2.5.3.1.tar.gz.  A couple of bugs are
fixed in this tarball, mainly related to the Octave binding, which now fully
works with Octave version 2.1.57.

This tarball should be very close to the final 5.3.1 release, scheduled for
the end of June. Reports on build and installation are welcome.

-- 
Rafael, Release Manager

[Plplot-general] First candidate tarball for 5.3.1 is released

[Rafael L. <rla...@us...>]

I am pleased to announce the first release candidate tarball for the
upcoming 5.3.1 version of PLplot.  It is available at:

    http://plplot.sourceforge.net/cvs-tarball/

The tarball is called plplot-5.3.0.rc1.5.3.1.tar.gz and has several
improvements as regards to the last CVS snapshot tarball
(5.3.0.cvs.20040520), witness the CVS ChangeLog attached below.

If everything goes well, we will be releasing the final 5.3.1 tarball around
the end of June.  

Reports on build and installation are welcome.

PLplotly yours,

-- 
Rafael, release manager

[Plplot-general] gd-related failure to install on Mac OS X (Fink)

[Duncan H. <ha...@Pr...>]

Hi

I have been trying to install the plplot  package on Mac OS X
10.3 using fink.

There is some problem with the gd library I cannot resolve:
get-drv-info  gd   fails always

creating get-drv-info
./get-drv-info `echo dg300.la | sed 's/.la//'` > dg300.rc
./get-drv-info `echo gd.la | sed 's/.la//'` > gd.rc
ltdl.c:2460: failed assertion `dirname'
make[2]: *** [gd.rc] Error 134
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
### execution of make failed, exit code 2
Failed: compiling plplot-5.3.0-12 failed

This happens when either the gd209  (gd-2.0.9-13) or the gd2  
(gd-2.0.23-25)
fink-10.3/unstable packages are installed.

Any ideas?

If anyone is able to install plplot from fink (please check if it still 
works)
could they let me know what gd packages they are using?

I would guess this  is a packaging problem, but I have no idea.

The problem is there when I run
"get-drv-info XX" with XX=gd
(XX=dg300 works)

(I tried putting strategic printf's in ltdl,.c and rebuilding it
and  get-drv-info, to see if I coulds understand what the
problem was, but that didnt clarify things...)

Thanks for any suggestions
(an in particular, any confirmation that someone
else can install plplot from fink/10.3/unstable today)

Duncan Haldane
  

Re: [Plplot-general] newbie question

[Arjen M. <arj...@wl...>]

Ali Muhammad wrote:
> 
> Hi Alan,
> 
> I think I did manage to generate the dll. Now I am trying to run the example
> programs. I am getting following when I run.
> 
> C:\>java plplot/examples/x01
> 
> Plotting Options:
>  < 1> win3       PLplot Win32 Window
>  < 2> plmeta     PLplot Native Meta-File
>  < 3> ps         PostScript File (monochrome)
>  < 4> psc        PostScript File (color)
>  < 5> null       Null device
> 
> Enter device number or keyword: 1
> 
> *** PLPLOT ERROR ***
> Unable to open or allocate memory for font file
> Program aborted
> 
> Please help
> 

I think this is due to the font files not being found.
The quickest way: copy the *.fnt files you find in the data
directory into the directory where you start your program
(there are also two environment variable you can set:
PLPLOT_LIB and PLPLOT_HOME). The new chapter on deployment
in the documentation holds more information about this.



Regards,

Arjen

Re: [Plplot-general] newbie question

[Ali M. <so...@sy...>]

Hi Alan,

I think I did manage to generate the dll. Now I am trying to run the example
programs. I am getting following when I run.

C:\>java plplot/examples/x01

Plotting Options:
 < 1> win3       PLplot Win32 Window
 < 2> plmeta     PLplot Native Meta-File
 < 3> ps         PostScript File (monochrome)
 < 4> psc        PostScript File (color)
 < 5> null       Null device

Enter device number or keyword: 1

*** PLPLOT ERROR ***
Unable to open or allocate memory for font file
Program aborted

Please help

Thank you

-Ali

----- Original Message ----- 
From: "Alan W. Irwin" <ir...@be...>
To: "Ali Muhammad" <so...@sy...>
Sent: Tuesday, May 18, 2004 8:49 PM
Subject: Re: [Plplot-general] newbie question


> On 2004-05-18 18:23-0400 Ali Muhammad wrote:
>
> > Dear Alan,
> >
> > Thank you for the advice. It was very useful. It helped me understand
the
> > build process. I did manage to generate the dll and also managed to
compile
> > the example java files. Right now I am struggling with running these
> > examples.
> >
> > C:\>java plplot.examples.x01
> > Native code library failed to load. See the chapter on Dynamic Linking
> > Problems
> > in the SWIG Java documentation for help.
> > java.lang.UnsatisfiedLinkError:
> > C:\plplot\plplot-5.3.0\sys\win32\msdev\plpdll: A
> > ccess is denied
> >
>
> Hi Ali:
>
> Sounds like you have made a lot of progress, and you are close to success.
>
> If that were Linux (which of course it isn't), I would say that error
> message means the file name doesn't exist (wrong spelling or directory
> location in config.java?), or it has the wrong permissions, or it was
built
> from the wrong source (plpdll should be built from plplotjavac_wrap.c), or
> it wasn't built quite right for a native dll that can be dynamically
loaded
> by java. If the first three simple explanations don't apply, it is
probably
> time to hit the java native dll tutorials for Windows to find out exactly
> what compile and link options you need for a dll loaded by the System.load
> command.  (That command is in PLStreamc.java, and you will note that it
uses
> libname, the variable you define in config.java.)
>
> Google found me some Linux java tutorials that mentioned System.load
> so I assume there are also similar tutorials for Windows.
>
> Good luck getting this final step to work and documenting the steps you
took.
>
> Furthermore, once you get java to work on Windows, I hope you can also
give
> me some advice on how to make a user-friendly interface. See my comments
> about this in bindings/java/README.javaAPI.  The truth is my java
experience
> is minimal and consists of rather mechanical extension of the java
examples
> based on what was done before. I inherited the hand-crafted interface from
> somebody else, but I knew enough about swig so I could generalize that
> hand-crafted approach quite easily with swig.  So I am proud we have a
Java
> interface and want to maintain it, but that interface and making sure all
> the examples work under Linux is the totality of my Java experience.
Thus,
> I am definitely open to advice from developers more expert than me in the
> Java language.
>
> Alan
> __________________________
> Alan W. Irwin
> email: ir...@be...
> phone: 250-727-2902
>
> Astronomical research affiliation with Department of Physics and
Astronomy,
> University of Victoria (astrowww.phys.uvic.ca).
>
> Programming affiliations with the PLplot scientific plotting software
> package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
> Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
> (lbproject.sf.net).
> __________________________
>
> Linux-powered Science
> __________________________

Re: [Plplot-general] newbie question

[Alan W. I. <ir...@be...>]

On 2004-05-17 21:41-0400 Ali Muhammad wrote:

> I am trying to build java binding on windows. The documentation provided
only talks about unix. I will appreciate it somebody can provide help.

I think building java binding on windows is possible since the java
interface C file and other essential java interface wrappers that are
generated by swig are all part of the tarball.  But as far as I know, a Java
interface to PLplot on windows has never been done before.  Also, I have no
windows expertise so you will have to infer what to do on the Windows side
from what is currently done on the Linux/Unix side of things which I will
now explain.

What is done with the tarball java files is pretty straightforward on
the Linux side so it shouldn't be that difficult to follow. In particular
our installation procedure creates the following tree in the install area.

./plplot/core/README.javaAPI
./plplot/core/PLStreamc.java
./plplot/core/config.java
./plplot/core/SWIGTYPE_p_p_char.java
./plplot/core/plplotjavac.java
./plplot/core/plplotjavacJNI.java
./plplot/core/plplotjavacConstants.java
./plplot/core/plplotjavac_wrap.so
./plplot/core/PLStreamc.class
./plplot/core/SWIGTYPE_p_p_char.class
./plplot/core/config.class
./plplot/core/plplotjavac.class
./plplot/core/plplotjavacConstants.class
./plplot/core/plplotjavacJNI.class
./plplot/examples/README.javademos
./plplot/examples/x01.java
./plplot/examples/x02.java
./plplot/examples/x03.java
./plplot/examples/x04.java
./plplot/examples/x05.java
./plplot/examples/x06.java
./plplot/examples/x07.java
./plplot/examples/x08.java
./plplot/examples/x09.java
./plplot/examples/x10.java
./plplot/examples/x11.java
./plplot/examples/x12.java
./plplot/examples/x13.java
./plplot/examples/x14.java
./plplot/examples/x15.java
./plplot/examples/x16.java
./plplot/examples/x17.java
./plplot/examples/x18.java
./plplot/examples/x19.java
./plplot/examples/x01.class
./plplot/examples/x02.class
./plplot/examples/x03.class
./plplot/examples/x04.class
./plplot/examples/x05.class
./plplot/examples/x06.class
./plplot/examples/x07.class
./plplot/examples/x08.class
./plplot/examples/x09.class
./plplot/examples/x10.class
./plplot/examples/x11.class
./plplot/examples/x12.class
./plplot/examples/x13.class
./plplot/examples/x14.class
./plplot/examples/x15.class
./plplot/examples/x16.class
./plplot/examples/x17.class
./plplot/examples/x18.class
./plplot/examples/x19.class

I assume you mostly know on the Windows side of thing how to create a
similar directory tree using the swig-generated java files from the tarball.
The only difficult file I am aware of is ./plplot/core/plplotjavac_wrap.so.
That is the actual wrapper library under Linux is created with the following
two commands under Linux on my system which happens to have a peculiar
location for java, hence all the strange -I options.

gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include -I../../include \
-I../../libltdl -I/home/software/java/IBMJava2-14/include \
-I/home/software/java/IBMJava2-14/include -g -O2 -mieee-fp -c \
plplotjavac_wrap.c -Wp,-MD,.deps/plplotjavac_wrap.TPlo  -fPIC -DPIC -o \
.libs/plplotjavac_wrap.o

gcc -shared  .libs/plplotjavac_wrap.o  -Wl,--rpath \
-Wl,/usr/local/plplot_at/lib \
-L/home/software/plplot_cvs/HEAD/plplot_devel/lib/nn/.libs \
-L/home/software/plplot_cvs/HEAD/plplot_devel/lib/csa/.libs -L/usr/lib/ \
-L/usr/local/plplot_at/lib -lplplotd -lm  -mieee-fp -Wl,-soname \
-Wl,plplotjavac_wrap.so -o .libs/plplotjavac_wrap.so

The first command compiles the tarball-supplied plplotjavac_wrap.c, and the
second command links the result into a shared object (I think you call that
a DLL on the windows side of things).

So you should be okay if you know how to compile and link on windows.  My
best guess for those steps is you should follow what is done to compile and
link libplplot itself on windows, and then install the result in the above
tree.  Then you will have to edit config.java.in from the tarball to point
to the correct absolute location of the shared object (or DLL?) you just installed.
For example, after Unix configuration (which puts the configured result from
config.java.in into config.java, my resulting config.java file has the
following line:

public static String libname = "/usr/local/plplot_at/lib/java/plplot/core/plplotjavac_wrap.so";

(My particular absolute installation prefix for the above plplot java
interface tree is /usr/local/plplot_at/lib/java so you will have to adjust
the line in your own config.java according to where you have installed the
above plplot tree containing the java interface.)

Finally, you will have to java compile all the *.java files in the above
tree that you copied from the tarball (or edited from config.java.in for the
config.java case).  But that should be straightforward so the only difficult
part is compiling and linking the shared object, and I hope I have given you
enough details on that so it should be straightforward for you.

Anyhow, good luck, and if you get the java interface to PLplot to work on
windows, please document what you did so we can add it to the windows
instructions for PLplot to make it easier for other windows users to follow
what you have done.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

[Plplot-general] newbie question

[Ali M. <so...@sy...>]

Hi,

I am trying to build java binding on windows. The documentation provided =
only talks about unix. I will appreciate it somebody can provide help.=20

Thanks

-Ali

Re: [Plplot-general] contour plotting problem

[Andrew R. <and...@us...>]

On Fri, May 14, 2004 at 07:03:29AM -0700, Alan Irwin wrote:
> On 2004-05-14 12:56+0200 Joachim Geiger wrote:
> 
> > Thanks, this indeed solves the problem.
> 
> That's great that Andrew Ross's new, contouring algorithm solved your
> problem.

I'm pleased to hear it too. Please let me know if you find any other
problems with the new scheme. It seems to be a definite improvement on
the old scheme but there may still be some teething problems.

Andrew

Re: [Plplot-general] contour plotting problem

[Alan W. I. <ir...@be...>]

On 2004-05-14 12:56+0200 Joachim Geiger wrote:

> Thanks, this indeed solves the problem.

That's great that Andrew Ross's new, contouring algorithm solved your
problem.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] contour plotting problem

[Joachim G. <joa...@ip...>]

Thanks, this indeed solves the problem.
Joachim

Alan W. Irwin wrote:

>On 2004-05-13 20:11+0200 Joachim Geiger wrote:
>
>  
>
>>(for version 5.3.0)
>>when you plot a function z(r) on the space x(r,theta),y(r,theta),
>>then it depends on the storage of the grid values for plcon2
>>whether you see the contour lines or not. Specifically,
>>x(r,theta) is stored in xg(i,j) (y and z the same way); if
>>i corresponds to r, then you will get nice nested contours,
>>if i corresponds to theta (thus j corresponds to r) then
>>no contours are seen.
>>    
>>
>
>To solve other contouring problems that have been found recently, there has
>been a replacement of our contouring algorithm since version 5.3.0, and it
>seems to have fixed the problem you have reported.  To be specific, for the
>latest PLplot (from cvs) your fortran code produced two identical looking
>plots with all contours plotted.
>
>Could you please try the latest cvs snapshot version of PLplot from
>http://plplot.sourceforge.net/cvs-tarball/ to see whether that fixes
>the problem for you?
>
>Alan
>__________________________
>Alan W. Irwin
>email: ir...@be...
>phone: 250-727-2902
>
>Astronomical research affiliation with Department of Physics and Astronomy,
>University of Victoria (astrowww.phys.uvic.ca).
>
>Programming affiliations with the PLplot scientific plotting software
>package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
>Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
>(lbproject.sf.net).
>__________________________
>
>Linux-powered Science
>__________________________
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: SourceForge.net Broadband
>Sign-up now for SourceForge Broadband and get the fastest
>6.0/768 connection for only $19.95/mo for the first 3 months!
>http://ads.osdn.com/?ad_id=2562&alloc_id=6184&op=click
>_______________________________________________
>Plplot-general mailing list
>Plp...@li...
>https://lists.sourceforge.net/lists/listinfo/plplot-general
>
>  
>


-- 
-----------------------------------------------@\\
Joachim Geiger                                  @\\
Institute for Plasma Physics, Euratom Association\\\
Department E3                                     \\\__
Wendelsteinstr. 1                                (  \\ \
D-17491 Greifswald, Germany                       \_ \\(_
Tel.: 03834/88-2327                                 ) \\ \
e-mail: joa...@ip...                   \  -  |
---------------------------------------------------  \____/

Re: [Plplot-general] contour plotting problem

[Alan W. I. <ir...@be...>]

On 2004-05-13 20:11+0200 Joachim Geiger wrote:

> (for version 5.3.0)
> when you plot a function z(r) on the space x(r,theta),y(r,theta),
> then it depends on the storage of the grid values for plcon2
> whether you see the contour lines or not. Specifically,
> x(r,theta) is stored in xg(i,j) (y and z the same way); if
> i corresponds to r, then you will get nice nested contours,
> if i corresponds to theta (thus j corresponds to r) then
> no contours are seen.

To solve other contouring problems that have been found recently, there has
been a replacement of our contouring algorithm since version 5.3.0, and it
seems to have fixed the problem you have reported.  To be specific, for the
latest PLplot (from cvs) your fortran code produced two identical looking
plots with all contours plotted.

Could you please try the latest cvs snapshot version of PLplot from
http://plplot.sourceforge.net/cvs-tarball/ to see whether that fixes
the problem for you?

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

[Plplot-general] contour plotting problem

[Joachim G. <joa...@ip...>]

Hello,
after some days of debugging (frustration included), I think
that the contour plotting routine used for plcon2 in the fortran
interface has a bug in it. The attached example shows the problem
and is taken from the expample x09f.f of version 5.3.0 and
is the following:
when you plot a function z(r) on the space x(r,theta),y(r,theta),
then it depends on the storage of the grid values for plcon2
whether you see the contour lines or not. Specifically,
x(r,theta) is stored in xg(i,j) (y and z the same way); if
i corresponds to r, then you will get nice nested contours,
if i corresponds to theta (thus j corresponds to r) then
no contours are seen.
If there is a dependence of z on r and theta, you will not
notice this asymmetry in the coordinates.
Has anyone an idea how to fix it, so that it does not
depend on the order of the coordinates that you might or
might not see contour lines?
By the way, I love plplot!!! Really great!
Joachim

-- 
-----------------------------------------------@\\
Joachim Geiger                                  @\\
Institute for Plasma Physics, Euratom Association\\\
Department E3                                     \\\__
Wendelsteinstr. 1                                (  \\ \
D-17491 Greifswald, Germany                       \_ \\(_
Tel.: 03834/88-2327                                 ) \\ \
e-mail: joa...@ip...                   \  -  |
---------------------------------------------------  \____/

Re: [Plplot-general] Problem with error bars (segfault)

[Alan W. I. <ir...@be...>]

On 2004-05-04 20:30+0100 Tiago Pereira wrote:

> Answering to myself, problem solved.
>
> On Tue, 4 May 2004, Tiago Pereira wrote:
>
> :     plerry(n,x,y-errminus,y+errplus)
>
> When getting the same error with plfill, I suspected. Fortunately there
> are examples with plfill! The problem seems to be bad documentation. When
> I read:
>
>  plerry (n, x, ymin, ymax);
>
> I thought that plerry would take 4 arguments. Apparently, I was wrong and
> "n" isn't an argument, so plerry only takes the other 3 arguments. This
> solved my problem.

Glad you figured it out.  I agree some short documentation of the python
interface would be useful.  Until somebody does that (are there any
volunteers?), you have to look at the extensive xw??.py python examples and
follow the spirit of those.  The general rule is all array dimensions are
dropped (since that information is redundant in python).  Our java interface
also has the same general rule for the same reason.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Problem with error bars (segfault)

[Tiago P. <tm...@me...>]

Answering to myself, problem solved.

On Tue, 4 May 2004, Tiago Pereira wrote:

:     plerry(n,x,y-errminus,y+errplus)

When getting the same error with plfill, I suspected. Fortunately there
are examples with plfill! The problem seems to be bad documentation. When
I read:

 plerry (n, x, ymin, ymax);

I thought that plerry would take 4 arguments. Apparently, I was wrong and
"n" isn't an argument, so plerry only takes the other 3 arguments. This
solved my problem.

Greetings,

Tiago Pereira
---

Ph.D Student @ IST, Portugal

Web: tiago.leftones.org
OpenPGP KeyID: F5D786D9
OpenPGP Public key: http://mega.ist.utl.pt/~tmdp/tiago.gpg

[Plplot-general] Problem with error bars (segfault)

[Tiago P. <tm...@me...>]

Hi there,

	I've been using plplot for a few weeks. Today I tried to include
error bars in a plot, without much success. Every call that I make to
plerry ends up invariably in a Segmentation Fault. I have tried to do some
simpler tests to plerry, but the result is always the same. Is
plerry/plerrx working? Could anybody post a working example?
	I am calling plerry in the following manner (python syntax):

    plsdev("xwin")
    plinit()

    [some code to create the arrays]

    plenv(xmin, xmax, ymin,ymax, 0, 0)

    plpoin(x,y, 20)
    plerry(n,x,y-errminus,y+errplus)

    pleop()
    plend()

	Help would be much appreciated. I'm really stuck on this one.

Thanks in advance,

Tiago Pereira
---

Ph.D student @ IST, Portugal

Web: tiago.leftones.org
OpenPGP KeyID: F5D786D9
OpenPGP Public key: http://mega.ist.utl.pt/~tmdp/tiago.gpg

Re: [Plplot-general] Static-linked, stand-alone executables?

[Alan W. I. <ir...@be...>]

On 2004-04-04 07:45-0700 ooOoo wrote:

>
> Thanks very much for the tips and helpful
> explanations.
> The following older message was also quite
> informative:
>
> http://sourceforge.net/mailarchive/message.php?msg_id=6293797
>
> In case others have a need to make the occasional
> stand-alone program for friends, I will post what
> worked for me (note that I have been slow to make good
> use of pkg-config to manage libraries).
>
> =============================================================
>
> 1. I created a separate plplot-5.3.0 installation just
>
>    for static linking.
>
>   I turned off everything I did not need by editing
>   configure.  Here are the options I used:
>
>   with_debug=no     with_opt=yes        with_double=no
>   with_profile=no   with_warn=no        with_pkgdir=
>   with_rpath=yes    with_nobraindead=no with_csa=no
>   with_freetype=no  with_qhull=no
>   with_pthreads=no  with_ltdlsystem=no
>
>   enable_shared=no  enable_drivers=yes
>   enable_dyndrivers=no
>   enable_f77=yes    enable_cxx=yes
>   enable_python=no
>   enable_tcl=no     enable_itcl=no      enable_java=no
>   enable_octave=no  enable_gnome=no     enable_ntk=no
>   enable_tkwin=no   enable_png=no       enable_jpeg=no
>   enable_cgm=no     enable_tk=no
> enable_xwin=yes
>
>   enable_ljii=yes   enable_ljiip=yes    enable_pbm=yes
>   enable_ps=yes     enable_pstex=yes
> enable_xfig=yes
>
One amendment I would suggest.  Try

./configure --help

to see the easy way to set all the above options without having to
edit the configure script.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Static-linked, stand-alone executables?

[ooOoo <jg3...@ya...>]

Thanks very much for the tips and helpful
explanations.
The following older message was also quite
informative:

http://sourceforge.net/mailarchive/message.php?msg_id=6293797

In case others have a need to make the occasional
stand-alone program for friends, I will post what
worked for me (note that I have been slow to make good
use of pkg-config to manage libraries).

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

1. I created a separate plplot-5.3.0 installation just

   for static linking.

  I turned off everything I did not need by editing
  configure.  Here are the options I used:

  with_debug=no     with_opt=yes        with_double=no
  with_profile=no   with_warn=no        with_pkgdir=
  with_rpath=yes    with_nobraindead=no with_csa=no   
  with_freetype=no  with_qhull=no           
  with_pthreads=no  with_ltdlsystem=no

  enable_shared=no  enable_drivers=yes  
  enable_dyndrivers=no
  enable_f77=yes    enable_cxx=yes        
  enable_python=no
  enable_tcl=no     enable_itcl=no      enable_java=no
  enable_octave=no  enable_gnome=no     enable_ntk=no
  enable_tkwin=no   enable_png=no       enable_jpeg=no
  enable_cgm=no     enable_tk=no       
enable_xwin=yes

  enable_ljii=yes   enable_ljiip=yes    enable_pbm=yes
  enable_ps=yes     enable_pstex=yes   
enable_xfig=yes

  I then did "./configure --prefix=/usr/local/static",

  "make",and "make install".

2. To link my f77 programs, I found that the following
   flags worked:

   PLTLIBS = -L /usr/local/static/lib -lplplotf77 \
             -lplplotcxx -lplplot
   XTRALIBS = -I/usr/local/static/include/plplot \
              -L/usr/X11R6/lib \
              -lX11 -lieee -lm -l pthread

   ... and I included the "-static" option on my link 
   command.

  Next, I used "strip" to reduce the size (from 2.5 M 
  to 1.5 M) of the executable files (the dynamically  
  linked executables were only 55 k).

3. At first, the programs failed on a non-plplot 
   computer, giving an error about missing font files.
 
   I copied the plplot font files
   (/usr/local/static/share/plplot5.3.0/plstnd5.fnt 
   & plxtnd5.fnt) to the working directory, and then 
   the programs ran fine.

From a happy plplot user - jg.



__________________________________
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway 
http://promotions.yahoo.com/design_giveaway/

Re: [Plplot-general] plplot-octave usage

[Alan W. I. <ir...@be...>]

On 2004-04-03 18:07+0900 Valery Pipin wrote:

> Dear Alan
>
> Here is my report,
>
> First I tryied the way which you recommended. I rebuilt plplot from
> tarball plplot-5.3.0.cvs.20040324.tar.gz.  I got  the same segmentation
> error .
>
> Then I downloaded the cvs tree and again rebuilt  the library, and found
> that everything's tip-top (ok!). So something was wrong with version
> 2004032.

When I downloaded 20040324, I also confirmed the problem with the PLplot
library (the only part of PLplot used by yplot).  Further checking with
valgrind showed a memory management problem in plcont.c was causing the
segfault.  Sure enough, in the cvs log messages for plcont.c as of March 26
(two days after the tarball) we have the following commit message from
Andrew Ross:

"Fix bounds check in plcont. Also (partially) fix problem contouring
flat surface."

So to summarize, Valery, you ran into a problem for the 20040324 tarball
(confirmed by me) and in the Debian unstable packages corresponding to that
tarball that has now been fixed in cvs (confirmed by both of us).  Sorry I
didn't clue in faster. Your report is much appreciated since it highlights
how important that plcont.c fix is.

As a matter of course, that important fix will be included in the next cvs
snapshot tarball and also the corresponding set of Debian packages since
they both currently track cvs HEAD.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] plplot-octave usage

[Valery P. <pi...@is...>]

Dear Alan

Here is my report,

First I tryied the way which you recommended. I rebuilt plplot from
tarball plplot-5.3.0.cvs.20040324.tar.gz.  I got  the same segmentation 
error .

Then I downloaded the cvs tree and again rebuilt  the library, and found 
that everything's tip-top (ok!). So something was wrong with version
2004032.

best rgds,
Valery
 

Right now, The bug (or something like this)  is
Alan W. Irwin wrote:

> On 2004-04-03 13:32+0900 Valery Pipin wrote:
>
>> Dear Alan
>>
>> Thanks very much for the help. My feeling is that my code is ok
>> and it works on plplot-5.3.0-2(debian sarge)
>> Please,  find archive in attachment.
>>
>> Best rgds
>> Valery
>
>
> Hi Valery:
>
> Both versions of your code work here on my Debian stable system with
> yplot-1.1.2 and the April 1st cvs version of PLplot.  I assume you are 
> using
> that version of yplot and the March 24 version of PLplot that you are 
> using
> should differ little from mine.
>
> When I remove the commentary from co.i, everything is fine, and the 
> plot is
> more detailed with no error messages.  I also ran both versions of co.i
> using valgrind to make sure there were no memory management problems 
> (which
> possibly could lead to a hidden error on my system and a segfault on 
> yours
> depending on the history of the memory in use).  It turns out there are
> absolutely no memory management problems so we can eliminate that 
> possible
> explanation of our differing results.
>
> So I cannot easily verify on my system the bug that you have found.  What
> that means is there is probably some subtle difference between our 
> systems
> that produces the yplot (or PLplot) error on yours and not on mine so to
> pursue this further we will have to do a lot of additional testing to 
> narrow
> down the cause of the problem.
>
> If I understand your message you are using the Debian unstable 
> packages for
> PLplot.  It is unlikely but still possible there are some problems 
> with that
> packaging.  To test that, try uninstalling it (apt-get --purge remove
> pkgname), and then do an ordinary build of 
> plplot-5.3.0.cvs.20040324.tar.gz
> taken from http://plplot.sourceforge.net/cvs-tarball/.  Before the build,
> please check there are no transmission errors when you download that 
> tarball
> by using gpg --verify plplot-5.3.0.cvs.20040324.tar.gz.asc (That detached
> signature file for plplot-5.3.0.cvs.20040324.tar.gz can also be 
> downloaded
> from the same place.)
>
> Let me know how it goes with the pure tarball approach.  If you still get
> the same error, then I will try the pure tarball approach here also, 
> and you
> may want to try other systems as well to test the possibility there is a
> system bug (gcc compiler or whatever) you are running into for the Debian
> unstable version that won't be present for other systems.
>
> Good luck, and let me know how it goes.
>
> Alan
> __________________________
> Alan W. Irwin
> email: ir...@be...
> phone: 250-727-2902
>
> Astronomical research affiliation with Department of Physics and 
> Astronomy,
> University of Victoria (astrowww.phys.uvic.ca).
>
> Programming affiliations with the PLplot scientific plotting software
> package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
> Loads of Linux Links project (loll.sf.net), and the Linux Brochure 
> Project
> (lbproject.sf.net).
> __________________________
>
> Linux-powered Science
> __________________________
>

Re: [Plplot-general] plplot-octave usage

[Alan W. I. <ir...@be...>]

On 2004-04-02 20:19+0900 Valery Pipin wrote:

> I found that the following piece of yplot code (at the end of message)
> fails on this version of plplot but it works on 5.3.0-2.[...]

> #######################
> plsdev,"xwin"
> plspage,400,400,400,400,1,1
> plscolbg,255,255,255
> plinit;
> pladv,0;
>
> plstick,3,1,1,1,2.,.5,2,1.5
> plwind,-1,0,-1.5,1.5
> pllsty,1
> plcol,7
> plcont,b,lbp,-x,y
....

Hi Valery:

I just rebuilt yplot on my own machine using the latest cvs version of
PLplot.  As you know, the yplot build includes building the documentation
and running many example scripts to produce all the standard postscript
plots.  No problems were encountered during this build so the problem you
have found is not triggered by the standard examples.

Thus, I tried to verify the error with your specific yplot script, but it
turns out that script is incomplete.  For example, b, lbp, x, and y are
undefined above (but there are other undefined quantities later in your
script as well).

I would like to figure out the source of the problem you have encountered so
could you send me privately a complete yplot script that works for
plplot-5.3.0, but not for the present plplot-5.3.0.cvs.20040324.tar.gz
snapshot at http://plplot.sourceforge.net/cvs-tarball/ ?

Thanks in advance.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

[Plplot-general] plplot-octave usage

[Valery P. <pi...@is...>]

I've tryed to use  the plplot  from octave.  I have  debian-sid.
I found that plplot can not be used from octave with usual octave 
(gnuplot) syntax.  The simple receipt given in "Usage.octave"
like,
plplot_toggle;
x=-pi:0.1:pi;
plot(x,sin(x));

does not work;

The only method is  to  immitate the  way which was used  in octave 
examples
x*.m copied from c dir. Is it correct?

Why am I asking  this?

Let me explain a bit. I usually use plplot from yplot interface.
Recently I've upgraded plplot from 5.3.0-2 to cvs version (as was proposed
by apt)

I found that the following piece of yplot code (at the end of message)  
fails on this version of plplot but it works on 5.3.0-2.   At first 
moment I  wanted to test it from octave but realised that it is not as 
simple.

best rgds,
Valery
#######################
plsdev,"xwin"
plspage,400,400,400,400,1,1
plscolbg,255,255,255
plinit;
pladv,0;

plstick,3,1,1,1,2.,.5,2,1.5
plwind,-1,0,-1.5,1.5
pllsty,1
plcol,7
plcont,b,lbp,-x,y
pllsty,2
plcol,13
plcont,b,lbn,-x,y
plcol,7
pllsty,1
plline,-x(1,1:tnum),y(1,1:tnum)
plline,-x(rnum,1:tnum),y(rnum,1:tnum)
plschr,2,2

plstick,3,1,2,1,7,1,4,1
plwind,0,1,-1,1
plwid,2
pllsty,1
plcol,1
plcont,o,lop,x,y
plcol,9
plcont,o,lon,x,y
pllsty,3
plwid,1
plcol,9
plcont,ob,levob,x,y
plcol,9
pllsty,1
plline,x(1,1:tnum),y(1,1:tnum)
plline,x(rnum,1:tnum),y(rnum,1:tnum)
plschr,2,2

//after  this line plplot  give
//ERROR (plcont) Segmentation violation interrupt (SIGSEGV)
//  LINE: 1188  FILE: /usr/share/yorick/1.5/i0/yplot.i
plstick,3,1,3,1,2,1,2,0
plwind,0,1.5,-1.5,1.5
pllsty,1
plcol,7
plcont,a, lap,x,y
pllsty,2
plcol,13
plcont,a, lan,x,y
plcol,7
pllsty,1
plline,x(1,1:tnum),y(1,1:tnum)
plline,x(rnum,1:tnum),y(rnum,1:tnum)
pllsty,1
plcol,7
plcont,aout, lap,x1e,y1e
pllsty,2
plcol,13
plcont,aout, lan,x1e,y1e

plend

Re: [Plplot-general] Static-linked, stand-alone executables?

[Alan W. I. <ir...@be...>]

On 2004-04-01 12:34-0800 ooOoo wrote:

> Is it possible to build statically-linked, stand-alone
> executable files with plplot?

Yes, although I don't much see the point since the executables end up
much larger than when you use the shared libraries.

However, assuming you have some exotic but justified reason to use static
libraries, (:-)) here is the cookbook:

Use 5.3.0 (since that is the only version we support).

./configure --disable-shared --disable-dyndrivers (+ other options such as --prefix)
make
make install

--disable-dyndrivers is important since the (default) dyndrivers option
does not work with static libraries.

Normally, both shared and static libraries are built, but we turn off the
shared ones (--disable-shared) to save you build time and also force the
executable build to use the static version of the libraries.  (Normally, the
linker tries to use the shared version of the library whenever possible.)

Build your executable just like the installed examples are built, e.g.,

gcc x01c.c -o x01c `plplot-config --cflags --libs`

Try plplot-config --help to see other options for that command.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________