plplot-general — PLplot user discussion group

Re: [Plplot-general] plplot and mac os x

[Koen v. d. D. <kvd...@ea...>]

On Dec 25, 2003, at 2:55 PM, Alan W. Irwin wrote:

> I am sorry everything is in a bit of disarray with Mac OS X and 
> PLplot, but
> we all three of Mac OS X, autotools, and PLplot are rapidly moving 
> targets.
> Also, bug fixing is slow because no PLplot developer has direct Mac OS 
> X
> access, and we must rely on indirect actions of users to help us debug 
> for
> that platform as a consequence.

I'll be glad to help - just don't have much time and resources 
(knowledge).


> In actual truth none of the above are needed.  For example, if all the 
> above
> are disabled, you could (theoretically, I haven't tried this myself) 
> still
> run any of the C examples from the command line using -dev psc (our 
> coloured
> postscript device) or -dev ps


I tried this, but didn;t get it to work, due to my experience with the 
command line (I don't have a unix but rather a Mac GUI background :).  
What is the correct syntax to run the c examples from the command line?


thanks,

- Koen.

Re: [Plplot-general] plplot and mac os x

[Alan W. I. <ai...@us...>]

On 2003-12-25 13:00-0500 Koen van der Drift wrote:

> On Dec 24, 2003, at 1:28 PM, Alan W. Irwin wrote:
>
> > Python:
> >
> > PY_VERSION=`python -c 'import sys ; print sys.version[0:3]'`
> > export PYTHON_INC_DIR=/usr/include/python${PY_VERSION}/
> >
> > The first command is just a trick to get the python version that is
> > normally
> > part of the include directory name, but if you know your python
> > version in
> > advance it won't be necessary.  The export command is the important
> > one.
> > Adjust that line (including the correct version number for your default
> > python) to wherever your python headers are located.
>
> Thanks for the input. From previous postings on Mac OSX and plplot I
> understand that I need to use --disable-shared.

That's only needed (as a temporary workaround) for the fortran case.  Also,
dynamic drivers don't work for all cases (we still haven't come up with a
fix for that bug) so you must always specify --disable-dyndrivers.  However,
if you specify --disable-f77 --disable-dyndrivers (and use the default
--enable-shared), shared libraries and python apparently (see discussions in
September on plplot-devel) work well.  Note, some patches had to be applied
to PLplot-5.2.1 to get that to work smoothly for old (10.1?) Mac OS X, but
the latest PLplot cvs does not work with Mac OS X 10.3 (still no fix for
that bug as well).

I am sorry everything is in a bit of disarray with Mac OS X and PLplot, but
we all three of Mac OS X, autotools, and PLplot are rapidly moving targets.
Also, bug fixing is slow because no PLplot developer has direct Mac OS X
access, and we must rely on indirect actions of users to help us debug for
that platform as a consequence.

> Besides python, X11 and tcltk, are there any other packages needed for
> plplot?

In actual truth none of the above are needed.  For example, if all the above
are disabled, you could (theoretically, I haven't tried this myself) still
run any of the C examples from the command line using -dev psc (our coloured
postscript device) or -dev ps (our gray-scale postscript device) and send
the file results directly to a printer without X, tcl/tk, or python being
installed on your system.

What is going on is the configure script has essentially every language
interface for PLplot and every device driver enabled by default. You can
turn off any of those with the --disable options, (run ./configure --help to
look at all options) and it will carry on with the subset of PLplot that you
have specified.  If configure cannot find needed system resources it will
disable even more options itself (with a warning message) and carry on.  So
if you want everything, look for those warning messages from configure and
install the appropriate development package. But you can choose to ignore
those warning messages, and the remaining subset of PLplot should work fine.

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 and mac os x

[Koen v. d. D. <kvd...@ea...>]

On Dec 24, 2003, at 1:28 PM, Alan W. Irwin wrote:

> Python:
>
> PY_VERSION=`python -c 'import sys ; print sys.version[0:3]'`
> export PYTHON_INC_DIR=/usr/include/python${PY_VERSION}/
>
> The first command is just a trick to get the python version that is 
> normally
> part of the include directory name, but if you know your python 
> version in
> advance it won't be necessary.  The export command is the important 
> one.
> Adjust that line (including the correct version number for your default
> python) to wherever your python headers are located.

Thanks for the input. From previous postings on Mac OSX and plplot I 
understand that I need to use --disable-shared. However, when shared is 
disabled, python will be disabled by the configure script. I could 
compile plplot with shared enabled, but I don't know yet whether the 
package is usable. I'll do some more testing in the next couple of days 
if I can find some time during the holidays.

Besides python, X11 and tcltk, are there any other packages needed for 
plplot?


thanks,

- Koen.

Re: [Plplot-general] plplot and mac os x

[Alan W. I. <ai...@us...>]

On 2003-12-24 07:00-0500 Koen van der Drift wrote:

> Hi,
>
> Thanks to some recent messages on this list, I was able to compile
> plplot (5.2.1) on Mac OS X 10.3.2. However, according to the table
> that's shown after the configure step, it doesn't compile for other
> languages than C++. I know that python and tcltk are standard on OS X.
> I checked the docs, but couldn't find how to configure the compilation
> (besides using --enable-python). What's the trick for doing this?
>

First make sure you have the development packages (those which include
the headers) installed for python and tcl/tk/itcl/itk.  Then
set environment variables before you execute ./configure to help that
script find the headers and libraries.

Python:

PY_VERSION=`python -c 'import sys ; print sys.version[0:3]'`
export PYTHON_INC_DIR=/usr/include/python${PY_VERSION}/

The first command is just a trick to get the python version that is normally
part of the include directory name, but if you know your python version in
advance it won't be necessary.  The export command is the important one.
Adjust that line (including the correct version number for your default
python) to wherever your python headers are located.

Tcl/Tk:

We have yet to find a decent autotools macro to help configure Tcl/Tk.  The
ones that come with our versions of Tcl/Tk seem way out of date (probably
because autotools have been such a moving target over the last year or so).
Thus, a lot more environment variables are currently required to help
./configure find tcl/tk stuff.  You may have to play a bit with the
following environment variables (see sysloc.in for how they are used) until
you get the right combination:

TCLINCDIR
TCLLIBDIR
TKINCDIR
TKLIBDIR

Once your configure output shows that both tcl and tk have been found you can
also control finding the include and library directories for itcl and itk
using

ITCLINCDIR
ITCLLIBDIR
ITKLIBDIR

(I have forgotten why we don't use ITKINCDIR, but apparently it is not
needed.) Also note you can do a lot in Tcl/Tk using PLplot without itcl and
itk so it is no disaster if those parts of the tcl/tk interface to PLplot
are turned off.

Let us 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
__________________________

[Plplot-general] plplot and mac os x

[Koen v. d. D. <kvd...@ea...>]

Hi,

Thanks to some recent messages on this list, I was able to compile 
plplot (5.2.1) on Mac OS X 10.3.2. However, according to the table 
that's shown after the configure step, it doesn't compile for other 
languages than C++. I know that python and tcltk are standard on OS X. 
I checked the docs, but couldn't find how to configure the compilation 
(besides using --enable-python). What's the trick for doing this?

thanks,


- Koen.

[Plplot-general] Re: I got a working plpot on Mac OS X 10.3

[Alan W. I. <ai...@us...>]

On 2003-12-21 18:54+0100 Michel Peyrard wrote:

>
>  Dear Alan,
>
> I finally got a working version of plplot on Mac OS X 10.3 with the
> tarfile 5.2.1, using --disable-shared and --disable-dyndrivers. Without
> any one of these options the compilation fails.
>
> The problem I had with undefined "restFP" and "saveFP" when compiling
> plplot fortran programs has been solved by a web search. The programs have
> to be linked with /usr/lib/libgcc.a and /usr/lib/libcc_dynamic.a
> (i.e. I must add -lgcc and -lcc_dynamic to my g77 compilation command)
>
> This solution is certainly not the best but it works on Mac OS X 10.3,
> with fortran, and this is really useful for me.

I am glad you have found a workaround for now.

Nevertheless, that leaves at least 3 PLplot problems to deal with for Mac OS
X 10.3 which I hope we will be able to sort out before our next release.

* The ./configure error you have with the latest cvs snapshots.

* The above linking problem for fortran.  I believe I have the solution to
this as discussed on plplot_devel.

* The ./get-drv-info error (Rob Managan also ran into this for an older Mac
OS X version).  The solution of this problem should allow you to drop the
--disable-dyndrivers limitation for Mac OS X.

I don't have the expertise to help with the first and 3rd problems, but I
hope others will be able to step in with some help.

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] I got a working plpot on Mac OS X 10.3

[Michel P. <Mic...@en...>]

 Dear Alan,

I finally got a working version of plplot on Mac OS X 10.3 with the
tarfile 5.2.1, using --disable-shared and --disable-dyndrivers. Without
any one of these options the compilation fails.

The problem I had with undefined "restFP" and "saveFP" when compiling
plplot fortran programs has been solved by a web search. The programs have
to be linked with /usr/lib/libgcc.a and /usr/lib/libcc_dynamic.a
(i.e. I must add -lgcc and -lcc_dynamic to my g77 compilation command)

This solution is certainly not the best but it works on Mac OS X 10.3,
with fortran, and this is really useful for me.

 Michel


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Michel Peyrard, Professeur       % Ph: +33 (0)4 7272 8374            %
% Laboratoire de Physique          % Fax: +33 (0)4 7272 8080           %
% Ecole Normale Sup=E9rieure de Lyon % e-mail:Mic...@en... %
% 46 all=E9e d'Italie                % perso.ens-lyon.fr/michel.peyrard  %
% 69364 Lyon Cedex 07, France      % (GPG key available on home page)  %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

[Plplot-general] Re: --disable-dyndrivers won't configure (fwd)

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

[Cc:ing to plplot-general]

* Alan W. Irwin <ir...@be...> [2003-12-20 08:30]:

> Rob Managan sailed through all the configuration part for Mac OS X without
> troubles in September.  Have we changed something relevant in our
> configuration since?

No I do not remember doing anything that would break building on MacOS X.
However, it may be appropriate to ask Rob Managan if he can build the
current cvs release of PLplot.

> I would appreciate you taking over here with Michel.  I don't even know the
> relevant questions to ask.
> 
> Alan
> ---------- Forwarded message ----------
> Date: Sat, 20 Dec 2003 10:29:11 +0100 (CET)
> From: Michel Peyrard <Mic...@en...>
> To: Alan W. Irwin <ai...@us...>
> Subject: --disable-dyndrivers won't configure
> 
> 
>  Dear Alan,
> 
> I tried to use ./configure --disable-dyndrives, but the command fails at
> the configuration level (see the details below).
> 
> Since it seemd to come from the cpp preprocessor which is not in /lib/cpp
> but in /usr/bin/cpp on Mac OS X, I tried to define the environment
> variable CPP='/usr/bin/cpp'
> With this variable the ./configure --disable-dyndrivers copmmand goes much
> further, but finally fails with another error.
> 
> In you preceding message you said that no developper has access to Mac OS
> X. I do not have expertise and time to be a developper, but, if you think
> that it is useful, I can be a "tester" of the new releases.
>
> [...]
> 
> With --disbale-dyndrivers, and the environment variable CPP='/usr/bin/cpp'
> -------------------------------------------------------------------------
> ... many successful steps ...
> 
> checking for vfork.h... no
> checking for fork... yes
> checking for vfork... yes
> checking for working fork... yes
> checking for working vfork... (cached) yes
> checking for popen... yes
> checking for usleep... yes
> checking for isinf... yes
> checking for finite... yes
> checking for isnan... yes
> checking for dynamic drivers...
> checking for perl... found
> configure: error: conditional "AMDEP" was never defined.
> Usually this means the macro was only invoked conditionally.

This is a very strange bug.  It looks like the /bin/sh program in your
system is failing with configure.  Could you please run the following line
at your shell and send the result to us (it should be either "t" or "f").

    echo `a=1;b=;if test -z "$a"&& test -z "$b";then echo t;else echo f; fi'|/bin/sh

Also, looking at the resulting config.log may also help.  Could you please
send it here?

-- 
Rafael

Re: [Plplot-general] plplot on macosx

[Alan W. I. <ai...@us...>]

On 2003-12-19 08:37+0100 Michel Peyrard wrote:

>
> Plplot on Mac OS X 10.3
> -----------------------
>
>  Unfortunately it does not seem to be as simple as that. I downloaded
> the CVS tar file and tried to install plplot on macosx 10.3 without
> success up to now. Attempts to disable dynamic drivers did not help.
>
>  Since I had problems with fortran, I tried ./configure --disable-f77
>
> The make still fails and I have not yet found a way to solve that.

I have now reviewed some of the messages on plplot_devel for late September
especially with the subject line "Fairly good success with Mac OS X"

It appears you have run into a known problem with dynamic drivers for Mac OS
X below.  In September when we ran into this "./get-drv-info" issue before,
full diagnostic output was generated for the developer who is knowledgable
about ./get-drv-info, but I don't think he has had time to look into it yet.
It's a difficult situation since no core developer has access to Mac OS X,
and we are looking for some knowledgable person to step forward to dig
through our code and autotools configuration to help us sort out any
problems for that platform.

For now, I advise using

./configure --disable-f77 --disable-dyndrivers

or (if you need fortran)

./configure --disable-shared --disable-dyndrivers

(From Rob Managan's ultimately successful tests in September, I recall there
was some problem with the default shared libraries and fortran common blocks
that could only be solved if you forced the use of static libraries as
above.)

I also strongly advise starting fresh every time (mv the old build directory
out of the way and start with a fresh unpack of the tarball) and also make
sure your prior installed results are completely removed.  The easiest way
to do the latter is to use a prefix such as --prefix=/usr/local/plplot.  The
old install can then be completely removed using rm -rf /usr/local/plplot.

Alan

> On Thu, 18 Dec 2003, Alan W. Irwin wrote:
>
> > On 2003-12-19 00:38+0100 Jens Frederich wrote:
> >
> > > hi,
> > >
> > > I will run plplot on macosx 10.3.
> > > When I try to compile plplot I become many error messages.
> > > Can anyone help me?
> > >
> > > /Jens
> >
> > This is a known libtool-1.4.x related problem with 5.2.1.  The solution was
> > for us to build our tarball with libtool-1.5.  At some point we will bring
> > out our next official release based on libtool-1.5, but meanwhile try a test
> > version of PLplot instead that has been built with libtool-1.5 and which is
> > located at http://people.debian.org/~rafael/plplot.html.  Others have had
> > good luck with our test versions and Mac OS X (and Cygwin as well which also
> > doesn't work for 5.2.1 because of libtool-1.4.x problems.).
> >
> > Alan
> > __________________________
> > Alan W. Irwin
> > email: ir...@be...
>
> -----------------------------------------------------------------
> Problem compiling PLPLOT on Mac OS X 10.3
> ------------------------------------------
> .... successful compilations ... until something goes wrong!
>
> /bin/sh ../libtool --mode=link gcc  -g -O2   -o xwin.la  -rpath /usr/local/lib/plplot5.2.1.cvs.20031030/data/../driversd -no-undefined -module -avoid-version  ../src/libplplotd.la -L/usr/X11R6/lib -lX11  -lm xwin_la-xwin.lo
> gcc  -o .libs/xwin.so -bundle  .libs/xwin_la-xwin.o  /usr/local/src/plplot-5.2.1.cvs.C/lib/csa/.libs/libcsirocsa.dylib ../src/.libs/libplplotd.dylib -L/usr/X11R6/lib -lX11 -lm
> ld: warning -prebind has no effect with -bundle
> ar cru .libs/xwin.a  xwin_la-xwin.o
> ranlib .libs/xwin.a
> creating xwin.la
> (cd .libs && rm -f xwin.la && ln -s ../xwin.la xwin.la)
> if gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../include -I../include   -I../libltdl  -I../libltdl -g -O2 -MT get_drv_info-get-drv-info.o -MD -MP -MF ".deps/get_drv_info-get-drv-info.Tpo" \
>   -c -o get_drv_info-get-drv-info.o `test -f 'get-drv-info.c' || echo './'`get-drv-info.c; \
> then mv -f ".deps/get_drv_info-get-drv-info.Tpo" ".deps/get_drv_info-get-drv-info.Po"; \
> else rm -f ".deps/get_drv_info-get-drv-info.Tpo"; exit 1; \
> fi
> /bin/sh ../libtool --mode=link gcc  -g -O2   -o get-drv-info  get_drv_info-get-drv-info.o ../libltdl/libltdlc.la
> gcc -g -O2 -o get-drv-info get_drv_info-get-drv-info.o  ../libltdl/.libs/libltdlc.a -ldl
> ./get-drv-info `echo dg300.la | sed 's/.la//'` > dg300.rc
> ltdl.c:2460: failed assertion `dirname'
> make[2]: *** [dg300.rc] Error 134
> make[1]: *** [all-recursive] Error 1
> make: *** [all] Error 2
>

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 on macosx

[Michel P. <Mic...@en...>]

Plplot on Mac OS X 10.3
-----------------------

 Unfortunately it does not seem to be as simple as that. I downloaded
the CVS tar file and tried to install plplot on macosx 10.3 without
success up to now. Attempts to disable dynamic drivers did not help.

 Since I had problems with fortran, I tried ./configure --disable-f77

The make still fails and I have not yet found a way to solve that.

I include below the result of the ./configure command and the error
message that I got when I tried to run make, in case someone can help me.


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Michel Peyrard, Professeur       % Ph: +33 (0)4 7272 8374            %
% Laboratoire de Physique          % Fax: +33 (0)4 7272 8080           %
% Ecole Normale Sup=E9rieure de Lyon % e-mail:Mic...@en... %
% 46 all=E9e d'Italie                % perso.ens-lyon.fr/michel.peyrard  %
% 69364 Lyon Cedex 07, France      % (GPG key available on home page)  %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

On Thu, 18 Dec 2003, Alan W. Irwin wrote:

> On 2003-12-19 00:38+0100 Jens Frederich wrote:
>
> > hi,
> >
> > I will run plplot on macosx 10.3.
> > When I try to compile plplot I become many error messages.
> > Can anyone help me?
> >
> > /Jens
>
> This is a known libtool-1.4.x related problem with 5.2.1.  The solution w=
as
> for us to build our tarball with libtool-1.5.  At some point we will brin=
g
> out our next official release based on libtool-1.5, but meanwhile try a t=
est
> version of PLplot instead that has been built with libtool-1.5 and which =
is
> located at http://people.debian.org/~rafael/plplot.html.  Others have had
> good luck with our test versions and Mac OS X (and Cygwin as well which a=
lso
> doesn't work for 5.2.1 because of libtool-1.4.x problems.).
>
> Alan
> __________________________
> Alan W. Irwin
> email: ir...@be...

-----------------------------------------------------------------
Problem compiling PLPLOT on Mac OS X 10.3
------------------------------------------
command:        ./configure --disable-f77
system:         powerpc-apple-darwin7.2.0
have_x:         yes
prefix:         /usr/local
CC:             gcc
CXX:            g++
LIB_TAG:        d
devices:         dg300 hp7470 hp7580 lj_hpgl imp ljii ljiip mem null pbm pl=
meta ps psc pstex xterm tek4010 tek4107 mskermit versaterm vlt conex tek401=
0f tek4107f xfig xwin

Available device drivers:
static:
dynamic:         dg300.la hpgl.la impress.la ljii.la ljiip.la mem.la null.l=
a pbm.la plmeta.la ps.la pstex.la tek.la xfig.la xwin.la

Compilation options:
with_debug:     no              with_opt:       yes
with_warn:      no              with_profile:   no

Library options:
enable_shared:  yes             enable_static:  yes
with_rpath:     yes             with_double:    yes

Optional libraries:
with_qhull:     no              with_csa:       yes
with_freetype:  no              with_pthreads:  no


Language Bindings:
enable_tcl:     no              enable_itcl:    no
enable_cxx:     yes             enable_f77:     no
enable_java:    no              enable_python:  no
enable_octave:  no

klee:/usr/local/src/plplot-5.2.1.cvs.C root# make
make  all-recursive
Making all in libltdl
make  all-am

=2E... successful compilations ... until something goes wrong!

/bin/sh ../libtool --mode=3Dlink gcc  -g -O2   -o xwin.la  -rpath /usr/loca=
l/lib/plplot5.2.1.cvs.20031030/data/../driversd -no-undefined -module -avoi=
d-version  ../src/libplplotd.la -L/usr/X11R6/lib -lX11  -lm xwin_la-xwin.lo
gcc  -o .libs/xwin.so -bundle  .libs/xwin_la-xwin.o  /usr/local/src/plplot-=
5.2.1.cvs.C/lib/csa/.libs/libcsirocsa.dylib ../src/.libs/libplplotd.dylib -=
L/usr/X11R6/lib -lX11 -lm
ld: warning -prebind has no effect with -bundle
ar cru .libs/xwin.a  xwin_la-xwin.o
ranlib .libs/xwin.a
creating xwin.la
(cd .libs && rm -f xwin.la && ln -s ../xwin.la xwin.la)
if gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../include -I../include   -I../libltd=
l  -I../libltdl -g -O2 -MT get_drv_info-get-drv-info.o -MD -MP -MF ".deps/g=
et_drv_info-get-drv-info.Tpo" \
  -c -o get_drv_info-get-drv-info.o `test -f 'get-drv-info.c' || echo './'`=
get-drv-info.c; \
then mv -f ".deps/get_drv_info-get-drv-info.Tpo" ".deps/get_drv_info-get-dr=
v-info.Po"; \
else rm -f ".deps/get_drv_info-get-drv-info.Tpo"; exit 1; \
fi
/bin/sh ../libtool --mode=3Dlink gcc  -g -O2   -o get-drv-info  get_drv_inf=
o-get-drv-info.o ../libltdl/libltdlc.la
gcc -g -O2 -o get-drv-info get_drv_info-get-drv-info.o  ../libltdl/.libs/li=
bltdlc.a -ldl
=2E/get-drv-info `echo dg300.la | sed 's/.la//'` > dg300.rc
ltdl.c:2460: failed assertion `dirname'
make[2]: *** [dg300.rc] Error 134
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

Re: [Plplot-general] plplot on macosx

[Alan W. I. <ai...@us...>]

On 2003-12-19 00:38+0100 Jens Frederich wrote:

> hi,
>
> I will run plplot on macosx 10.3.
> When I try to compile plplot I become many error messages.
> Can anyone help me?
>
> /Jens

This is a known libtool-1.4.x related problem with 5.2.1.  The solution was
for us to build our tarball with libtool-1.5.  At some point we will bring
out our next official release based on libtool-1.5, but meanwhile try a test
version of PLplot instead that has been built with libtool-1.5 and which is
located at http://people.debian.org/~rafael/plplot.html.  Others have had
good luck with our test versions and Mac OS X (and Cygwin as well which also
doesn't work for 5.2.1 because of libtool-1.4.x problems.).

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 on macosx

[Jens F. <jfr...@ma...>]

hi,

I will run plplot on macosx 10.3.
When I try to compile plplot I become many error messages.
Can anyone help me?

/Jens

--
advTHANKSance, Jens Frederich
je...@fr...

Re: [Plplot-general] Problem compiling plplot-5.2.1

[Alan W. I. <ai...@us...>]

On 2003-12-04 14:25-0500 James B. Hiller wrote:

> The last problem that I'm having (though it doesn't prohibit the build)
> is that configure seems to not be finding my python (2.3.2) installation.

You can override everything python related.

For example, I believe the following should work for you:

PY_VERSION=`python -c 'import sys ; print sys.version[0:3]'`
export PYTHON_INC_DIR=/usr/local/include/python${PY_VERSION}/

./configure ....

> I found the sysloc.in file, and the list of directories to look
> in, and added python2.3 to that list.  Also no help.

hmmm.  That should work, but that method requires that you have all the
correct autotools versions installed.

If the export trick above does not work, do you have the correct Numeric for
your python version installed as well?  (See below for where to get it.)

> Is this a case where python-2.3.2 is simply too new for the current
> version of plplot to deal with?

My experience is we have had to make virtually no changes from
python-1.5/Numeric-15 on since both the python and numeric teams are so
careful to keep things backwards compatible.  So I doubt python-2.3 or
Numeric-2 (which you can find as Numeric-23.1.tar.gz at
http://sourceforge.net/project/showfiles.php?group_id=1369) are going to be
a problem.  I haven't tried python-2.3 myself, but IIRC, plplot builds fine
on Debian unstable with python2.3.

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 compiling plplot-5.2.1

[James B. H. <jh...@vi...>]

Hi!  Thx for answering so fast!

> | undefined reference to `errno' collect2: ld returned 1 exit status
> 
> Hi, 
> This is a known problem in some architectures.
> You must  add
> 
>     #include <errno.h>
> 
> to lib/csa/csa.c

Much appreciated!

The last problem that I'm having (though it doesn't prohibit the build)
is that configure seems to not be finding my python (2.3.2) installation.

I have python headers installed in /usr/local/lib/python2.3, and the
library files in /usr/local/lib/python2.3.  I rummaged through configure,
config.log, and config.status and saw that it wasn't finding Python.h,
which IS in /usr/local/include/python2.3.  So I included both a
CFLAGS and CPPFLAGS value, and LDFLAGS value to configure; no help.
And I found the sysloc.in file, and the list of directories to look
in, and added python2.3 to that list.  Also no help.

Is this a case where python-2.3.2 is simply too new for the current
version of plplot to deal with?

thx,
jbh

Re: [Plplot-general] Problem compiling plplot-5.2.1

[<jc...@fe...>]

On Thursday 04 December 2003 17:54, James B. Hiller wrote:
| Hi.
|
| Am trying to compile subject version, and am getting stuck with:
|
| gcc -g -O2 -mieee-fp -o .libs/plserver plserver.o 
| ../tcl/.libs/libplplottcltkd.so -L/usr/local/lib -litcl3.2 -litk3.2
| -ltk8.4
| /usr/local/packages/incoming/plplot-5.2.1/src/.libs/libplplotd.so
| /usr/local/lib/libfreetype.so -lz
| /usr/local/packages/incoming/plplot-5.2.1/lib/csa/.libs/libcsa.so -lm
| -ldl
| /usr/local/packages/incoming/plplot-5.2.1/bindings/tcl/.libs/libtclma
|trixd.so -ltcl8.4 -Wl,--rpath -Wl,/usr/local/lib
| /usr/local/packages/incoming/plplot-5.2.1/lib/csa/.libs/libcsa.so:
| undefined reference to `errno' collect2: ld returned 1 exit status

Hi, 
This is a known problem in some architectures.
You must  add

    #include <errno.h>

to lib/csa/csa.c

Thanks
Joao

| make[3]: *** [plserver] Error 1
| make[3]: Leaving directory
| `/usr/local/packages/incoming/plplot-5.2.1/bindings/tk' make[2]: ***
| [all-recursive] Error 1
| make[2]: Leaving directory
| `/usr/local/packages/incoming/plplot-5.2.1/bindings' make[1]: ***
| [all-recursive] Error 1
| make[1]: Leaving directory
| `/usr/local/packages/incoming/plplot-5.2.1' make: *** [all] Error 2
|
| From where it was dying, I thought maybe something was wrong with my
| tcl or tk install, so I did explicit --disable-tcl and --disable-tk
| in the configuration, and the compilation got past this area, but
| then died with the same error during compilation of the csa library.
|
| The last thing I tried was simply a plain ./configure, to see if
| anything else would get rid of it, and that died as well.
|
| Am using linux-2.4.22, gcc 3.2.3, and glibc-2.3.2.
|
| Any insight anyone could provide would be most appreciated.
|
| thx,
| jbh
|
|
| -------------------------------------------------------
| This SF.net email is sponsored by: SF.net Giveback Program.
| Does SourceForge.net help you be more productive?  Does it
| help you create better code?  SHARE THE LOVE, and help us help
| YOU!  Click Here: http://sourceforge.net/donate/
| _______________________________________________
| Plplot-general mailing list
| Plp...@li...
| https://lists.sourceforge.net/lists/listinfo/plplot-general

[Plplot-general] Problem compiling plplot-5.2.1

[James B. H. <jh...@vi...>]

Hi.

Am trying to compile subject version, and am getting stuck with:

gcc -g -O2 -mieee-fp -o .libs/plserver plserver.o  ../tcl/.libs/libplplottcltkd.so -L/usr/local/lib -litcl3.2 -litk3.2 -ltk8.4 /usr/local/packages/incoming/plplot-5.2.1/src/.libs/libplplotd.so /usr/local/lib/libfreetype.so -lz /usr/local/packages/incoming/plplot-5.2.1/lib/csa/.libs/libcsa.so -lm -ldl /usr/local/packages/incoming/plplot-5.2.1/bindings/tcl/.libs/libtclmatrixd.so -ltcl8.4 -Wl,--rpath -Wl,/usr/local/lib
/usr/local/packages/incoming/plplot-5.2.1/lib/csa/.libs/libcsa.so: undefined reference to `errno'
collect2: ld returned 1 exit status
make[3]: *** [plserver] Error 1
make[3]: Leaving directory `/usr/local/packages/incoming/plplot-5.2.1/bindings/tk'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory `/usr/local/packages/incoming/plplot-5.2.1/bindings'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/local/packages/incoming/plplot-5.2.1'
make: *** [all] Error 2

From where it was dying, I thought maybe something was wrong with my tcl or
tk install, so I did explicit --disable-tcl and --disable-tk in the
configuration, and the compilation got past this area, but then died with
the same error during compilation of the csa library.

The last thing I tried was simply a plain ./configure, to see if anything
else would get rid of it, and that died as well.

Am using linux-2.4.22, gcc 3.2.3, and glibc-2.3.2.

Any insight anyone could provide would be most appreciated.

thx,
jbh

Re: [Plplot-general] plplot on MacOS X

[Alan W. I. <ai...@us...>]

On 2003-11-20 10:45+0100 Michel Peyrard wrote:

>
> I tried to compile the plplot-5.2.1 on a Mac OS X (V. 1.3) computer and
> failed. The compilation goes through but I have a problem with ld (see
> message below).
>
> I need a fortran_enabled version.
>
> Has any one succeeded?  Is there by chance a binary package available (fink
> does not have it I think).

Here is the short story:

* The autotools versions we used to generate the 5.2.1 tarball do not
support Mac OS X.  Instead, use the experimental tarball available (built
with the latest autotools) at http://people.debian.org/~rafael/plplot.html.

* Rob Managan found the experimental tarball mostly worked on Mac OS X
(see his posts and my replies on plplot-general and plplot-devel), but
IIRC there were problems with fortran and using shared libraries due to
a macosx limitation on common blocks in shared libraries.  So I would suggest
trying the static library approach (the --disable-shared ./configure option)
first.  Remember, ./configure --help is your friend.

Good luck, and let us 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
__________________________

[Plplot-general] plplot on MacOS X

[Michel P. <Mic...@en...>]

I tried to compile the plplot-5.2.1 on a Mac OS X (V. 1.3) computer and
failed. The compilation goes through but I have a problem with ld (see
message below).

I need a fortran_enabled version.

Has any one succeeded? Is there by chance a binary package available (fink
does not have it I think).

Thanks.

 Michel Peyrard

ld: warning prebinding disabled because dependent library: /usr/local/plplo=
t-5.2.1/lib/libplplotd.5.dylib is not prebound
ld: warning can't open dynamic library: /usr/local/plplot-5.2.1/lib/libcsa.=
0.dylib (checking for undefined symbols may be affected) (No such file or d=
irectory, errno =3D 2)
ld: common symbols not allowed with MH_DYLIB output format with the -multi_=
module option
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _plplot_ (s=
ize 48)
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _zzplstr1_ =
(size 320)
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _zzplstr2_ =
(size/usr/bin/libtool: internal link edit command failed
320)
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _zzplstr3_ =
(size 320)
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _zzplstr4_ =
(size 320)
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _zzplstr5_ =
(size 320)
=2Elibs/libplplotf77d.5.3.0.dylib-master.o definition of common _zzplstr6_ =
(size 320)
make[3]: *** [libplplotf77d.la] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Michel Peyrard, Professeur       % Ph: +33 (0)4 7272 8374            %
% Laboratoire de Physique          % Fax: +33 (0)4 7272 8080           %
% Ecole Normale Sup=E9rieure de Lyon % e-mail:Mic...@en... %
% 46 all=E9e d'Italie                % perso.ens-lyon.fr/michel.peyrard  %
% 69364 Lyon Cedex 07, France      % (GPG key available on home page)  %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Re: [Plplot-general] -g option for compilers

[Rafael L. <ra...@de...>]

* Joachim Geiger <joa...@ip...> [2003-11-17 10:38]:

> Thanks alot, this worked. I followed Alan's suggestion giving make the 
> appropriate compilation flags. Additionally, I would vote for the suggested
> TODO entry. 

Investigating further this issue (thanks, Alan, for bugging me!), I
discovered that Automake already includes this option.  In order to install
stripped versions of the libraries, all you have to do is:

    make install-strip

This should work in a cross-platform way.    
 
-- 
Rafael

Re: [Plplot-general] -g option for compilers

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

Rafael Laboissiere wrote:

>* Alan W. Irwin <ai...@us...> [2003-11-14 12:19]:
>
>  
>
>>On 2003-11-14 19:05+0100 Joachim Geiger wrote:
>>
>>    
>>
>>>... I was not able to get rid of this stupid -g flag.
>>>      
>>>
>>Here is what is going on according to the documentation of the AC_PROG_CC
>>you get from 'info autoconf'.  By default it chooses the -g flag if your
>>compiler will accept that flag (which is apparently the case for you.) But
>>if you want to override that default you must set the environment variable
>>CFLAGS to whatever option you like that is appropriate for your compiler,
>>i.e., use
>>
>>make 'CFLAGS = '
>>
>>or
>>
>>make 'CFLAGS = -O2'
>>
>>I notice in the default Makefile generated on my system that CXXFLAGS and
>>FFLAGS also contain the -g option so you will have to set those variables to
>>whatever compile options are appropriate for your c++ and fortran compilers
>>as well.
>>
>>Rafael has suggested stripping your libraries, and my guess is that would
>>work as well.
>>    
>>
>
>It would be nice to provide a configure option for forcing the stripping of
>the libraries.  This could be implemented in an install-hook target in
>src/Makefile.am.  Should I write a TODO entry for this?
>
>  
>
Thanks alot, this worked. I followed Alan's suggestion giving make the 
appropriate
compilation flags. Additionally, I would vote for the suggested TODO entry.
Best regards,
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] -g option for compilers

[Rafael L. <lab...@ps...>]

* Alan W. Irwin <ai...@us...> [2003-11-14 12:19]:

> On 2003-11-14 19:05+0100 Joachim Geiger wrote:
> 
> > ... I was not able to get rid of this stupid -g flag.
> 
> Here is what is going on according to the documentation of the AC_PROG_CC
> you get from 'info autoconf'.  By default it chooses the -g flag if your
> compiler will accept that flag (which is apparently the case for you.) But
> if you want to override that default you must set the environment variable
> CFLAGS to whatever option you like that is appropriate for your compiler,
> i.e., use
> 
> make 'CFLAGS = '
> 
> or
> 
> make 'CFLAGS = -O2'
> 
> I notice in the default Makefile generated on my system that CXXFLAGS and
> FFLAGS also contain the -g option so you will have to set those variables to
> whatever compile options are appropriate for your c++ and fortran compilers
> as well.
> 
> Rafael has suggested stripping your libraries, and my guess is that would
> work as well.

It would be nice to provide a configure option for forcing the stripping of
the libraries.  This could be implemented in an install-hook target in
src/Makefile.am.  Should I write a TODO entry for this?

-- 
Rafael

Re: [Plplot-general] -g option for compilers

[Alan W. I. <ai...@us...>]

On 2003-11-14 19:42+0100 Rafael Laboissiere wrote:

> * Rafael Laboissiere <ra...@de...> [2003-11-14 19:30]:
>
> > Maybe you could try to do the same using the strip command (or better:
> > change from RH to Debian :-)
>
> I forgot to give you the reference:
>
>     http://lwn.net/Articles/56151/

That's a good article and the comments afterward are well worth reading
as well.  Thanks for bringing that URL to our attention.

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] -g option for compilers

[Alan W. I. <ai...@us...>]

On 2003-11-14 19:05+0100 Joachim Geiger wrote:

> ... I was not able to get rid of this stupid -g flag.

Here is what is going on according to the documentation of the AC_PROG_CC
you get from 'info autoconf'.  By default it chooses the -g flag if your
compiler will accept that flag (which is apparently the case for you.) But
if you want to override that default you must set the environment variable
CFLAGS to whatever option you like that is appropriate for your compiler,
i.e., use

make 'CFLAGS = '

or

make 'CFLAGS = -O2'

I notice in the default Makefile generated on my system that CXXFLAGS and
FFLAGS also contain the -g option so you will have to set those variables to
whatever compile options are appropriate for your c++ and fortran compilers
as well.

Rafael has suggested stripping your libraries, and my guess is that would
work as well.

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] -g option for compilers

[Rafael L. <ra...@de...>]

* Rafael Laboissiere <ra...@de...> [2003-11-14 19:30]:

> Maybe you could try to do the same using the strip command (or better:
> change from RH to Debian :-)

I forgot to give you the reference:

    http://lwn.net/Articles/56151/
    
-- 
Rafael

Re: [Plplot-general] -g option for compilers

[Rafael L. <ra...@de...>]

* Joachim Geiger <joa...@ip...> [2003-11-14 19:05]:

> I tried to compile and install the 5.2.1-version of plplot
> on my RedHat8.0 machine enabling f77 using the Lahey-
> Fujitsu-compiler f95f. Compilation and intallation went fine
> but when linking to my other codes I had a hard time upto
> the point when I noticed that the whole plplot was compiled
> with the -g flag (I think this is debugging) and that I also
> had to compile my other code with -g. I tried to configure
> plplot with --disable-debugging but I was not able to get
> rid of this stupid -g flag. I don't want to run all of my codes
> in debug-mode. Any ideas what to do?

I do not know if what I am going to say matters in this case, but the Debian
packages install stripped versions of the library files:

$ for f in  src/.libs/libplplotd.so.9.0.0 /usr/lib/libplplotd.so.9.0.0 ; do file $f ; ls -l $f ; done
src/.libs/libplplotd.so.9.0.0: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), not stripped
-rwxr-xr-x    1 rafael   rafael     776087 2003-10-30 09:55 src/.libs/libplplotd.so.9.0.0
/usr/lib/libplplotd.so.9.0.0: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), stripped
-rw-r--r--    1 root     root       222224 2003-10-29 22:25 /usr/lib/libplplotd.so.9.0.0

Maybe you could try to do the same using the strip command (or better:
change from RH to Debian :-)

-- 
Rafael