plplot-general — PLplot user discussion group

Re: [Plplot-general] Am I going about this the best way?

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

On 2015-03-31 09:25-0700 Alan W. Irwin wrote:

> Hi Aaron.
>
> There has been a very recent development.  One of our core developers
> has just run into the same bug of being unable to build plplotqt for
> the static library case because the compile flag -DUSINGDLL has been
> incorrectly used for that case.  I have fixed it on my local system,
> but SourceForge has a complete outage at this time (ugh) so I cannot
> propagate that fix (yet) to a place where you can use it.
>
> Once SF is back up, if you haven't done so already, I would sign up
> for our git feed so you will get e-mail notice of each PLplot git
> commit to the master branch (which, for example, would let you know
> when my fix is pushed to master).  In order to do that signup login to
> Sourceforge (once SF is back up) go to our SF project page -> Code,
> and sign up for the git feed (using one of the menu items on that Code
> page).

Hi Aaron:

SF now working again, and I have pushed this bug fix (commit id
b6b60d1).  Please test that this solves the issue.

Alan
__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Am I going about this the best way?

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

Hi Aaron.

There has been a very recent development.  One of our core developers
has just run into the same bug of being unable to build plplotqt for
the static library case because the compile flag -DUSINGDLL has been
incorrectly used for that case.  I have fixed it on my local system,
but SourceForge has a complete outage at this time (ugh) so I cannot
propagate that fix (yet) to a place where you can use it.

Once SF is back up, if you haven't done so already, I would sign up
for our git feed so you will get e-mail notice of each PLplot git
commit to the master branch (which, for example, would let you know
when my fix is pushed to master).  In order to do that signup login to
Sourceforge (once SF is back up) go to our SF project page -> Code,
and sign up for the git feed (using one of the menu items on that Code
page).

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Am I going about this the best way?

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

On 2015-03-31 07:45-0500 Aaron Hexamer wrote:

> Alan,
>
> The version I'm using is from git, but after doing a remote fetch, I can see
> that I'm a few commits behind.  I'm working off of 0bfe721.  Happy to
> provide as many details as you need.  Should I advance first and see if its
> fixed?

Yes, please.

> Is there a bug submission std I can check out to see how much detail
> you'd like?

Yes. Look at the last part of
<https://sourceforge.net/p/plplot/wiki/Building_PLplot/>.

Thanks in advance.

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] plplot - mem/memcairo repeated, > render-to-memory?

[Hazen B. <hba...@ma...>]

>
> Hi everyone.
> Just to follow up on this briefly, I have solved the problem to my satisfaction by a small modification to the ltdl_win32.c file (maps dynamic loader functions to their win32 equivalents). The driver DLLs are now simply kept in memory for the entire process life time by skipping the FreeLibrary call. It turned out that loading/freeing shared libs is the major cycle hog (everything else done in plinit() is negligible in comparison), even if the modules in question have previously been mapped into the proces. Now LoadLibrary() just returns the handle to the already mapped module.
>
> This isn't 100% clean of course, but it'll do nicely for me.
>
> Thanks to all for their suggestions.
> bests,
> Dietmar

You might try using plend1() instead of plend() when you finish the 
plot. This also keeps all the shared libs in memory so that plinit() 
does not have to reload them.

-Hazen

Re: [Plplot-general] Am I going about this the best way?

[Aaron H. <he...@co...>]

Alan,

The version I'm using is from git, but after doing a remote fetch, I can see
that I'm a few commits behind.  I'm working off of 0bfe721.  Happy to
provide as many details as you need.  Should I advance first and see if its
fixed?  Is there a bug submission std I can check out to see how much detail
you'd like?

Thanks,

Aaron.

-----Original Message-----
From: Alan W. Irwin [mailto:ir...@be...] 
Sent: Tuesday, March 31, 2015 1:22 AM
To: Aaron Hexamer
Cc: plp...@li...
Subject: RE: [Plplot-general] Am I going about this the best way?

Hi Aaron:

On 2015-03-30 22:29-0500 Aaron Hexamer wrote:

> Alan,
>
> Thanks for the hint on Qt4 vs Qt5 - probably spared me a few hours of
pain!
> I set out to trial the memqt driver and ran into some issues with the 
> build on MSVC 2012.
>
> 1) For some reason, which I had never encountered before, the CMake 
> system generates incorrect build files when the source path includes 
> spaces.  Not a big deal really, I can just move to a path without 
> spaces.  I don't recall the exact issue, but I think it broke up 
> include paths into separate entries.  I am using a fairly old version 
> of CMake (2.8.12.2)

That's an issue with our own build system and not CMake.  From time to time
we try and clean up some aspect of this issue, but it is a low priority
because it is so easy to use PATH's without spaces in them.

>
> I started with Qt configured for dynamic linking
>
> 2) I tried to build PLplot as a shared lib, then realized that was not 
> really going to work for me since my application accessed some other 
> normally private functions, not exported in the DLL (pldtik, pldprec,
etc).
>
> 3) I tried to build PLplot as a static lib (which I normally do for my 
> app anyway).  But then I started getting incompatible linkage warnings 
> and some errors about trying to initialize dllimport'ed data
> (\plplot-plplot\bindings\qt_gui\plqt.cpp(28): error C2491: 'vectorize' :
> definition of dllimport data not allowed).  After finding/following 
> the thread here (http://ehc.ac/p/plplot/mailman/message/33603310/), I 
> began to suspect that the USINGDLL macro might be an issue.  Despite 
> the fact that I configured PLplot for static build it seems that the 
> cmake scripts still applied USINGDLL to the plqt.c file. Is that 
> expected (I don't totally understand the use of USINGDLL in all the 
> contexts)?  Is it perhaps related to the issue described in that thread on
the devel list?

Are you using PLplot-5.10.0 or the git master tip version (which is about to
turn into PLplot-5.11.0)?  I would prefer you to try the master tip version
if you haven't been doing that already because it has been a very long time
since 5.10.0 was released, I am much more familiar with the master tip
version, and I would like to get the master tip version tested as well as
possible before the 5.11.0 release.  For example, if you spot a USINGDLL
issue for that version (such as USINGDLL being defined for the static
library case), please let me know with lots of details.

Alan
__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); the Time Ephemerides project
(timeephem.sf.net); PLplot scientific plotting software package
(plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux
Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] plplot - mem/memcairo repeated render-to-memory?

[Dietmar N. <po...@di...>]

Hi everyone.
Just to follow up on this briefly, I have solved the problem to my satisfaction by a small modification to the ltdl_win32.c file (maps dynamic loader functions to their win32 equivalents). The driver DLLs are now simply kept in memory for the entire process life time by skipping the FreeLibrary call. It turned out that loading/freeing shared libs is the major cycle hog (everything else done in plinit() is negligible in comparison), even if the modules in question have previously been mapped into the proces. Now LoadLibrary() just returns the handle to the already mapped module. 

This isn't 100% clean of course, but it'll do nicely for me.

Thanks to all for their suggestions.
bests,
Dietmar


Hazen Babcock schrieb am 27.03.2015 15:00:

>>
>> Hello all.
>>
>> After having tried quite a few different different options I feel I'm
>> running out of ideas how to achive my goal, so I decided to ask around.
>> Apologies if this is a trivial issue ...
>>
>> What I'm trying to do is get plplot to render a graph into a memory
>> buffer, which is then blitted to a window whenver i) that window changes
>> in size, ii) the window needs to be updated after being invalidated or
>> iii) the data underlying the graph changes.
>>
>> A small test application I prepared (platform: Windows X64) worked well
>> enough, I set the driver to "memcairo", passed a frame buffer memory
>> block via plsmem(), issued the plotting commands and then dumped the
>> buffer to inspect its contents. Right enough, it was rendered just the
>> way I need it, ready to be displayed.
>>
>> But my attempts to incorporate that approach into the real application
>> failed. I was planning to have one plstream per window, created and
>> initialized upon window creation. A new memory buffer is passed in by
>> plsmem() whenever the window size changes, to reflect the current
>> drawing surface. In the window paint routine, I'm trying to simply
>> render the plot, then blit the buffer.
>>
>> However that only works exactly once, all subsequent paint calls fail
>> with an access violation in a pl...() command.
>>
>> The rendering commands were stripped down to
>>
>> plenv0(...)
>> plline()
>>
>> But that doesn't seem to make any difference. Basically any call to
>> plenv() seems to change something in an unwanted way, causing subsequent
>> calls to write to invalid memory.
>>
>> Do I need to skip plenv(), and replace it with individual calls for
>> drawing a box and setting a viewport?
>>
>> I played with pladv(), but to no avail.
>>
>> calling plinit(...) for each render cycle to set up a new enviroment
>> seems to work, but that is so slow as to be unusable, so that's no option.
>>
>> Is there anything I'm doing wrong, or are there some examples I failed
>> to notice? Help is appreciated.
> 
> Unfortunately, as you have noticed, you can't provide change the plot 
> memory once you've specified it. You basically do need to start from 
> scratch each time as the memory is passed to the plot driver when the 
> driver is initialized. I'm surprised that this is too slow, as I've used 
> a similar approach (extcairo driver), and it seemed to work ok for me.
> 
> One option might be to pass in an area as large as you think you will 
> ever need, then just plot into and blit the sub-section of the area that 
> matches the current window size?
> 
> -Hazen
> 

Re: [Plplot-general] Build Plplot with ifort on OS X

[Arjen M. <Arj...@de...>]

Hi Alan, Hiro,



> -----Original Message-----
> From: Alan W. Irwin [mailto:ir...@be...]


>
> @Arjen:
>
> In my experience running pkg-config twice like that with the separate options should
> give exactly the same result as running it once with the two options.
>
If that is indeed the case, it is even more puzzling.

> @Hiro:
>
> We are desperate here to find a solution to this extremely puzzling issue so by all
> means try any experiment you can think of including this one suggested by Arjen or
> the one I suggested previously to cut and paste pkg-config results so that your hand-
> crafted build command doesn't use pkg-config directly.
>
If this does not work, I will post a question on the appropriate Intel forum about it.

Regards,

Arjen



DISCLAIMER: This message is intended exclusively for the addressee(s) and may contain confidential and privileged information. If you are not the intended recipient please notify the sender immediately and destroy this message. Unauthorized use, disclosure or copying of this message is strictly prohibited. The foundation 'Stichting Deltares', which has its seat at Delft, The Netherlands, Commercial Registration Number 41146461, is not liable in any way whatsoever for consequences and/or damages resulting from the improper, incomplete and untimely dispatch, receipt and/or content of this e-mail.

Re: [Plplot-general] Am I going about this the best way?

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

Hi Aaron:

On 2015-03-30 22:29-0500 Aaron Hexamer wrote:

> Alan,
>
> Thanks for the hint on Qt4 vs Qt5 - probably spared me a few hours of pain!
> I set out to trial the memqt driver and ran into some issues with the build
> on MSVC 2012.
>
> 1) For some reason, which I had never encountered before, the CMake system
> generates incorrect build files when the source path includes spaces.  Not a
> big deal really, I can just move to a path without spaces.  I don't recall
> the exact issue, but I think it broke up include paths into separate
> entries.  I am using a fairly old version of CMake (2.8.12.2)

That's an issue with our own build system and not CMake.  From time to
time we try and clean up some aspect of this issue, but it is a low
priority because it is so easy to use PATH's without spaces in them.

>
> I started with Qt configured for dynamic linking
>
> 2) I tried to build PLplot as a shared lib, then realized that was not
> really going to work for me since my application accessed some other
> normally private functions, not exported in the DLL (pldtik, pldprec, etc).
>
> 3) I tried to build PLplot as a static lib (which I normally do for my app
> anyway).  But then I started getting incompatible linkage warnings and some
> errors about trying to initialize dllimport'ed data
> (\plplot-plplot\bindings\qt_gui\plqt.cpp(28): error C2491: 'vectorize' :
> definition of dllimport data not allowed).  After finding/following the
> thread here (http://ehc.ac/p/plplot/mailman/message/33603310/), I began to
> suspect that the USINGDLL macro might be an issue.  Despite the fact that I
> configured PLplot for static build it seems that the cmake scripts still
> applied USINGDLL to the plqt.c file. Is that expected (I don't totally
> understand the use of USINGDLL in all the contexts)?  Is it perhaps related
> to the issue described in that thread on the devel list?

Are you using PLplot-5.10.0 or the git master tip version (which is
about to turn into PLplot-5.11.0)?  I would prefer you to try the
master tip version if you haven't been doing that already because it
has been a very long time since 5.10.0 was released, I am much more
familiar with the master tip version, and I would like to get the
master tip version tested as well as possible before the 5.11.0
release.  For example, if you spot a USINGDLL issue for that version
(such as USINGDLL being defined for the static library case), please
let me know with lots of details.

Alan
__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Am I going about this the best way?

[Aaron H. <he...@co...>]

Alan,

Thanks for the hint on Qt4 vs Qt5 - probably spared me a few hours of pain!
I set out to trial the memqt driver and ran into some issues with the build
on MSVC 2012.

1) For some reason, which I had never encountered before, the CMake system
generates incorrect build files when the source path includes spaces.  Not a
big deal really, I can just move to a path without spaces.  I don't recall
the exact issue, but I think it broke up include paths into separate
entries.  I am using a fairly old version of CMake (2.8.12.2)

I started with Qt configured for dynamic linking

2) I tried to build PLplot as a shared lib, then realized that was not
really going to work for me since my application accessed some other
normally private functions, not exported in the DLL (pldtik, pldprec, etc).

3) I tried to build PLplot as a static lib (which I normally do for my app
anyway).  But then I started getting incompatible linkage warnings and some
errors about trying to initialize dllimport'ed data
(\plplot-plplot\bindings\qt_gui\plqt.cpp(28): error C2491: 'vectorize' :
definition of dllimport data not allowed).  After finding/following the
thread here (http://ehc.ac/p/plplot/mailman/message/33603310/), I began to
suspect that the USINGDLL macro might be an issue.  Despite the fact that I
configured PLplot for static build it seems that the cmake scripts still
applied USINGDLL to the plqt.c file. Is that expected (I don't totally
understand the use of USINGDLL in all the contexts)?  Is it perhaps related
to the issue described in that thread on the devel list?

Thanks,

Aaron.

-----Original Message-----
From: Alan W. Irwin [mailto:ir...@be...] 
Sent: Thursday, March 26, 2015 1:58 AM
To: Aaron Hexamer
Cc: plp...@li...
Subject: RE: [Plplot-general] Am I going about this the best way?

On 2015-03-26 00:08-0500 Aaron Hexamer wrote:

> Thanks, Alan.  I'll check out Qt. Looks like it has a mem driver too.  
> I'll just have to try and see which dependencies are lightest.

P.S.  At this stage I would definitely recommend the latest version of
Qt4 rather than experimenting with Qt5.  The latter does mostly work with
our qt device driver and qt_example (using the experimental
-DPLPLOT_USE_QT5=ON option), but is just not nearly as reliable as
Qt4 in our experience.  I am sure Qt5 will eventually get there, but it just
needs a lot more time to mature properly.

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); the Time Ephemerides project
(timeephem.sf.net); PLplot scientific plotting software package
(plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux
Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Build Plplot with ifort on OS X

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

On 2015-03-30 14:26-0000 Arjen Markus wrote:

> Hello Hiro,
>
>
>
> I am not an expert on pkg-config, but I guess that the problem is due to you specifying two keywords for it. Could you try:
>
>
>
> export PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig
> /usr/bin/ifort  x00f.f90 -o x00f `pkg-config  --cflags plplotd-f95` `pkg-config --libs plplotd-f95` -lplf95demolibd
>
> (I moved the setting of PKG_CONFIG_PATH to a separate command for convenience.)
>
> That way pkg-config is called twice, once for the cflags property and once for the libs property (or whatever it is called ;)).

@Arjen:

In my experience running pkg-config twice like that with the separate
options should give exactly the same result as running it once with
the two options.

@Hiro:

We are desperate here to find a solution to this extremely puzzling
issue so by all means try any experiment you can think of including
this one suggested by Arjen or the one I suggested previously to cut
and paste pkg-config results so that your hand-crafted build command
doesn't use pkg-config directly.

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Build Plplot with ifort on OS X

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

On 2015-03-30 22:38+0900 Hiroyasu Yasuda wrote:

> Hello Alan,
>
> Thank you for your kindness. I ran your suggested commands and that results as below:
>
> hiro$ PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" pkg-config --cflags plplotd-f95
> -I/Users/hiro/Desktop/plplot_ifort/include/plplot -I/Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot -I/Users/hiro/Desktop/plplot_ifort/include/plplot
>
> hiro$ make x00f
> /usr/bin/ifort  x00f.f90 -o x00f  `PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" pkg-config  --cflags --libs plplotd-f95` -lplf95demolibd
> x00f.f90(24): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [PLPLOT_STR]
>    use plf95demolib
> --------^
> compilation aborted for x00f.f90 (code 1)
> make: *** [x00f] Error 1
>
> Also I intend a file list in specified fold with pkg-config as below:
>
> hiro$ ls -l /Users/hiro/Desktop/plplot_ifort/include/plplot
> total 560
> -rw-r--r--  1 hiro  staff   3255 Mar 30 17:49 disptab.h
> -rw-r--r--  1 hiro  staff   7284 Mar 30 17:49 drivers.h
> -rw-r--r--  1 hiro  staff   5398 Mar 30 17:49 gcw.h
> -rw-r--r--  1 hiro  staff   4043 Mar 30 17:49 pdf.h
> -rw-r--r--  1 hiro  staff   3304 Mar 30 17:51 plConfig.h
> -rw-r--r--  1 hiro  staff   2802 Mar 30 17:51 plDevs.h
> -rw-r--r--  1 hiro  staff   2763 Mar 30 17:49 pldebug.h
> -rw-r--r--  1 hiro  staff   5761 Mar 30 17:51 pldll.h
> -rw-r--r--  1 hiro  staff   7509 Mar 30 17:49 plevent.h
> -rw-r--r--  1 hiro  staff  77370 Mar 30 17:49 plplot.h
> -rw-r--r--  1 hiro  staff  32642 Mar 30 17:49 plplotP.h
> -rw-r--r--  1 hiro  staff  28357 Mar 30 17:49 plplotcanvas.h
> -rw-r--r--  1 hiro  staff  42350 Mar 30 17:49 plstream.h
> -rw-r--r--  1 hiro  staff  30762 Mar 30 17:49 plstrm.h
> -rw-r--r--  1 hiro  staff   4466 Mar 30 17:49 plxwd.h
> -rw-r--r--  1 hiro  staff   2633 Mar 30 17:49 qsastime.h
> -rw-r--r--  1 hiro  staff   1612 Mar 30 17:49 qsastimedll.h
>
> hiro$ ls -l /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot
> total 392
> -rw-r--r--  1 hiro  staff    2447 Mar 30 17:51 plf95demolib.mod
> -rw-r--r--  1 hiro  staff  125957 Mar 30 17:51 plplot.mod
> -rw-r--r--  1 hiro  staff     629 Mar 30 17:51 plplot_flt.mod
> -rw-r--r--  1 hiro  staff   62576 Mar 30 17:51 plplotp.mod
>

> Those results seem to be working fine except error #7002. Since
error #7002 is an intrinsic error message, the built result of *.mod
with ifort is suspected. If there are some ifort compiling options to
obtain a normal built result, an built result of *.mod with ifort will
be equaled to an built results with gfortran. Are there some compiling
option for option?

I could be wrong, but my interpretation of the error message is ifort
cannot find the fortran modules (specifically the plf95demolib module
file) despite the -I option pointing right to them.  That is an
extremely puzzling result. That is also similar to a result another
user got with ifort back in 2011, see
<https://sourceforge.net/p/plplot/bugs/107/>.

According to <http://inside.mines.edu/mio/man/ifort.html> the -I
option should be recognized by ifort to help that compiler find the
modules, but since that appears not to be working for your platform I
wonder if you have a different version of ifort that does not
recognize this option? What does your own ifort manual say about the
option that should be used to find modules?

Also, I suggest you experiment with the build command to see what you
have to modify to get it to work.

Currently, according to your result above, the Makefile generates the
following command.

/usr/bin/ifort  x00f.f90 -o x00f  \
`PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" \
pkg-config --cflags --libs plplotd-f95` -lplf95demolibd

I would run that by hand (outside the Makefile) to confirm you get the
same error, then cut and paste the result of


PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" \
pkg-config --cflags --libs plplotd-f95

to remove pkg-config from the result.  Then modify the relevant -I
option according to what your ifort documentation says to do to see if
you can convince ifort to find your installed modules directory at
/Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot.

Sorry I cannot be of more explicit help, but this is a tough issue
that I think can only be figured out by experimentation on your part.

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Build Plplot with ifort on OS X

[Arjen M. <Arj...@de...>]

Hello Hiro,



I am not an expert on pkg-config, but I guess that the problem is due to you specifying two keywords for it. Could you try:



export PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig
/usr/bin/ifort  x00f.f90 -o x00f `pkg-config  --cflags plplotd-f95` `pkg-config --libs plplotd-f95` -lplf95demolibd

(I moved the setting of PKG_CONFIG_PATH to a separate command for convenience.)

That way pkg-config is called twice, once for the cflags property and once for the libs property (or whatever it is called ;)).

Regards,

Arjen

> -----Original Message-----
> From: Hiroyasu Yasuda [mailto:hi...@gs...]
> Sent: Monday, March 30, 2015 3:38 PM
> To: Alan W. Irwin
> Cc: plplot_general
> Subject: Re: [Plplot-general] Build Plplot with ifort on OS X
>
> Hello Alan,
>
> Thank you for your kindness. I ran your suggested commands and that results as
> below:
>
> hiro$ PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" pkg-
> config --cflags plplotd-f95 -I/Users/hiro/Desktop/plplot_ifort/include/plplot -
> I/Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot -
> I/Users/hiro/Desktop/plplot_ifort/include/plplot
>
> hiro$ make x00f
> /usr/bin/ifort  x00f.f90 -o x00f
> `PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" pkg-config  --
> cflags --libs plplotd-f95` -lplf95demolibd
> x00f.f90(24): error #7002: Error in opening the compiled module file.  Check
> INCLUDE paths.   [PLPLOT_STR]
>     use plf95demolib
> --------^
> compilation aborted for x00f.f90 (code 1)
> make: *** [x00f] Error 1
>
> Also I intend a file list in specified fold with pkg-config as below:
>
>  hiro$ ls -l /Users/hiro/Desktop/plplot_ifort/include/plplot
> total 560
> -rw-r--r--  1 hiro  staff   3255 Mar 30 17:49 disptab.h
> -rw-r--r--  1 hiro  staff   7284 Mar 30 17:49 drivers.h
> -rw-r--r--  1 hiro  staff   5398 Mar 30 17:49 gcw.h
> -rw-r--r--  1 hiro  staff   4043 Mar 30 17:49 pdf.h
> -rw-r--r--  1 hiro  staff   3304 Mar 30 17:51 plConfig.h
> -rw-r--r--  1 hiro  staff   2802 Mar 30 17:51 plDevs.h
> -rw-r--r--  1 hiro  staff   2763 Mar 30 17:49 pldebug.h
> -rw-r--r--  1 hiro  staff   5761 Mar 30 17:51 pldll.h
> -rw-r--r--  1 hiro  staff   7509 Mar 30 17:49 plevent.h
> -rw-r--r--  1 hiro  staff  77370 Mar 30 17:49 plplot.h
> -rw-r--r--  1 hiro  staff  32642 Mar 30 17:49 plplotP.h
> -rw-r--r--  1 hiro  staff  28357 Mar 30 17:49 plplotcanvas.h
> -rw-r--r--  1 hiro  staff  42350 Mar 30 17:49 plstream.h
> -rw-r--r--  1 hiro  staff  30762 Mar 30 17:49 plstrm.h
> -rw-r--r--  1 hiro  staff   4466 Mar 30 17:49 plxwd.h
> -rw-r--r--  1 hiro  staff   2633 Mar 30 17:49 qsastime.h
> -rw-r--r--  1 hiro  staff   1612 Mar 30 17:49 qsastimedll.h
>
> hiro$ ls -l /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot
> total 392
> -rw-r--r--  1 hiro  staff    2447 Mar 30 17:51 plf95demolib.mod
> -rw-r--r--  1 hiro  staff  125957 Mar 30 17:51 plplot.mod
> -rw-r--r--  1 hiro  staff     629 Mar 30 17:51 plplot_flt.mod
> -rw-r--r--  1 hiro  staff   62576 Mar 30 17:51 plplotp.mod
>
> Those results seem to be working fine except error #7002. Since error #7002 is an
> intrinsic error message, the built result of *.mod with ifort is suspected. If there are
> some ifort compiling options to obtain a normal built result, an built result of *.mod
> with ifort will be equaled to an built results with gfortran. Are there some compiling
> option for option?
>
> Sincerely,
> Hiro
>
>
> > On Mar 30, 2015, at 01:57, Alan W. Irwin <ir...@be...> wrote:
> >
> > On 2015-03-29 14:26+0900 Hiroyasu Yasuda wrote:
> >
> >> Hello Alan,
> >>
> >> Thank you for the reply.
> >>
> >>> If all appears to be well with that option, then please send the
> >>> Makefile located at $PREFIX/plplot5.10.0/examples/f95/Makefile
> >>> as well as the results from
> >>>
> >>> make x00f >& x00f.out
> >>>
> >>> in that directory.
> >>
> >> At my last mail, according to your mail as above I attached an output of result
> when I ran Make in $PREFIX/plplot5.10.0/examples/f95/ as “make x00f >& x00f.out”.
> If your intention misunderstood me, please let me know your request as specific
> commands again.
> >
> > Hi Hiro:
> >
> > I was referring to a different part of that e-mail which I replicate
> > here for your convenience:
> >
> > ===========
> > If you look carefully at the Makefile, the command it uses to find the
> > necessary compile flags to build one of the f95 examples is
> >
> > PKG_CONFIG_PATH=$PREFIX/lib/pkg-config pkgconfig --cflags plplot-f95
> >
> > When I run that command here with appropriate $PREFIX, here are the
> > results:
> >
> > -I<PREFIX>/include/plplot -I<PREFIX>/lib/fortran/modules/plplot
> >
> > where the <PREFIX> result should be the same as $PREFIX.
> >
> > What are the results of that command there?
> > ===========
> >
> > It's that last question I need answered.
> >
> > Alan
> > __________________________
> > Alan W. Irwin
> >
> > Astronomical research affiliation with Department of Physics and
> > Astronomy, University of Victoria (astrowww.phys.uvic.ca).
> >
> > Programming affiliations with the FreeEOS equation-of-state
> > implementation for stellar interiors (freeeos.sf.net); the Time
> > Ephemerides project (timeephem.sf.net); PLplot scientific plotting
> > software package (plplot.sf.net); the libLASi project
> > (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
> > and the Linux Brochure Project (lbproject.sf.net).
> > __________________________
> >
> > Linux-powered Science
> > __________________________
>
>
>
> - - - - -
> 安田 浩保
> 〒950-2181 新潟市西区五十嵐二の町8050
> 新潟大学 災害・復興科学研究所
> 総合研究棟（環境エネルギ系）213号室
> TEL : 025-262-7053  FAX : 025-262-7050
> E-mail : hi...@gs...
>
> Hiroyasu YASUDA
> Laboratory of River Research
> Research Institute for Natural Hazards & Disaster Recovery Niigata University
> Ikarashi 2-no-cho, Nis-ku, Niigata, 950-2181, Japan Phone +81-25-262-7053
> Facsimile +81-25-262-7050 E-mail hi...@gs... URL
> http://rde.nhdr.niigata-u.ac.jp/lab/
>
> .´￣`.
> : (` :
> `. `´   :
>   `・--´
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored by
> Intel and developed in partnership with Slashdot Media, is your hub for all things
> parallel software development, from weekly thought leadership blogs to news, videos,
> case studies, tutorials and more. Take a look and join the conversation now.
> http://goparallel.sourceforge.net/
> _______________________________________________
> Plplot-general mailing list
> Plp...@li...
> https://lists.sourceforge.net/lists/listinfo/plplot-general

DISCLAIMER: This message is intended exclusively for the addressee(s) and may contain confidential and privileged information. If you are not the intended recipient please notify the sender immediately and destroy this message. Unauthorized use, disclosure or copying of this message is strictly prohibited. The foundation 'Stichting Deltares', which has its seat at Delft, The Netherlands, Commercial Registration Number 41146461, is not liable in any way whatsoever for consequences and/or damages resulting from the improper, incomplete and untimely dispatch, receipt and/or content of this e-mail.

Re: [Plplot-general] Build Plplot with ifort on OS X

[Hiroyasu Y. <hi...@gs...>]

Hello Alan, 

Thank you for your kindness. I ran your suggested commands and that results as below:

hiro$ PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" pkg-config --cflags plplotd-f95
-I/Users/hiro/Desktop/plplot_ifort/include/plplot -I/Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot -I/Users/hiro/Desktop/plplot_ifort/include/plplot 

hiro$ make x00f
/usr/bin/ifort  x00f.f90 -o x00f  `PKG_CONFIG_PATH="/Users/hiro/Desktop/plplot_ifort/lib/pkgconfig" pkg-config  --cflags --libs plplotd-f95` -lplf95demolibd
x00f.f90(24): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [PLPLOT_STR]
    use plf95demolib
--------^
compilation aborted for x00f.f90 (code 1)
make: *** [x00f] Error 1

Also I intend a file list in specified fold with pkg-config as below:

 hiro$ ls -l /Users/hiro/Desktop/plplot_ifort/include/plplot
total 560
-rw-r--r--  1 hiro  staff   3255 Mar 30 17:49 disptab.h
-rw-r--r--  1 hiro  staff   7284 Mar 30 17:49 drivers.h
-rw-r--r--  1 hiro  staff   5398 Mar 30 17:49 gcw.h
-rw-r--r--  1 hiro  staff   4043 Mar 30 17:49 pdf.h
-rw-r--r--  1 hiro  staff   3304 Mar 30 17:51 plConfig.h
-rw-r--r--  1 hiro  staff   2802 Mar 30 17:51 plDevs.h
-rw-r--r--  1 hiro  staff   2763 Mar 30 17:49 pldebug.h
-rw-r--r--  1 hiro  staff   5761 Mar 30 17:51 pldll.h
-rw-r--r--  1 hiro  staff   7509 Mar 30 17:49 plevent.h
-rw-r--r--  1 hiro  staff  77370 Mar 30 17:49 plplot.h
-rw-r--r--  1 hiro  staff  32642 Mar 30 17:49 plplotP.h
-rw-r--r--  1 hiro  staff  28357 Mar 30 17:49 plplotcanvas.h
-rw-r--r--  1 hiro  staff  42350 Mar 30 17:49 plstream.h
-rw-r--r--  1 hiro  staff  30762 Mar 30 17:49 plstrm.h
-rw-r--r--  1 hiro  staff   4466 Mar 30 17:49 plxwd.h
-rw-r--r--  1 hiro  staff   2633 Mar 30 17:49 qsastime.h
-rw-r--r--  1 hiro  staff   1612 Mar 30 17:49 qsastimedll.h

hiro$ ls -l /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot
total 392
-rw-r--r--  1 hiro  staff    2447 Mar 30 17:51 plf95demolib.mod
-rw-r--r--  1 hiro  staff  125957 Mar 30 17:51 plplot.mod
-rw-r--r--  1 hiro  staff     629 Mar 30 17:51 plplot_flt.mod
-rw-r--r--  1 hiro  staff   62576 Mar 30 17:51 plplotp.mod

Those results seem to be working fine except error #7002. Since error #7002 is an intrinsic error message, the built result of *.mod with ifort is suspected. If there are some ifort compiling options to obtain a normal built result, an built result of *.mod with ifort will be equaled to an built results with gfortran. Are there some compiling option for option?

Sincerely,
Hiro


> On Mar 30, 2015, at 01:57, Alan W. Irwin <ir...@be...> wrote:
> 
> On 2015-03-29 14:26+0900 Hiroyasu Yasuda wrote:
> 
>> Hello Alan,
>> 
>> Thank you for the reply.
>> 
>>> If all appears to be well with that option, then please send the
>>> Makefile located at $PREFIX/plplot5.10.0/examples/f95/Makefile
>>> as well as the results from
>>> 
>>> make x00f >& x00f.out
>>> 
>>> in that directory.
>> 
>> At my last mail, according to your mail as above I attached an output of result when I ran Make in $PREFIX/plplot5.10.0/examples/f95/ as “make x00f >& x00f.out”. If your intention misunderstood me, please let me know your request as specific commands again.
> 
> Hi Hiro:
> 
> I was referring to a different part of that e-mail which I replicate
> here for your convenience:
> 
> ===========
> If you look carefully at the Makefile, the command it uses to find the
> necessary compile flags to build one of the f95 examples is
> 
> PKG_CONFIG_PATH=$PREFIX/lib/pkg-config pkgconfig --cflags plplot-f95
> 
> When I run that command here with appropriate $PREFIX, here are the
> results:
> 
> -I<PREFIX>/include/plplot -I<PREFIX>/lib/fortran/modules/plplot
> 
> where the <PREFIX> result should be the same as $PREFIX.
> 
> What are the results of that command there?
> ===========
> 
> It's that last question I need answered.
> 
> Alan
> __________________________
> Alan W. Irwin
> 
> Astronomical research affiliation with Department of Physics and Astronomy,
> University of Victoria (astrowww.phys.uvic.ca).
> 
> Programming affiliations with the FreeEOS equation-of-state
> implementation for stellar interiors (freeeos.sf.net); the Time
> Ephemerides project (timeephem.sf.net); PLplot scientific plotting
> software package (plplot.sf.net); the libLASi project
> (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
> and the Linux Brochure Project (lbproject.sf.net).
> __________________________
> 
> Linux-powered Science
> __________________________



- - - - -
安田 浩保
〒950-2181 新潟市西区五十嵐二の町8050
新潟大学 災害・復興科学研究所
総合研究棟（環境エネルギ系）213号室
TEL : 025-262-7053  FAX : 025-262-7050
E-mail : hi...@gs...

Hiroyasu YASUDA
Laboratory of River Research
Research Institute for Natural Hazards & Disaster Recovery
Niigata University
Ikarashi 2-no-cho, Nis-ku, Niigata, 950-2181, Japan 
Phone +81-25-262-7053  Facsimile +81-25-262-7050 
E-mail hi...@gs... 
URL http://rde.nhdr.niigata-u.ac.jp/lab/

.´￣`.
: (` :  
`. `´   :
  `・--´　

Re: [Plplot-general] Build Plplot with ifort on OS X

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

On 2015-03-29 14:26+0900 Hiroyasu Yasuda wrote:

> Hello Alan,
>
> Thank you for the reply.
>
>> If all appears to be well with that option, then please send the
>> Makefile located at $PREFIX/plplot5.10.0/examples/f95/Makefile
>> as well as the results from
>>
>> make x00f >& x00f.out
>>
>> in that directory.
>
> At my last mail, according to your mail as above I attached an output of result when I ran Make in $PREFIX/plplot5.10.0/examples/f95/ as “make x00f >& x00f.out”. If your intention misunderstood me, please let me know your request as specific commands again.

Hi Hiro:

I was referring to a different part of that e-mail which I replicate
here for your convenience:

===========
If you look carefully at the Makefile, the command it uses to find the
necessary compile flags to build one of the f95 examples is

PKG_CONFIG_PATH=$PREFIX/lib/pkg-config pkgconfig --cflags plplot-f95

When I run that command here with appropriate $PREFIX, here are the
results:

-I<PREFIX>/include/plplot -I<PREFIX>/lib/fortran/modules/plplot

where the <PREFIX> result should be the same as $PREFIX.

What are the results of that command there?
===========

It's that last question I need answered.

Alan
__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Build Plplot with ifort on OS X

[Hiroyasu Y. <hi...@gs...>]

Hello Alan,

Thank you for the reply.

> If all appears to be well with that option, then please send the
> Makefile located at $PREFIX/plplot5.10.0/examples/f95/Makefile
> as well as the results from
> 
> make x00f >& x00f.out
> 
> in that directory.

At my last mail, according to your mail as above I attached an output of result when I ran Make in $PREFIX/plplot5.10.0/examples/f95/ as “make x00f >& x00f.out”. If your intention misunderstood me, please let me know your request as specific commands again.

Sincerely,
Hiro



> On Mar 29, 2015, at 03:05, Alan W. Irwin <ir...@be...> wrote:
> 
> On 2015-03-28 17:31+0900 Hiroyasu Yasuda wrote:
> 
>>> (2) pkg-config issue?
>> 
>> I would like to send you a build result as attached filed: x00f.out.
>> 
> 
> Hi Hiro:
> 
> Please review my original reply where I asked you a question about the
> pkg-config output on your system.  I need that information to proceed
> further.
> 
> Alan
> 
> __________________________
> Alan W. Irwin
> 
> Astronomical research affiliation with Department of Physics and Astronomy,
> University of Victoria (astrowww.phys.uvic.ca).
> 
> Programming affiliations with the FreeEOS equation-of-state
> implementation for stellar interiors (freeeos.sf.net); the Time
> Ephemerides project (timeephem.sf.net); PLplot scientific plotting
> software package (plplot.sf.net); the libLASi project
> (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
> and the Linux Brochure Project (lbproject.sf.net).
> __________________________
> 
> Linux-powered Science
> __________________________<x00f.out>
> 
>> On Mar 28, 2015, at 02:18, Alan W. Irwin <ir...@be...> 
> wrote:
> 
>> 
> 
>> On 2015-03-27 19:34+0900 Hiroyasu Yasuda wrote:
> 
>> 
> 
>>> Hello all
> 
>>> 
> 
>>> I’ve tried to build plplot with ifort on OS X as below 
> follows:
> 
>>> 
> 
>>> export CC="gcc -O2"
> 
>>> export CXX="g++ -O2"
> 
>>> export FC=“ifort -O2"
> 
>>> 
> 
>>> cmake -DCMAKE_INSTALL_PREFIX=~/Desktop/plplot_ifort 
> -DENABLE_tcl=OFF -DENABLE_tk=OFF -DENABLE_python=OFF 
> -DENABLE_qt=OFF -DENABLE_java=OFF -DENABLE_lua=OFF 
> -DENABLE_wxwidgets=OFF ..
> 
>>> 
> 
>>> make
> 
>>> 
> 
>>> make install
> 
>>> 
> 
>> 
> 
>>> Actually processes of build and install are seemingly success
> 
>> according to attached cmake.t, make.t and make.install.t. However when
> 
>> build example codes of f95 in share/plplot5.10.0/examples/f95, build
> 
>> process stopped with as below error messages:
> 
>> 
> 
>>> x00f.f90(24): error #7002: Error in opening the compiled module file. 
> Check INCLUDE paths.   [PLPLOT_STR]
> 
>> 
> 
>>> If latest plplot support the ifort as fortran compilers, would you 
> like to tell me a way to build plplot with ifort?
> 
>> 
> 
>> Hi Hiro:
> 
>> 
> 
>> Thanks for your interest in PLplot.  To answer your question, it 
> appears all went well with
> 
>> the core build and install of PLplot, but the subsequent build of the
> 
>> installed examples is giving you some trouble.
> 
>> 
> 
>> 
> 
>> We actually have two separate build systems for installed examples.
> 
>> The traditional one uses make and pkg-config to build the examples,
> 
>> and we have also implemented a CMake-based build system for the
> 
>> installed examples as well.  If you changed directory to
> 
>> $PREFIX/plplot5.10.0/examples/f95 (where $PREFIX is your install
> 
>> prefix ~/Desktop/plplot_ifort) and ran the "make" command, then
> 
>> you are using the traditional build system.  Some relevant questions
> 
>> in that case are (1) are the module files installed properly, and
> 
>> (2) if so, what pkg-config issue is preventing those from being
> 
>> seen by the ifort compiler?
> 
>> 
> 
>> (1) Module files installed properly?
> 
>> 
> 
>> On my (Debian stable) system, the relevant Fortran module files are
> 
>> installed at $PREFIX/lib/fortran/modules/plplot/. For my gfortran case
> 
>> those files are called
> 
>> 
> 
>> software@raven> ls -l
> 
>> total 576
> 
>> -rw-r--r-- 1 software software 246325 Mar 26 12:12 plf95demolib.mod
> 
>> -rw-r--r-- 1 software software 243851 Mar 26 12:12 plplot.mod
> 
>> -rw-r--r-- 1 software software   3394 Mar 26 12:12 plplot_graphics.mod
> 
>> -rw-r--r-- 1 software software   6764 Mar 26 12:12 plplot_str.mod
> 
>> -rw-r--r-- 1 software software   2896 Mar 26 12:12 plplot_strutils.mod
> 
>> -rw-r--r-- 1 software software   1318 Mar 26 12:12 plplot_types.mod
> 
>> -rw-r--r-- 1 software software  72858 Mar 26 12:12 plplotp.mod
> 
>> 
> 
>> For ifort you should have equivalent files (probably with the same
> 
>> names, but with very different content since ifort uses a different
> 
>> module file format than gfortran) installed in that location for all
> 
>> these modules.  In fact, if you look at your make.install.t file, the
> 
>> following lines appear
> 
>> 
> 
>> -- Installing: 
> /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plplot.mod
> 
>> -- Installing: 
> /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plplotp.mod
> 
>> -- Installing: 
> /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plplot_flt.mod
> 
> 
>> -- Installing: 
> /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plf95demolib.m
> od
> 
>> 
> 
>> I think you are likely fine, and the difference between the two
> 
>> platforms is simply I am using the master tip version of PLplot (which
> 
>> has more Fortran modules) while you are using the 5.10.0 version.
> 
>> (By the way, PLplot-5.11.0 is coming soon based on the git master tip
> 
>> version of PLplot.)
> 
>> 
> 
>> (2) pkg-config issue?
> 
>> 
> 
>> When you execute the "make" command above, that should automatically 
> invoke
> 
>> pkg-config to find the relevant compile options for the ifort case.
> 
>> 
> 
>> If you look carefully at the Makefile, the command it uses to find the
> 
>> necessary compile flags to build one of the f95 examples is
> 
>> 
> 
>> PKG_CONFIG_PATH=$PREFIX/lib/pkg-config pkgconfig --cflags plplot-f95
> 
>> 
> 
>> When I run that command here with appropriate $PREFIX, here are the
> 
>> results:
> 
>> 
> 
>> -I<PREFIX>/include/plplot -I<PREFIX>/lib/fortran/modules/plplot
> 
>> 
> 
>> where the <PREFIX> result should be the same as $PREFIX.
> 
>> 
> 
>> What are the results of that command there?
> 
>> 
> 
>> Note the reference above to -I<PREFIX>/lib/fortran/modules/plplot
> 
>> which your results should have as well. That is the compiler option 
> that
> 
>> guarantees that gfortran (in my case) and ifort (in your case) finds
> 
>> the relevant installed module files.
> 
>> 
> 
>> If all appears to be well with that option, then please send the
> 
>> Makefile located at $PREFIX/plplot5.10.0/examples/f95/Makefile
> 
>> as well as the results from
> 
>> 
> 
>> make x00f >& x00f.out
> 
>> 
> 
> 
> 
> 
> 
> 
> 
>> in that directory.
> 
>> 
> 
>> Alan
> 
>> __________________________
> 
>> Alan W. Irwin
> 
>> 
> 
>> Astronomical research affiliation with Department of Physics and 
> Astronomy,
> 
>> University of Victoria (astrowww.phys.uvic.ca).
> 
>> 
> 
>> Programming affiliations with the FreeEOS equation-of-state
> 
>> implementation for stellar interiors (freeeos.sf.net); the Time
> 
>> Ephemerides project (timeephem.sf.net); PLplot scientific plotting
> 
>> software package (plplot.sf.net); the libLASi project
> 
>> (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
> 
>> and the Linux Brochure Project (lbproject.sf.net).
> 
>> __________________________
> 
>> 
> 
>> Linux-powered Science
> 
>> __________________________
> 
> 
> 
> 
> 
> 
> 
> .´￣`.
> 
> : (` :  
> 
> `. `´   :
> 
>  `・--´　
> 
> 
> 



- - - - -
安田 浩保
〒950-2181 新潟市西区五十嵐二の町8050
新潟大学 災害・復興科学研究所
総合研究棟（環境エネルギ系）213号室
TEL : 025-262-7053  FAX : 025-262-7050
E-mail : hi...@gs...

Hiroyasu YASUDA
Laboratory of River Research
Research Institute for Natural Hazards & Disaster Recovery
Niigata University
Ikarashi 2-no-cho, Nis-ku, Niigata, 950-2181, Japan 
Phone +81-25-262-7053  Facsimile +81-25-262-7050 
E-mail hi...@gs... 
URL http://rde.nhdr.niigata-u.ac.jp/lab/

.´￣`.
: (` :  
`. `´   :
  `・--´　

Re: [Plplot-general] Build Plplot with ifort on OS X

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

On 2015-03-28 17:31+0900 Hiroyasu Yasuda wrote:

>> (2) pkg-config issue?
>
> I would like to send you a build result as attached filed: x00f.out.
>

Hi Hiro:

Please review my original reply where I asked you a question about the
pkg-config output on your system.  I need that information to proceed
further.

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Build Plplot with ifort on OS X

[Hiroyasu Y. <hi...@gs...>]

Hello Alan,

Thank you for the reply. I confirm your suggestions as below:

> (1) Module files installed properly?


I checked out the tree of $PREFIX/lib/fortran/modules/plplot when FC was specified the ifort:

total 392
-rw-r--r--  1 hiro  staff    2447 Mar 28 16:58 plf95demolib.mod
-rw-r--r--  1 hiro  staff  125957 Mar 28 16:58 plplot.mod
-rw-r--r--  1 hiro  staff     629 Mar 28 16:58 plplot_flt.mod
-rw-r--r--  1 hiro  staff   62576 Mar 28 16:58 plplotp.mod

Also I checked out the tree of $PREFIX/lib/fortran/modules/plplot when FC was specified the gfortan:

total 1080
-rw-r--r--  1 hiro  staff  236690 Mar 27 19:21 plf95demolib.mod
-rw-r--r--  1 hiro  staff  234190 Mar 27 19:21 plplot.mod
-rw-r--r--  1 hiro  staff    1108 Mar 27 19:21 plplot_flt.mod
-rw-r--r--  1 hiro  staff   73302 Mar 27 19:21 plplotp.mod

Moreover I indicated the tree of $PREFIX/lib/fortran/modules/plplot when FC was specified the gfortan and Macports was employed as an installer:

total 1080
-rw-r--r--  1 root  wheel  236790 Nov 28 10:59 plf95demolib.mod
-rw-r--r--  1 root  wheel  234290 Nov 28 10:59 plplot.mod
-rw-r--r--  1 root  wheel    1208 Nov 28 10:59 plplot_flt.mod
-rw-r--r--  1 root  wheel   73402 Nov 28 10:59 plplotp.mod


> (2) pkg-config issue?

I would like to send you a build result as attached filed: x00f.out.

Please let me know a way to build plplot with ifort.

Sincerely,
Hiro

[Plplot-general] Testing of git master tip requested

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

I have just completed running the scripts/comprehensive_test.sh script
without issues on Linux (Debian testing) for the master tip version of
PLplot as part of the preparation for the forthcoming release of
PLplot-5.11.0.  So to test this forthcoming release on as wide a
variety of platforms as possible before the release, I ask those on
this list to run that script on any platform of interest for them.

In order to participate in this test you need all the software
installed on your platform that is normally used to help build PLplot.
You also need access to bash (since the shell scripts that are used to
comprehensively test PLplot have bashisms in them).  Such access is
easy to obtain on Linux and Mac OS X.  Windows is a little more
problematic, but bash.exe is provided by any of MinGW/MSYS,
MinGW-w64/MSYS2, or Cygwin on Windows platforms.  So comprehensive
testing on any of those Unix-like Windows platforms should be
straightforward if you have installed bash.exe.

In theory you should also be able to comprehensivly test MSVC with
that script if you simply put one of those bash versions on your PATH
and specify appropriate options for scripts/comprehensive_test.sh to
use nmake to build the software (using the "NMake Makefiles"
generator), and also exclude testing the traditional build of the
installed examples (which requires both make and pkg-config which are
not consistent with nmake and MSVC) from the comprehensive testing for
the MSVC case.

To learn more about the many options possible for that script run

scripts/comprehensive_test.sh --help

I also highly recommend that you consult
<https://sourceforge.net/p/plplot/wiki/Testing_PLplot/#Comprehensive%20testing>
for more overview details for running that script.

Comprehensive testing results for Linux are always welcome, but such
results are particularly welcome for Mac OS X, Cygwin, MinGW/MSYS,
MinGW-w64/MSYS2, and MSVC as well since those platforms are currently
not comprehensively tested at all or only tested with a subset
of the tests that are possible with scripts/comprehensive_test.sh.

Alan
__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] Build Plplot with ifort on OS X

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

On 2015-03-27 19:34+0900 Hiroyasu Yasuda wrote:

> Hello all
>
> I’ve tried to build plplot with ifort on OS X as below follows:
>
> export CC="gcc -O2"
> export CXX="g++ -O2"
> export FC=“ifort -O2"
>
> cmake -DCMAKE_INSTALL_PREFIX=~/Desktop/plplot_ifort -DENABLE_tcl=OFF -DENABLE_tk=OFF -DENABLE_python=OFF -DENABLE_qt=OFF -DENABLE_java=OFF -DENABLE_lua=OFF -DENABLE_wxwidgets=OFF ..
>
> make
>
> make install
>

> Actually processes of build and install are seemingly success
according to attached cmake.t, make.t and make.install.t. However when
build example codes of f95 in share/plplot5.10.0/examples/f95, build
process stopped with as below error messages:

> x00f.f90(24): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [PLPLOT_STR]

> If latest plplot support the ifort as fortran compilers, would you like to tell me a way to build plplot with ifort?

Hi Hiro:

Thanks for your interest in PLplot.  To answer your question, it appears all went well with
the core build and install of PLplot, but the subsequent build of the
installed examples is giving you some trouble.


We actually have two separate build systems for installed examples.
The traditional one uses make and pkg-config to build the examples,
and we have also implemented a CMake-based build system for the
installed examples as well.  If you changed directory to
$PREFIX/plplot5.10.0/examples/f95 (where $PREFIX is your install
prefix ~/Desktop/plplot_ifort) and ran the "make" command, then
you are using the traditional build system.  Some relevant questions
in that case are (1) are the module files installed properly, and
(2) if so, what pkg-config issue is preventing those from being
seen by the ifort compiler?

(1) Module files installed properly?

On my (Debian stable) system, the relevant Fortran module files are
installed at $PREFIX/lib/fortran/modules/plplot/. For my gfortran case
those files are called

software@raven> ls -l
total 576
-rw-r--r-- 1 software software 246325 Mar 26 12:12 plf95demolib.mod
-rw-r--r-- 1 software software 243851 Mar 26 12:12 plplot.mod
-rw-r--r-- 1 software software   3394 Mar 26 12:12 plplot_graphics.mod
-rw-r--r-- 1 software software   6764 Mar 26 12:12 plplot_str.mod
-rw-r--r-- 1 software software   2896 Mar 26 12:12 plplot_strutils.mod
-rw-r--r-- 1 software software   1318 Mar 26 12:12 plplot_types.mod
-rw-r--r-- 1 software software  72858 Mar 26 12:12 plplotp.mod

For ifort you should have equivalent files (probably with the same
names, but with very different content since ifort uses a different
module file format than gfortran) installed in that location for all
these modules.  In fact, if you look at your make.install.t file, the
following lines appear

-- Installing: /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plplot.mod
-- Installing: /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plplotp.mod
-- Installing: /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plplot_flt.mod
-- Installing: /Users/hiro/Desktop/plplot_ifort/lib/fortran/modules/plplot/plf95demolib.mod

I think you are likely fine, and the difference between the two
platforms is simply I am using the master tip version of PLplot (which
has more Fortran modules) while you are using the 5.10.0 version.
(By the way, PLplot-5.11.0 is coming soon based on the git master tip
version of PLplot.)

(2) pkg-config issue?

When you execute the "make" command above, that should automatically invoke
pkg-config to find the relevant compile options for the ifort case.

If you look carefully at the Makefile, the command it uses to find the
necessary compile flags to build one of the f95 examples is

PKG_CONFIG_PATH=$PREFIX/lib/pkg-config pkgconfig --cflags plplot-f95

When I run that command here with appropriate $PREFIX, here are the
results:

-I<PREFIX>/include/plplot -I<PREFIX>/lib/fortran/modules/plplot

where the <PREFIX> result should be the same as $PREFIX.

What are the results of that command there?

Note the reference above to -I<PREFIX>/lib/fortran/modules/plplot
which your results should have as well. That is the compiler option that
guarantees that gfortran (in my case) and ifort (in your case) finds
the relevant installed module files.

If all appears to be well with that option, then please send the
Makefile located at $PREFIX/plplot5.10.0/examples/f95/Makefile
as well as the results from

make x00f >& x00f.out

in that directory.

Alan
__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Re: [Plplot-general] plplot - mem/memcairo repeated render-to-memory?

[Hazen B. <hba...@ma...>]

>
> Hello all.
>
> After having tried quite a few different different options I feel I'm
> running out of ideas how to achive my goal, so I decided to ask around.
> Apologies if this is a trivial issue ...
>
> What I'm trying to do is get plplot to render a graph into a memory
> buffer, which is then blitted to a window whenver i) that window changes
> in size, ii) the window needs to be updated after being invalidated or
> iii) the data underlying the graph changes.
>
> A small test application I prepared (platform: Windows X64) worked well
> enough, I set the driver to "memcairo", passed a frame buffer memory
> block via plsmem(), issued the plotting commands and then dumped the
> buffer to inspect its contents. Right enough, it was rendered just the
> way I need it, ready to be displayed.
>
> But my attempts to incorporate that approach into the real application
> failed. I was planning to have one plstream per window, created and
> initialized upon window creation. A new memory buffer is passed in by
> plsmem() whenever the window size changes, to reflect the current
> drawing surface. In the window paint routine, I'm trying to simply
> render the plot, then blit the buffer.
>
> However that only works exactly once, all subsequent paint calls fail
> with an access violation in a pl...() command.
>
> The rendering commands were stripped down to
>
> plenv0(...)
> plline()
>
> But that doesn't seem to make any difference. Basically any call to
> plenv() seems to change something in an unwanted way, causing subsequent
> calls to write to invalid memory.
>
> Do I need to skip plenv(), and replace it with individual calls for
> drawing a box and setting a viewport?
>
> I played with pladv(), but to no avail.
>
> calling plinit(...) for each render cycle to set up a new enviroment
> seems to work, but that is so slow as to be unusable, so that's no option.
>
> Is there anything I'm doing wrong, or are there some examples I failed
> to notice? Help is appreciated.

Unfortunately, as you have noticed, you can't provide change the plot 
memory once you've specified it. You basically do need to start from 
scratch each time as the memory is passed to the plot driver when the 
driver is initialized. I'm surprised that this is too slow, as I've used 
a similar approach (extcairo driver), and it seemed to work ok for me.

One option might be to pass in an area as large as you think you will 
ever need, then just plot into and blit the sub-section of the area that 
matches the current window size?

-Hazen

Re: [Plplot-general] Build Plplot with ifort on OS X

[Arjen M. <Arj...@de...>]

Hello Hiro,

See my comments below.

From: Hiroyasu Yasuda [mailto:hi...@gs...]
Sent: Friday, March 27, 2015 11:34 AM
To: plplot_general
Subject: [Plplot-general] Build Plplot with ifort on OS X

Hello all

I've tried to build plplot with ifort on OS X as below follows:

export CC="gcc -O2"
export CXX="g++ -O2"
export FC="ifort -O2"

cmake -DCMAKE_INSTALL_PREFIX=~/Desktop/plplot_ifort -DENABLE_tcl=OFF -DENABLE_tk=OFF -DENABLE_python=OFF -DENABLE_qt=OFF -DENABLE_java=OFF -DENABLE_lua=OFF -DENABLE_wxwidgets=OFF ..

make

make install

Actually processes of build and install are seemingly success according to attached cmake.t, make.t and make.install.t. However when build example codes of f95 in share/plplot5.10.0/examples/f95, build process stopped with as below error messages:

x00f.f90(24): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [PLPLOT_STR]

If latest plplot support the ifort as fortran compilers, would you like to tell me a way to build plplot with ifort?

---
It appears that the module files (*.mod) are not located in a directory that the compiler is told about, at least when compiling the examples.

I have no access to OSX, so I cannot check where they are located, but could you search for them? You must have a file PLPLOT_STR.mod or something similar somewhere. I will check the output files you sent later. They may have the clue as to what is going on.

Regards,

Arjen

DISCLAIMER: This message is intended exclusively for the addressee(s) and may contain confidential and privileged information. If you are not the intended recipient please notify the sender immediately and destroy this message. Unauthorized use, disclosure or copying of this message is strictly prohibited. The foundation 'Stichting Deltares', which has its seat at Delft, The Netherlands, Commercial Registration Number 41146461, is not liable in any way whatsoever for consequences and/or damages resulting from the improper, incomplete and untimely dispatch, receipt and/or content of this e-mail.

[Plplot-general] Build Plplot with ifort on OS X

[Hiroyasu Y. <hi...@gs...>]

Hello all 

I’ve tried to build plplot with ifort on OS X as below follows:

export CC="gcc -O2"
export CXX="g++ -O2"
export FC=“ifort -O2"

cmake -DCMAKE_INSTALL_PREFIX=~/Desktop/plplot_ifort -DENABLE_tcl=OFF -DENABLE_tk=OFF -DENABLE_python=OFF -DENABLE_qt=OFF -DENABLE_java=OFF -DENABLE_lua=OFF -DENABLE_wxwidgets=OFF ..

make

make install

Actually processes of build and install are seemingly success according to attached cmake.t, make.t and make.install.t. However when build example codes of f95 in share/plplot5.10.0/examples/f95, build process stopped with as below error messages:

x00f.f90(24): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [PLPLOT_STR]

If latest plplot support the ifort as fortran compilers, would you like to tell me a way to build plplot with ifort?

Sincerely,
Hiro

.´￣`.
: (` :  
`. `´   :
  `・--´　

Re: [Plplot-general] plplot - mem/memcairo repeated render-to-memory?

[<he...@co...>]

Dietmar,<br><br>It seems strange that plinit is so slow. What you're doing in your application is very similar to what I'm doing in my windows app - main difference is that my repaints are due to zoom operations.  I do call plinit each repaint, and that all happens pretty fast - at least subjectively.  I'd say its less than 1/2 second to update the buffer and BLT to screen.  Maybe your perception that its slow is also considering the drawing time - which I would think would be constant even if you find a way to avoid calling plinit.  Having stepped through plinit for debugging purposes before, it seems pretty short.<br><br>Aaron<br><br>Sent from XFINITY Connect Mobile App<br>-----Original Message-----<br><br>From: po...@di...<br>To: plp...@li...<br>Cc: <br>Sent: 2015-03-26 13:19:31 GMT<br>Subject: [Plplot-general] plplot - mem/memcairo repeated render-to-memory?<br><br>Hello all.<br><br>After having tried quite a few different different options I feel I'm <br>running out of ideas how to achive my goal, so I decided to ask around. <br>Apologies if this is a trivial issue ...<br><br>What I'm trying to do is get plplot to render a graph into a memory <br>buffer, which is then blitted to a window whenver i) that window changes <br>in size, ii) the window needs to be updated after being invalidated or <br>iii) the data underlying the graph changes.<br><br>A small test application I prepared (platform: Windows X64) worked well <br>enough, I set the driver to "memcairo", passed a frame buffer memory <br>block via plsmem(), issued the plotting commands and then dumped the <br>buffer to inspect its contents. Right enough, it was rendered just the <br>way I need it, ready to be displayed.<br><br>But my attempts to incorporate that approach into the real application <br>failed. I was planning to have one plstream per window, created and <br>initialized upon window creation. A new memory buffer is passed in by <br>plsmem() whenever the window size changes, to reflect the current <br>drawing surface. In the window paint routine, I'm trying to simply <br>render the plot, then blit the buffer.<br><br>However that only works exactly once, all subsequent paint calls fail <br>with an access violation in a pl...() command.<br><br>The rendering commands were stripped down to<br><br>plenv0(...)<br>plline()<br><br>But that doesn't seem to make any difference. Basically any call to <br>plenv() seems to change something in an unwanted way, causing subsequent <br>calls to write to invalid memory.<br><br>Do I need to skip plenv(), and replace it with individual calls for <br>drawing a box and setting a viewport?<br><br>I played with pladv(), but to no avail.<br><br>calling plinit(...) for each render cycle to set up a new enviroment <br>seems to work, but that is so slow as to be unusable, so that's no option.<br><br>Is there anything I'm doing wrong, or are there some examples I failed <br>to notice? Help is appreciated.<br><br>TIA<br>best,<br>Dietmar<br><br><br>------------------------------------------------------------------------------<br>Dive into the World of Parallel Programming The Go Parallel Website, sponsored<br>by Intel and developed in partnership with Slashdot Media, is your hub for all<br>things parallel software development, from weekly thought leadership blogs to<br>news, videos, case studies, tutorials and more. Take a look and join the <br>conversation now. http://goparallel.sourceforge.net/<br>_______________________________________________<br>Plplot-general mailing list<br>Plp...@li...<br>https://lists.sourceforge.net/lists/listinfo/plplot-general<br>

[Plplot-general] plplot - mem/memcairo repeated render-to-memory?

[Dietmar N. <po...@di...>]

Hello all.

After having tried quite a few different different options I feel I'm 
running out of ideas how to achive my goal, so I decided to ask around. 
Apologies if this is a trivial issue ...

What I'm trying to do is get plplot to render a graph into a memory 
buffer, which is then blitted to a window whenver i) that window changes 
in size, ii) the window needs to be updated after being invalidated or 
iii) the data underlying the graph changes.

A small test application I prepared (platform: Windows X64) worked well 
enough, I set the driver to "memcairo", passed a frame buffer memory 
block via plsmem(), issued the plotting commands and then dumped the 
buffer to inspect its contents. Right enough, it was rendered just the 
way I need it, ready to be displayed.

But my attempts to incorporate that approach into the real application 
failed. I was planning to have one plstream per window, created and 
initialized upon window creation. A new memory buffer is passed in by 
plsmem() whenever the window size changes, to reflect the current 
drawing surface. In the window paint routine, I'm trying to simply 
render the plot, then blit the buffer.

However that only works exactly once, all subsequent paint calls fail 
with an access violation in a pl...() command.

The rendering commands were stripped down to

plenv0(...)
plline()

But that doesn't seem to make any difference. Basically any call to 
plenv() seems to change something in an unwanted way, causing subsequent 
calls to write to invalid memory.

Do I need to skip plenv(), and replace it with individual calls for 
drawing a box and setting a viewport?

I played with pladv(), but to no avail.

calling plinit(...) for each render cycle to set up a new enviroment 
seems to work, but that is so slow as to be unusable, so that's no option.

Is there anything I'm doing wrong, or are there some examples I failed 
to notice? Help is appreciated.

TIA
best,
Dietmar

Re: [Plplot-general] Am I going about this the best way?

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

On 2015-03-26 00:08-0500 Aaron Hexamer wrote:

> Thanks, Alan.  I'll check out Qt. Looks like it has a mem driver too.  I'll
> just have to try and see which dependencies are lightest.

P.S.  At this stage I would definitely recommend the latest version of
Qt4 rather than experimenting with Qt5.  The latter does mostly
work with our qt device driver and qt_example (using the experimental
-DPLPLOT_USE_QT5=ON option), but is just not nearly as reliable as
Qt4 in our experience.  I am sure Qt5 will eventually get there, but
it just needs a lot more time to mature properly.

Alan

__________________________
Alan W. Irwin

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

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________