plplot-general — PLplot user discussion group

Re: [Plplot-general] Another way to change the plotting font

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

On 2017-10-19 08:33+0100 p.d...@gm... wrote:

>
>
> Sent from my Windows 10 phone
>
> From: Alan W. Irwin
> Sent: 19 October 2017 03:01
> To: Frédéric
> Cc: plp...@li...
>
>> what do you mean by "set up fontconf"? isn't fontconfig just for
>> POSIX?
>
> No.  For example, the Windows version of GTK+ depends on fontconfig.
>
>
> We have had this discussion before. It is practically impossible for a mere mortal to build the gtk stack on Windows. I have tried and failed more than once. Gimp uses it, so it must be possible, but I have no idea how. Maybe using mingw?

For your information, there is at least two well-known native Windows
binary downloads for fontconfig, and also a variant
of fontconfig that works on Windows with the aid of the cygwin dll.

1. The GTK+ developers themselves produce a binary version of the
fontconfig package as part of their set of GTK+ packages for Windows.
That used to be the best source of binary packages for all GTK+
components, but I feel it has been superseded now by the next choice.

2. The MinGW-w64/MSYS2 distribution provides a binary package for
fontconfig.  Also, apparently anybody that has installed
MinGW-w64/MSYS2 can rebuild any of its packages (presumably including
fontconfig and the rest of the GTK+ stack) from source as well with no
muss/no fuss.  So my conclusion is that MinGW-w64/MSYS is Windows
users' best source of free software libraries right now, and I suggest
you would want to look into that possibility further rather than
attempting the difficult task of trying to build complex sets of free
software libraries such as GTK+ yourself from scratch.

3. Cygwin also provides binary packages for fontconfig, the rest of
the GTK+ stack, and many other free software libraries, but all those
free software libraries depend on the cygwin dll in order to work on
Windows so I am not as keen on the Cygwin distribution of free
software as I am on MinGW-w64/MSYS distribution of free software.

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] Another way to change the plotting font

[<p.d...@gm...>]

Sent from my Windows 10 phone

From: Alan W. Irwin
Sent: 19 October 2017 03:01
To: Frédéric
Cc: plp...@li...

> what do you mean by "set up fontconf"? isn't fontconfig just for
> POSIX?

No.  For example, the Windows version of GTK+ depends on fontconfig.


We have had this discussion before. It is practically impossible for a mere mortal to build the gtk stack on Windows. I have tried and failed more than once. Gimp uses it, so it must be possible, but I have no idea how. Maybe using mingw?

Re: [Plplot-general] Another way to change the plotting font

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

On 2017-10-17 13:27+0200 Frédéric wrote:

>> So getting back to your original request, if you are not satisfied
>> with how generic fonts look on your platform for -dev wxwidgets then
>> you might want to try to set up fontconfig a different way to favor
>> your preferred fonts.
>
> what do you mean by "set up fontconf"? isn't fontconfig just for
> POSIX?

No.  For example, the Windows version of GTK+ depends on fontconfig.

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] Another way to change the plotting font

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

On 2017-10-17 15:48+0100 p.d...@gm... wrote:

> Sorry,  thought I sent the message below, but I just found it in my drafts.
>
> wxWidgets accepts a font name, then uses the style for fallback if that named font doesn't exist. In the plplot wxWidgets backend we don't use the name because plplot doesn't include the ability to pass the name in as you have found.
>
> I personally would like to see PLplot follow a similar pattern – in fact I think this was one of my first feature requests a long time ago. As well as being generally useful, I think some journals have specific fonts that should be used. Unfortunately I’m not sure if other drivers are able to function in this way and an API change is often difficult as it requires updating all the supported languages.
>
> However, I relatively easily add this to the wxWidgets API only, as a method of the wxPLplotstream class.
>

> Alan would you support such a change? The call itself could accept a
pointer to a wxFont object as a wxWidgets native way to set the font.

Hi Phil:

Our long PLplot tradition is those who code get the last say.  So none
of us including me have any veto power here.

However, that said, I would advise you not to develop an additional
specific font feature for the wxwidgets device mainly because my
experiences with specific font approaches has been mostly bad.  For
one example of this, look at the "More on the reason for the missing
PDF glyphs" heading in doc/docbook/README.developers.  That was a case
where I was very careful to optimize the choice of specific fonts as
best I could.  But more typical is the user case where bad specific
fonts are chosen with many missing glyphs as a result whenever the
user attempts to use anything outside the ascii subset of UTF-8.  This
is one of the reasons I have deprecated the plfreetype approach (which
uses specific fonts and which is currently used by the wingcc device
and our old version of the wxwidgets-related code).  In sum, I just
feel pretty strongly that generic fonts (which are already working
well for the wxwidgets device driver) are a much easier approach for
us to support and users to use.

If despite this advice you do decide to go ahead and provide an
additional feature to our wxwidgets-related code that supports user
choice of specific fonts, I would not object further, but please
document that added feature (including the missing glyph problems that
might result).

Of course, regardless of your decision on this, I feel the most
important current issues for our wxwidgets related code are the
interactive problems and inefficiency issues I have documented
recently for the non-IPC3 case for -dev wxwidgets on the Linux
platform.  So your verifications and fixes for those issues (using
your access to debugging facilities on Linux platforms) would be much
appreciated.

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] Another way to change the plotting font

[Frédéric <ufo...@gm...>]

> However, I relatively easily add this to the wxWidgets API only, as a method
> of the wxPLplotstream class.

do you mean that you are able to use the wxWidgets fonts with plplot?
That's interesting.

I would like to use plplot to print the plots on the screen but also
in a pdf document with wxPdfDocument. And it is there where the font
is more important to me because the rest of the document is printed
with a single font so I would be happy if the plots could also use the
same font.

I set the font in the pdf with its name:
pdf.SetFont(font_name, font_style, size);

To register the custom fonts, I do the following:
auto font_manager = wxPdfFontManager::GetFontManager();
font_manager->AddSearchPath(font_path);
font_manager->SetDefaultEmbed(true);
font_manager->RegisterFontDirectory(font_path);
font_manager->RegisterFont("OpenSans-Regular.xml", font_name, 0);

It required me to build some special files OpenSans-Regular.{ctg.z,xml,z}:
http://wxcode.sourceforge.net/docs/wxpdfdoc/makefont.html

F

Re: [Plplot-general] Another way to change the plotting font

[<p.d...@gm...>]

Sorry,  thought I sent the message below, but I just found it in my drafts.

wxWidgets accepts a font name, then uses the style for fallback if that named font doesn't exist. In the plplot wxWidgets backend we don't use the name because plplot doesn't include the ability to pass the name in as you have found. 

I personally would like to see PLplot follow a similar pattern – in fact I think this was one of my first feature requests a long time ago. As well as being generally useful, I think some journals have specific fonts that should be used. Unfortunately I’m not sure if other drivers are able to function in this way and an API change is often difficult as it requires updating all the supported languages.

However, I relatively easily add this to the wxWidgets API only, as a method of the wxPLplotstream class.

Alan would you support such a change? The call itself could accept a pointer to a wxFont object as a wxWidgets native way to set the font.

Phil

From: Alan W. Irwin
Sent: 13 October 2017 23:59
To: Alaric Senat
Cc: plp...@li...
Subject: Re: [Plplot-general] Another way to change the plotting font

On 2017-10-13 15:26-0700 Alan W. Irwin wrote:

> On 2017-10-13 16:11+0200 Alaric Senat wrote:
>
>> Hi,
>> 
>> I'm trying to simply change the font of the texts in plplot by another one,
>> I found a lot of things in the doc' to change the font in the way to make 
>> it
>> Italic or Serif etc. But I'm wondering if there is no way to specify a font
>> file
>> like a .ttf, for example. It's good to know that I'm using Plplot with the
>> wxwidgets
>> Driver.
>> 
>>> From what I read in the documentation it seem's that i'm limited by the
>> 30 possibilities of plplot FCI system.
>
> Yes.  PLplot uses generic fonts.  I got that idea from the SVG
> graphics format, and all of the pango/cairo, Qt, and wxWidgets suites of
> graphical libraries.  See the first part of
> <http://docs.wxwidgets.org/3.1/overview_font.html> for the generic
> fonts used by wxwidgets.
>
> What generic fonts mean in practice is for each (unicode)
> glyph requested by the PLplot users via the UTF-8 encoded string to be
> plotted, the operating system supplies that glyph using the "best"
> system font of the particular generic kind (Sans, Serif, etc.)
> requested by the PLplot user. Where "best" is typically defined by how
> fontconfig is set up.
>
> The results are normally good.  For example, see
> <http://plplot.org/examples.php?demo=02> and click
> on the screenshots of the two pages of that example
> which show what the (PLplot default) sans serif, upright,
> normal weight font looks like.  (Those screenshots
> were generated by -dev pngcairo, but all cairo and
> qt results look like this for that example.)
>
> To check what wxwidgets does in this regard I
> ran
>
> examples/c/x02c -dev wxwidgets
>
> and the result is an italic font, i.e., not what
> is intended at all for that example!  So as a result
> of this discussion, I have found a bug in the font
> selection by -dev wxwidgets, and once that is sorted
> out, I suspect you will be much more satisfied with
> the generic font choice for -dev wxwidgets.
>
> To be continued by discussion in the plplot-devel mailing
> list, but once the bug is resolved, I will let you
> know here.

Oops.  Never mind.  There is no bug.  That example calls plfont(4)
which corresponds (for unicode fonts) to asking for a generic script
font which -dev wxwidgets supplies. In this case, the other device
drivers (e.g., qt, cairo) do not have generic script font capabilities
implemented (likely due to library limitations) so they fall back to
using generic sans for this case which obviously looks very different
than what is produced by -dev wxwidgets.

So getting back to your original request, if you are not satisfied
with how generic fonts look on your platform for -dev wxwidgets then
you might want to try to set up fontconfig a different way to favor
your preferred fonts.

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
__________________________

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Plplot-general mailing list
Plp...@li...
https://lists.sourceforge.net/lists/listinfo/plplot-general

Re: [Plplot-general] Another way to change the plotting font

[Frédéric <ufo...@gm...>]

> So getting back to your original request, if you are not satisfied
> with how generic fonts look on your platform for -dev wxwidgets then
> you might want to try to set up fontconfig a different way to favor
> your preferred fonts.

what do you mean by "set up fontconf"? isn't fontconfig just for
POSIX? what would then happens on Windows?
I am also interested to use my own truetype font with plplot on linux,
osx and windows with wxWidgets. Is there a way to do that
cross-platformly?

Kind regards,

F

Re: [Plplot-general] Another way to change the plotting font

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

On 2017-10-13 15:26-0700 Alan W. Irwin wrote:

> On 2017-10-13 16:11+0200 Alaric Senat wrote:
>
>> Hi,
>> 
>> I'm trying to simply change the font of the texts in plplot by another one,
>> I found a lot of things in the doc' to change the font in the way to make 
>> it
>> Italic or Serif etc. But I'm wondering if there is no way to specify a font
>> file
>> like a .ttf, for example. It's good to know that I'm using Plplot with the
>> wxwidgets
>> Driver.
>> 
>>> From what I read in the documentation it seem's that i'm limited by the
>> 30 possibilities of plplot FCI system.
>
> Yes.  PLplot uses generic fonts.  I got that idea from the SVG
> graphics format, and all of the pango/cairo, Qt, and wxWidgets suites of
> graphical libraries.  See the first part of
> <http://docs.wxwidgets.org/3.1/overview_font.html> for the generic
> fonts used by wxwidgets.
>
> What generic fonts mean in practice is for each (unicode)
> glyph requested by the PLplot users via the UTF-8 encoded string to be
> plotted, the operating system supplies that glyph using the "best"
> system font of the particular generic kind (Sans, Serif, etc.)
> requested by the PLplot user. Where "best" is typically defined by how
> fontconfig is set up.
>
> The results are normally good.  For example, see
> <http://plplot.org/examples.php?demo=02> and click
> on the screenshots of the two pages of that example
> which show what the (PLplot default) sans serif, upright,
> normal weight font looks like.  (Those screenshots
> were generated by -dev pngcairo, but all cairo and
> qt results look like this for that example.)
>
> To check what wxwidgets does in this regard I
> ran
>
> examples/c/x02c -dev wxwidgets
>
> and the result is an italic font, i.e., not what
> is intended at all for that example!  So as a result
> of this discussion, I have found a bug in the font
> selection by -dev wxwidgets, and once that is sorted
> out, I suspect you will be much more satisfied with
> the generic font choice for -dev wxwidgets.
>
> To be continued by discussion in the plplot-devel mailing
> list, but once the bug is resolved, I will let you
> know here.

Oops.  Never mind.  There is no bug.  That example calls plfont(4)
which corresponds (for unicode fonts) to asking for a generic script
font which -dev wxwidgets supplies. In this case, the other device
drivers (e.g., qt, cairo) do not have generic script font capabilities
implemented (likely due to library limitations) so they fall back to
using generic sans for this case which obviously looks very different
than what is produced by -dev wxwidgets.

So getting back to your original request, if you are not satisfied
with how generic fonts look on your platform for -dev wxwidgets then
you might want to try to set up fontconfig a different way to favor
your preferred fonts.

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] Another way to change the plotting font

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

On 2017-10-13 16:11+0200 Alaric Senat wrote:

> Hi,
>
> I'm trying to simply change the font of the texts in plplot by another one,
> I found a lot of things in the doc' to change the font in the way to make it
> Italic or Serif etc. But I'm wondering if there is no way to specify a font
> file
> like a .ttf, for example. It's good to know that I'm using Plplot with the
> wxwidgets
> Driver.
>
>> From what I read in the documentation it seem's that i'm limited by the
> 30 possibilities of plplot FCI system.

Yes.  PLplot uses generic fonts.  I got that idea from the SVG
graphics format, and all of the pango/cairo, Qt, and wxWidgets suites of
graphical libraries.  See the first part of
<http://docs.wxwidgets.org/3.1/overview_font.html> for the generic
fonts used by wxwidgets.

What generic fonts mean in practice is for each (unicode)
glyph requested by the PLplot users via the UTF-8 encoded string to be
plotted, the operating system supplies that glyph using the "best"
system font of the particular generic kind (Sans, Serif, etc.)
requested by the PLplot user. Where "best" is typically defined by how
fontconfig is set up.

The results are normally good.  For example, see
<http://plplot.org/examples.php?demo=02> and click
on the screenshots of the two pages of that example
which show what the (PLplot default) sans serif, upright,
normal weight font looks like.  (Those screenshots
were generated by -dev pngcairo, but all cairo and
qt results look like this for that example.)

To check what wxwidgets does in this regard I
ran

examples/c/x02c -dev wxwidgets

and the result is an italic font, i.e., not what
is intended at all for that example!  So as a result
of this discussion, I have found a bug in the font
selection by -dev wxwidgets, and once that is sorted
out, I suspect you will be much more satisfied with
the generic font choice for -dev wxwidgets.

To be continued by discussion in the plplot-devel mailing
list, but once the bug is resolved, I will let you
know here.

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
__________________________

[Plplot-general] Another way to change the plotting font

[Alaric S. <sen...@gm...>]

Hi,

I'm trying to simply change the font of the texts in plplot by another one,
I found a lot of things in the doc' to change the font in the way to make it
Italic or Serif etc. But I'm wondering if there is no way to specify a font
file
like a .ttf, for example. It's good to know that I'm using Plplot with the
wxwidgets
Driver.

>From what I read in the documentation it seem's that i'm limited by the
30 possibilities of plplot FCI system.

Regards,

Alaric

Re: [Plplot-general] setting PLPLOT_LIB by a function at the start of the program

[Tom S. <tom...@me...>]

Hi

I can confirm that Arjen is right as I use the putenv PLPLOT_LIB trick myself on macOS and Windows. 

On Linux this is generally not necessary assuming PLplot is installed using the distribution package manager. 

Best

Tom

On 12 Oct 2017, at 10:23, Frédéric <ufo...@gm...> wrote:

>> There is no need to worry about such interference: the environment is
>> private to the process and is inherited from whatever process started it.
> 
> Is this true on all OS (Windows, OSX and Linux)?
> If so, I just call putenv at the beginning of my program, that's simple.
> 
> Thanks,
> 
> F
> 
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Plplot-general mailing list
> Plp...@li...
> https://lists.sourceforge.net/lists/listinfo/plplot-general

Re: [Plplot-general] setting PLPLOT_LIB by a function at the start of the program

[Frédéric <ufo...@gm...>]

> I can confirm that Arjen is right as I use the putenv PLPLOT_LIB trick myself on macOS and Windows.

Apparently putenv is not standard but getenv is! putenv is POSIX.

However, I can find putenv in <cstdlib> on linux/windows/osx.
So that works.

Thanks,

F

Re: [Plplot-general] setting PLPLOT_LIB by a function at the start of the program

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

Hi Frédéric,

> -----Original Message-----
> From: Frédéric [mailto:ufo...@gm...]
> Sent: Thursday, October 12, 2017 10:24 AM
> > There is no need to worry about such interference: the environment is
> > private to the process and is inherited from whatever process started it.
>
> Is this true on all OS (Windows, OSX and Linux)?
> If so, I just call putenv at the beginning of my program, that's simple.
>
I know it is true for Linux and OSX (as UNIX descendants). On Windows the situation may be less straightforward if you include system-defined environment variables, but I have been forced to restart some program that would start other programs more than once, because the changed environment was not automatically copied into the child processes. If somebody else has evidence to the converse, I will stand corrected :).

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] setting PLPLOT_LIB by a function at the start of the program

[Frédéric <ufo...@gm...>]

> There is no need to worry about such interference: the environment is
> private to the process and is inherited from whatever process started it.

Is this true on all OS (Windows, OSX and Linux)?
If so, I just call putenv at the beginning of my program, that's simple.

Thanks,

F

Re: [Plplot-general] setting PLPLOT_LIB by a function at the start of the program

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

Hi Frédéric,

> -----Original Message-----
> From: Frédéric [mailto:ufo...@gm...]
> Sent: Thursday, October 12, 2017 10:10 AM
>
> Yes, I could do that (I am using C++). But does this interacts with the OS then?
> If so, I do not like it. Imagine I have 2 independant programs that run on the same
> system and I want each of them to point to a different directory. There will be a
> conflict.
>
> This is why I would prefer to set the directory with a function so that it write to an
> internal variable of the library.
>
> I can probably implement that and propose a patch if you tell me where it is best to
> start.
>
There is no need to worry about such interference: the environment is private to the process and is inherited from whatever process started it.

Such a function would need to be called before any PLplot function, a possible set-up is (just a sketch):

void plinit_with_path( char *path ) {

               static char full_string[500];

               sprintf( full_string, "%s%s", "PLPLOT_LIB=", path );

               putenv( full_string );

               plinit();

}

Or put it in plparseopts(). Perhaps that is even better.

Mind you, any new API function will have to be made available in all supported languages.

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] setting PLPLOT_LIB by a function at the start of the program

[Frédéric <ufo...@gm...>]

> Sure that would be a possibility, but you could write it yourself ;). You
> don’t mention the programming language you’re using, but it should not be
> too difficult – all that is required is that you call the putenv() C
> function to set the environment variable from within your program to a value
> indicating that directory. Possibly some more work is needed, as that
> directory is likely to be relative to the location of your program.

Yes, I could do that (I am using C++). But does this interacts with the OS then?
If so, I do not like it. Imagine I have 2 independant programs that
run on the same system and I want each of them to point to a different
directory. There will be a conflict.

This is why I would prefer to set the directory with a function so
that it write to an internal variable of the library.

I can probably implement that and propose a patch if you tell me where
it is best to start.

F

Re: [Plplot-general] setting PLPLOT_LIB by a function at the start of the program

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

Hi Frédéric,



Sure that would be a possibility, but you could write it yourself ;). You don't mention the programming language you're using, but it should not be too difficult - all that is required is that you call the putenv() C function to set the environment variable from within your program to a value indicating that directory. Possibly some more work is needed, as that directory is likely to be relative to the location of your program.



Regards,



Arjen



> -----Original Message-----
> From: Frédéric [mailto:ufo...@gm...]
> Sent: Thursday, October 12, 2017 9:24 AM
> To: plp...@li...
> Subject: [Plplot-general] setting PLPLOT_LIB by a function at the start of the
> program
>
> Hi,
>
> I am studying the possibility to use plplot in my project.
> I am using static linking and would like to store the font files in a specific sub-
> directory of my program directory.
> I do not like to rely on an environment variable (PLPLOT_LIB*) to point to this
> directory (anybody could remove it thinking it should not be there) and would prefer
> to call an initialize() function at the start of the program where I provide the path to
> the lib directory.
>
> Could we imagine adding this to the library?
>
> Thanks,
>
> F
>
> * http://plplot.sourceforge.net/docbook-manual/plplot-html-5.13.0/deploying.html
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most engaging tech
> sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> 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.

[Plplot-general] setting PLPLOT_LIB by a function at the start of the program

[Frédéric <ufo...@gm...>]

Hi,

I am studying the possibility to use plplot in my project.
I am using static linking and would like to store the font files in a
specific sub-directory of my program directory.
I do not like to rely on an environment variable (PLPLOT_LIB*) to
point to this directory (anybody could remove it thinking it should
not be there) and would prefer to call an initialize() function at the
start of the program where I provide the path to the lib directory.

Could we imagine adding this to the library?

Thanks,

F

* http://plplot.sourceforge.net/docbook-manual/plplot-html-5.13.0/deploying.html

[Plplot-general] trycompile issue with clang

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

On 2017-10-05 18:06+0200 Alaric Senat wrote:

[...]
> Today I also had another issue while attempting to specify an Osx version
> of clang to cmake, it failed to compile a test C program
> whith the osx clang version whereas I can compile mannually my own C files
> with it.
> You can find the CmakeError.log in attachment.
> It seems that the ar comand fail but I don't really get why and all the
> files in CmakeTmp are destroyed after the execution of Cmake.
> So have you an idea about where I can find this compilation test to try it
> manually to find out what's wrong ?

Hi Alaric:

I have changed the subject line to something more appropriate for this
topic.  Also, I don't mind answering occasional general cmake
questions like this one on the plplot-general mailing list for the
benefit of PLplot users, but a better place to ask such general CMake
questions is the cmake mailing list which is typically quite helpful
and which naturally brings to the attention of the CMake developers
any CMake bug you might have found.

To track down such trycompile issues like you have described I suggest
you use the latest CMake version (which is 3.9.4) (in case the
trycompile issue you have discovered is due to a CMake bug that has
been fixed in that latest version).  Also, you should use the simplest
possible CMake project which illustrates the trycompile issue, e.g., a
source directory containing just the following CMakeLists.txt file:

cmake_minimum_required(VERSION 3.9.4 FATAL_ERROR)
project(test_c_compiler C)

I haven't tried that simple project myself, but this project
should attempt a trycompile for your C compiler and that
is about all.  You should be able to
test that the above two-lines of CMake logic works in general by simply trying the gcc
compiler.  And if that test succeeds. then trying it with the clang
compiler will likely fail according to the information you have
supplied above.

Once, you have gone through all of that (gcc success + clang failure
for such a simple project), then you can debug that clang failure by
using, e.g., the cmake option --debug-trycompile. That option is only
useful for one trycompile at a time, but that is exactly what is done
in the above simple project.  Assuming this advice lets you see the
exact issue, I cannot help you further with that exact issue (I have no
experience with clang or Mac OS X), but those on the cmake mailing
list probably can help especially when you are presenting them with
such a simple test case that demonstrates 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] Cmake Cross-compilation issue

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

On 2017-10-05 18:06+0200 Alaric Senat wrote:

> Hi Alan,
>
> Thank you for answering me as quickly !
> I finally found out the problem thanks to your answer, it was indeed that
> my CMAKE_NATIVE_BINARY_DIR did not
> specify a built binary directory (I didn't unserstand it needed to be
> built, it was quite simple actually).
> So it cross-compile well for windows, I just needed to modify a bit some
> wxWidgets macro in the source code
> probably because of my compilator.
>
> Also, the -DNaNAwareCCompiler option is totaly ignored by Cmake on my
> computer.

Hi Alaric:

It was good to hear that you got cross-compilation to work following our old
wiki item at
<https://sourceforge.net/p/plplot/wiki/Building_PLplot_with_a_cross-compiler/>.
Is there anything in that wiki item you think we should update?

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] Cmake Cross-compilation issue

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

On 2017-10-04 18:12+0200 Alaric Senat wrote:

> Hello,
>
> I'm trying to compile plplot with mingw64 on linux and I've got a strange
> issue since this morning.
> Cmake is trying to include non-generated files

[....]
>
> *if(NOT CMAKE_CROSSCOMPILING)  add_executable(tai-utc-gen
> ${CMAKE_CURRENT_SOURCE_DIR}/tai-utc-gen.c)  if(MATH_LIB)
> target_link_libraries(tai-utc-gen ${MATH_LIB})  endif(MATH_LIB)
> add_executable(deltaT-gen ${CMAKE_CURRENT_SOURCE_DIR}/deltaT-gen.c
> ${CMAKE_CURRENT_SOURCE_DIR}/dspline.c)  export(TARGETS tai-utc-gen FILE
> ${CMAKE_CURRENT_BINARY_DIR}/tai-utc-gen.cmake )  export(TARGETS deltaT-gen
> FILE ${CMAKE_CURRENT_BINARY_DIR}/deltaT-gen.cmake )else(NOT
> CMAKE_CROSSCOMPILING)  # Generated by initial non-cross build above.
> include(${CMAKE_NATIVE_BINARY_DIR}/lib/qsastime/tai-utc-gen.cmake)
> include(${CMAKE_NATIVE_BINARY_DIR}/lib/qsastime/deltaT-gen.cmake)endif(NOT
> CMAKE_CROSSCOMPILING)*
> So, I don't really understand how the files included in the else statement
> are supposed to be generated when I'm cross-compiling. I tried to generate
> them by removing the condition  but of course the compilation of doesn't
> pass in the makefile !
>
> Sorry if my problem is a kind of weird or obvious to resolve and thank's a
> lot for reading !

Hi Alaric:

Thanks for your interest in cross-compiling PLplot.
Note our cross-compilation did work back in 2009 (see
<https://sourceforge.net/p/plplot/wiki/Building_PLplot_with_a_cross-compiler/>),
but as far as I know there has been little testing of that capability
since.  For example, I never did learn how to cross-compile just to
test that capability continues to work although that wiki item states
cross-compilation of PLplot should be straightforward on my (Linux)
platform.

So it is certainly possible some errors have crept in since 2009, but (with your
cooperation in reporting issues that you find) I will make every
effort to resolve those since cross-compilation is a PLplot capability
I would like to preserve.  And once everything works for you, I would
appreciate your review of the above wiki item in (the likely) case we
need to update it.

So the first question for you is did you follow those wiki
instructions, which for example, tell you how to set up the
all-important CMAKE_NATIVE_BINARY_DIR variable which helps to access
selected results from your native build (as in the CMake logic above)
to be used in your subsequent cross-platform build.

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
__________________________

[Plplot-general] Cmake Cross-compilation issue

[Alaric S. <sen...@gm...>]

Hello,

I'm trying to compile plplot with mingw64 on linux and I've got a strange
issue since this morning.
Cmake is trying to include non-generated files, here is the error:











*CMake Error at lib/qsastime/CMakeLists.txt:53 (include):  include could
not find load file:
/home/alaric/plplot-5.13.0//lib/qsastime/tai-utc-gen.cmakeCMake Error at
lib/qsastime/CMakeLists.txt:54 (include):  include could not find load
file:    /home/alaric/plplot-5.13.0//lib/qsastime/deltaT-gen.cmakeCMake
Error at include/CMakeLists.txt:45 (INCLUDE):  INCLUDE could not find load
file:    /home/alaric/plplot-5.13.0//include/ImportExecutables.cmake*

When I'm looking at 'lib/qsastime/CMakeLists.txt' this block appears:

# Routine for generating a header containing data to be used by the
# TAI-UTC transformation code.














*if(NOT CMAKE_CROSSCOMPILING)  add_executable(tai-utc-gen
${CMAKE_CURRENT_SOURCE_DIR}/tai-utc-gen.c)  if(MATH_LIB)
target_link_libraries(tai-utc-gen ${MATH_LIB})  endif(MATH_LIB)
add_executable(deltaT-gen ${CMAKE_CURRENT_SOURCE_DIR}/deltaT-gen.c
${CMAKE_CURRENT_SOURCE_DIR}/dspline.c)  export(TARGETS tai-utc-gen FILE
${CMAKE_CURRENT_BINARY_DIR}/tai-utc-gen.cmake )  export(TARGETS deltaT-gen
FILE ${CMAKE_CURRENT_BINARY_DIR}/deltaT-gen.cmake )else(NOT
CMAKE_CROSSCOMPILING)  # Generated by initial non-cross build above.
include(${CMAKE_NATIVE_BINARY_DIR}/lib/qsastime/tai-utc-gen.cmake)
include(${CMAKE_NATIVE_BINARY_DIR}/lib/qsastime/deltaT-gen.cmake)endif(NOT
CMAKE_CROSSCOMPILING)*
So, I don't really understand how the files included in the else statement
are supposed to be generated when I'm cross-compiling. I tried to generate
them by removing the condition  but of course the compilation of doesn't
pass in the makefile !

Sorry if my problem is a kind of weird or obvious to resolve and thank's a
lot for reading !
Sincerely,
Alaric SENAT

Re: [Plplot-general] Win installation

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

On 2017-09-27 23:47+1300 Cao Murrano wrote:

> Ok, cleaning the directory did help. All three commands (
> cmake -DCMAKE_INSTALL_PREFIX=install ../;
> make; make install) completed with no error, and dll's are in place.
>
> Now I need to do some testing with Ada code, but the compilation issue is
> closed.

Hi Cao:

Just to interject here, I was glad to hear your build issues were
solved by making a clean start.  However, could you double-check your
configure and build steps for the Ada-related PLplot components
because as far as I know, nobody has ever gotten our Ada bindings to
configure properly (much less build properly) on any Windows platform
before.

For example, our records of Arjen's last Cygwin comprehensive test at
<https://sourceforge.net/p/plplot/wiki/Testing_PLplot/#Testing
Reports> had the following note whose Ada-related part is

(G1) Ada automatically disabled because gnat library not found

Similarly, his last comprehensive test on the "MinGW-w64/MSYS2" platform
had the following note whose Ada-related part is

(G3) gnatmake cannot compile simple Ada test programme

So could you capture the complete cmake output from a fresh build
start and send it to us (one easy way to do that is

cmake <options including -DBUILD_TEST=ON> <path to top-level directory of source tree> >& cmake.out

) to confirm there are no Ada-related
configuration issues for your particular Cygwin installation?

And if that cmake.out file shows the net result of your configuration is

ENABLE_ada:             ON

i.e., there are no Ada-related configuration issues, then could you go
on and run-time test our Ada examples by putting the build-tree dll
subdirectory on your PATH and then running

make VERBOSE=1 test_ada_psc >& test_ada_psc.out

?

N.B. the above -DBUILD_TEST=ON cmake option enables general run-time
testing (including the above test_ada_psc target which builds our Ada
binding, builds each of our Ada standard examples, and then runs each
of those built examples using -dev psc). The VERBOSE=1 option on the
make command gives many additional details (e.g., compile, link, and
gnatmake options) concerning each of the build and run-time steps to
let us know if our build system got any of those details wrong.

Note my Ada expertise is pretty limited.  So when I wrote the
unofficial CMake-based language support files for Ada that are
included with PLplot I had to rely on following pretty closely how
other languages (such as C++ and Fortran) are officially supported by
CMake.  The result works well on Linux so my Ada language support is
likely to be close to working well on Cygwin. However, I might have
gotten an essential compile, link, or gnatmake option wrong for that
case so if you have any Ada expertise yourself, I will have to rely on
you to help me look at the above VERBOSE=1 make results to figure out what
changes (if any) in those options need to be done.

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] Fwd: PLplot capabilities

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

On 2017-09-27 17:34+0900 Don Warren wrote:

> This is either a feature request or "RTFM"-bait.  I'm trying to figure out
> if PLplot can do the following two things:
> 
> (1) Logarithmic ticks on an axis, but only labeling every Nth decade (when
> your Y axis goes from 10^-60 to 10^-20, you don't want every single decade
> to have a number next to it!)
> 
> (2) Moving tick mark numbers towards/away from the axis they're associated
> with.  It looks like PLplot can do this with the axis labels, using the
> plmtex command.  Is there no way to do this with the numbers associated
> with major tick marks?
> 
> I know both of these could be done by hand with sufficient usage of the
> plmtex command (or rather, I know how to do it in PGPLOT, which I'm
> currently stuck using for legacy reasons).  But then I am individually
> labeling each major tick on the axis, and placing them all independently.
> I would rather not have 20 extra "annotation" blocks in my plotting scripts
> if it's avoidable, not to mention the time spent making sure each number is
> correctly positioned.

Hi Don:

This is my second attempt to answer you since the first attempt sent
earlier today did not appear to be distributed to the mailing list....

Sorry we didn't respond the first time you asked your questions, but I
will attempt to do that now.

Note, I give you a number of plplot.org references below, but you will
have to wait to access those because of a general Sourceforge outage
that started today for the many websites they supply for projects
(like ours) that they host.  Apparently they are working hard to
fix whatever the issue is, but there is no ETA about when SF-related
websites like ours will be available again.

With regard to your question (1), PLplot does have detailed log axis
labelling (the plbox "l" option) to do that, but that is only
suitable for data whose range is a couple of decades (see
<http://plplot.org/examples.php?demo=04> for an example of
that capability) where you want to label unequally spaced logarithmic
subticks between the decades, corresponding to 1, 2, 3, etc., times
some power of 10.  So for your huge range in data, instead of using
that option, I suggest you simply take the log_10 of your data and use
ordinary labelling of those equally spaced major and minor tickmarks
where you have the choice of both the major and minor equally spaced
tick intervals.  See
<http://plplot.org/docbook-manual/plplot-html-5.12.0/plbox.html>
for more details.

With regard to your question (2), I don't think PLplot currently has
this capability.  However, our users appear to be satisfied with the
current positioning of numerical axis labels, and that might turn out
to be true for you as well once you try PLplot. So to evaluate the spacing we 
provide
(even before you try PLplot) I suggest you look carefully at the
tick-mark labelling results for our many standard examples at
<http://plplot.org/examples.php>.

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] Win installation

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

Hi Cao,



Ah, good to hear that! Let me know what you find - I can then update my installation and we could check the Ada binding on Cygwin.



In general it is best to start with a clean slate, but I was not certain it could have this kind of effects ;).



Regards,



Arjen

From: Cao Murrano [mailto:mu...@gm...]
Sent: Wednesday, September 27, 2017 12:47 PM
To: Arjen Markus
Cc: plp...@li...
Subject: Re: [Plplot-general] Win installation

Ok, cleaning the directory did help. All three commands (cmake -DCMAKE_INSTALL_PREFIX=install ../;
make; make install) completed with no error, and dll's are in place.
Now I need to do some testing with Ada code, but the compilation issue is closed.
Thanks again.
Cao

On Wed, Sep 27, 2017 at 9:57 PM, Arjen Markus <Arj...@de...<mailto:Arj...@de...>> wrote:

Hi Cao,



Wrt Ada library: that seems to require some fine-tuning on my system, but I do not use Ada myself, so I have not got much experience with that :).



The CMake problem you report is rather odd. CMake tries to find out all manner of things via small test programs, but the messages indicate that that is failing.



Do you start in a clean directory? Should not really matter for this problem, but perhaps it does. I have no ready solution for this, so I am merely searching for a possible workaround/solution.



Regards,



Arjen




From: Cao Murrano [mailto:mu...@gm...<mailto:mu...@gm...>]
Sent: Wednesday, September 27, 2017 10:48 AM
To: Arjen Markus
Cc: plp...@li...<mailto:plp...@li...>
Subject: Re: [Plplot-general] Win installation

Thanks Arjen,
Right, I didn't have the "make". Now I do, and it goes somewhere, but still throws a fatal error:

----------------------------------------------------------------------------------------------------------------------
Determining if the C compiler works failed with the following output:
Change Dir: /cygdrive/d/ada/plplot/buildcygwin/CMakeFiles/CMakeTmp

Run Build Command:"/usr/bin/make.exe" "cmTC_76d2f/fast"
/usr/bin/make -f CMakeFiles/cmTC_76d2f.dir/build.make CMakeFiles/cmTC_76d2f.dir/build
make[1]: Entering directory '/cygdrive/d/ada/plplot/buildcygwin/CMakeFiles/CMakeTmp'
Building C object CMakeFiles/cmTC_76d2f.dir/testCCompiler.c.o
/cygdrive/c/GNAT/msys64/mingw64/bin/cc.exe     -o CMakeFiles/cmTC_76d2f.dir/testCCompiler.c.o   -c /cygdrive/d/ada/plplot/buildcygwin/CMakeFiles/CMakeTmp/testCCompiler.c
cc.exe: error: /cygdrive/d/ada/plplot/buildcygwin/CMakeFiles/CMakeTmp/testCCompiler.c: No such file or directory
cc.exe: fatal error: no input files
compilation terminated.
make[1]: *** [CMakeFiles/cmTC_76d2f.dir/build.make:66: CMakeFiles/cmTC_76d2f.dir/testCCompiler.c.o] Error 1
make[1]: Leaving directory '/cygdrive/d/ada/plplot/buildcygwin/CMakeFiles/CMakeTmp'
make: *** [Makefile:126: cmTC_76d2f/fast] Error 2
-------------------------------------------------------------------------------------------------------------------------
I'm not so worried about the ada bindings. These normally compile ok under GNAT once the library is in place.
Cao

On Tue, Sep 26, 2017 at 10:00 PM, Arjen Markus <Arj...@de...<mailto:Arj...@de...>> wrote:

Hi Cao,



Have you installed the make program? In the Cygwin setup program that would be in the category Devel, "make: The GNU version of the 'make' utility" (version 4.2.1-2). You may have to specify more information to build the Ada binding - the GNU Ada compiler on Cygwin has presented us with sufficient problems for me to be a bit wary. I could not find the latest status in my notes just now.



Regards,



Arjen



From: Cao Murrano [mailto:mu...@gm...<mailto:mu...@gm...>]
Sent: Tuesday, September 26, 2017 10:51 AM
To: plp...@li...<mailto:plp...@li...>
Subject: [Plplot-general] Win installation

Hello,
Is it possible to have some more thorough instructions for Win compilation?
I was following the posted instructions, and here' s what I'm getting:

D:\ada\plplot\buildcygwin>cmake -DCMAKE_INSTALL_PREFIX=install ../
CMake Error: CMake was unable to find a build program corresponding to "Unix Makefiles".  CMAKE_MAKE_PROGRAM is not set.  You probably need
to select a different build tool.
-- Configuring incomplete, errors occurred!
My system is Win 10/Cygwin/GNAT 2016 (Ada).
Thanks in advance.
Cao
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.

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.

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.