plplot-devel

[Plplot-devel] Using Window's raw API for shapes and text

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

Hi Aaron:

I am putting this further discussion on the plplot-devel mailing list
since some developers there have more device expertise than I do (and
also a lot more Windows expertise) and therefore will likely want to
contribute to the discussion.

On 2015-05-13 07:37-0500 Aaron Hexamer wrote:

> [....] I've been wondering is if it
> would be feasible to make a driver that uses more of Window's raw API for
> shapes and text, thus avoiding the need for libs like Qt, Cairo-Pango, etc.
> - one would still get portability through other drivers on other platforms.
> It looks like wingcc is somewhat like that, but preferred to use Freetype.
> I gather from looking at the history that text rendering is challenging,
> especially around vertical alignment topics.

For your information, the plfreetype approach is strongly deprecated. 
The primary reason for this is the user can only control the selection
of the needed font files at cmake time or else at run-time via
environment variables which is a very clumsy approach.  A secondary
reason is the plfreetype approach only works for simple text layout
(left-to-right) languages.  Instead, our preferred approach is to rely
on external libraries such as pango/cairo/fontconfig, Qt, or wxwidgets
to automatically select the best system font (of the sans, serif,
normal weight, bold, etc., generic classes of fonts that PLplot
supports) to render each unicode glyph encountered (which
automatically allows multi-language plots such as example 24) and to
do the required (complex) text layout.

Thus, I agree that the plfreetype approach used by wingcc (the last
device driver that still uses that approach) should be completely
replaced by calling the appropriate native Windows API for selecting
the best system font to render each unicode glyph that is encountered
and to do the text layout.

I think in the past that Arjen Markus has commented this approach
should be possible, but he has not had time to pursue it further.
Also, my understanding is that Jim Dishaw has worked on implementing a
Windows device driver.  I am not sure what the status of that project
is, but he might have some comments also about the feasibility of
modifying wingcc this way.

Anyhow, if you feel such modification of wingcc is possible, I would
encourage you to give it a try following, say, the broad outline of
the alt_unicode text handling that is done in drivers/cairo.c.

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-devel] Using Window's raw API for shapes and text

[Jim D. <ji...@di...>]

I still have the revised windows driver that I can blow the dust off and offer as a possible replacement. 



> On May 13, 2015, at 1:59 PM, Alan W. Irwin <ir...@be...> wrote:
> 
> Hi Aaron:
> 
> I am putting this further discussion on the plplot-devel mailing list
> since some developers there have more device expertise than I do (and
> also a lot more Windows expertise) and therefore will likely want to
> contribute to the discussion.
> 
>> On 2015-05-13 07:37-0500 Aaron Hexamer wrote:
>> 
>> [....] I've been wondering is if it
>> would be feasible to make a driver that uses more of Window's raw API for
>> shapes and text, thus avoiding the need for libs like Qt, Cairo-Pango, etc.
>> - one would still get portability through other drivers on other platforms.
>> It looks like wingcc is somewhat like that, but preferred to use Freetype.
>> I gather from looking at the history that text rendering is challenging,
>> especially around vertical alignment topics.
> 
> For your information, the plfreetype approach is strongly deprecated. 
> The primary reason for this is the user can only control the selection
> of the needed font files at cmake time or else at run-time via
> environment variables which is a very clumsy approach.  A secondary
> reason is the plfreetype approach only works for simple text layout
> (left-to-right) languages.  Instead, our preferred approach is to rely
> on external libraries such as pango/cairo/fontconfig, Qt, or wxwidgets
> to automatically select the best system font (of the sans, serif,
> normal weight, bold, etc., generic classes of fonts that PLplot
> supports) to render each unicode glyph encountered (which
> automatically allows multi-language plots such as example 24) and to
> do the required (complex) text layout.
> 
> Thus, I agree that the plfreetype approach used by wingcc (the last
> device driver that still uses that approach) should be completely
> replaced by calling the appropriate native Windows API for selecting
> the best system font to render each unicode glyph that is encountered
> and to do the text layout.
> 
> I think in the past that Arjen Markus has commented this approach
> should be possible, but he has not had time to pursue it further.
> Also, my understanding is that Jim Dishaw has worked on implementing a
> Windows device driver.  I am not sure what the status of that project
> is, but he might have some comments also about the feasibility of
> modifying wingcc this way.
> 
> Anyhow, if you feel such modification of wingcc is possible, I would
> encourage you to give it a try following, say, the broad outline of
> the alt_unicode text handling that is done in drivers/cairo.c.
> 
> 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
> __________________________
> 
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud 
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Plplot-devel mailing list
> Plp...@li...
> https://lists.sourceforge.net/lists/listinfo/plplot-devel

Re: [Plplot-devel] Using Window's raw API for shapes and text

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

On 2015-05-13 14:03-0400 Jim Dishaw wrote:

> I still have the revised windows driver that I can blow the dust off and offer as a possible replacement.

Hi Jim:

As I recall, you had a lot of interesting ideas for your Windows
device driver so I hope you do integrate that into PLplot in the near
future regardless of whether or not it uses the deprecated plfreetype
approach to render text. Of course, it would be great if it uses a
native Windows unicode API to render text using whatever Windows
system fonts happen to be installed since that approach could
be straightforwardly copied over to wingcc to get rid of the
last use of plfreetype.

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-devel] Using Window's raw API for shapes and text

[Jim D. <ji...@di...>]

> On May 13, 2015, at 6:06 PM, Alan W. Irwin <ir...@be...> wrote:
> 
>> On 2015-05-13 14:03-0400 Jim Dishaw wrote:
>> 
>> I still have the revised windows driver that I can blow the dust off and offer as a possible replacement.
> 
> Hi Jim:
> 
> As I recall, you had a lot of interesting ideas for your Windows
> device driver so I hope you do integrate that into PLplot in the near
> future regardless of whether or not it uses the deprecated plfreetype
> approach to render text. Of course, it would be great if it uses a
> native Windows unicode API to render text using whatever Windows
> system fonts happen to be installed since that approach could
> be straightforwardly copied over to wingcc to get rid of the
> last use of plfreetype.
> 
> Alan

I agree. I think another thing to look at for version 6 is to have a consistent menubar and keybindings for the GUI drivers. 

> 
> __________________________
> 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-devel] Using Window's raw API for shapes and text

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

Alan,

I'm a little confused by the statement about copying code from Jim's driver
to wingcc.  I just assumed that wingcc was named as such because maybe
PLplot was first ported to windows with support only for the gcc compiler?
I.e. is there really anything specific to gcc about wingcc anymore.  I
recall using it with MSVC when I first got started with PLplot. If Jim's
driver did everything that wingcc does, but without plfreetype, would there
be a need for both?

Aaron.

-----Original Message-----
From: Alan W. Irwin [mailto:ir...@be...] 
Sent: Wednesday, May 13, 2015 5:06 PM
To: Jim Dishaw
Cc: Aaron Hexamer; PLplot development list
Subject: Re: [Plplot-devel] Using Window's raw API for shapes and text

On 2015-05-13 14:03-0400 Jim Dishaw wrote:

> I still have the revised windows driver that I can blow the dust off and
offer as a possible replacement.

Hi Jim:

As I recall, you had a lot of interesting ideas for your Windows device
driver so I hope you do integrate that into PLplot in the near future
regardless of whether or not it uses the deprecated plfreetype approach to
render text. Of course, it would be great if it uses a native Windows
unicode API to render text using whatever Windows system fonts happen to be
installed since that approach could be straightforwardly copied over to
wingcc to get rid of the last use of plfreetype.

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-devel] Using Window's raw API for shapes and text

[Jim D. <ji...@di...>]

> On May 13, 2015, at 6:17 PM, Aaron Hexamer <he...@co...> wrote:
> 
> Alan,
> 
> I'm a little confused by the statement about copying code from Jim's driver
> to wingcc.  I just assumed that wingcc was named as such because maybe
> PLplot was first ported to windows with support only for the gcc compiler?
> I.e. is there really anything specific to gcc about wingcc anymore.  I
> recall using it with MSVC when I first got started with PLplot. If Jim's
> driver did everything that wingcc does, but without plfreetype, would there
> be a need for both?
> 
> Aaron.
> 

I think the smart thing to do is to have one Windows driver. I can't recall any differences that could not be handled without an #ifdef block. 

> -----Original Message-----
> From: Alan W. Irwin [mailto:ir...@be...] 
> Sent: Wednesday, May 13, 2015 5:06 PM
> To: Jim Dishaw
> Cc: Aaron Hexamer; PLplot development list
> Subject: Re: [Plplot-devel] Using Window's raw API for shapes and text
> 
>> On 2015-05-13 14:03-0400 Jim Dishaw wrote:
>> 
>> I still have the revised windows driver that I can blow the dust off and
> offer as a possible replacement.
> 
> Hi Jim:
> 
> As I recall, you had a lot of interesting ideas for your Windows device
> driver so I hope you do integrate that into PLplot in the near future
> regardless of whether or not it uses the deprecated plfreetype approach to
> render text. Of course, it would be great if it uses a native Windows
> unicode API to render text using whatever Windows system fonts happen to be
> installed since that approach could be straightforwardly copied over to
> wingcc to get rid of the last use of plfreetype.
> 
> 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-devel] Using Window's raw API for shapes and text

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

On 2015-05-13 18:27-0400 Jim Dishaw wrote:

>
>> On May 13, 2015, at 6:17 PM, Aaron Hexamer <he...@co...> wrote:
>>
>> Alan,
>>
>> I'm a little confused by the statement about copying code from Jim's driver
>> to wingcc.  I just assumed that wingcc was named as such because maybe
>> PLplot was first ported to windows with support only for the gcc compiler?
>> I.e. is there really anything specific to gcc about wingcc anymore.  I
>> recall using it with MSVC when I first got started with PLplot. If Jim's
>> driver did everything that wingcc does, but without plfreetype, would there
>> be a need for both?
>>
>> Aaron.
>>
>

> I think the smart thing to do is to have one Windows driver. I can't
recall any differences that could not be handled without an #ifdef
block.

Hi Jim:

Any way you decide to implement your driver (i.e., separately or as
part of wingcc) is fine with me.

@Aaron:

Your question about the wingcc name got me curious so I looked at git
log results for drivers/wingcc.c, and indeed the initial commit (back
in 2004 by Andrew Roach was the originator of the plfreetype approach
but who has since retired from PLplot development) was for "users of
GCC windows compilers such as mingw".  But clearly the wingcc name is
now a misnomer.  Of course, you don't want to break backwards
compatibility on this name now unless there is a really good reason,
but one possible good reason is if someone could think up a really
compelling name for this device driver that we all immediately like.

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-devel] Using Window's raw API for shapes and text

[Jim D. <ji...@di...>]

> On May 13, 2015, at 8:43 PM, Alan W. Irwin <ir...@be...> wrote:
> 
>> On 2015-05-13 18:27-0400 Jim Dishaw wrote:
>> 
>> 
>>> On May 13, 2015, at 6:17 PM, Aaron Hexamer <he...@co...> wrote:
>>> 
>>> Alan,
>>> 
>>> I'm a little confused by the statement about copying code from Jim's driver
>>> to wingcc.  I just assumed that wingcc was named as such because maybe
>>> PLplot was first ported to windows with support only for the gcc compiler?
>>> I.e. is there really anything specific to gcc about wingcc anymore.  I
>>> recall using it with MSVC when I first got started with PLplot. If Jim's
>>> driver did everything that wingcc does, but without plfreetype, would there
>>> be a need for both?
>>> 
>>> Aaron.
> 
>> I think the smart thing to do is to have one Windows driver. I can't
> recall any differences that could not be handled without an #ifdef
> block.
> 
> Hi Jim:
> 
> Any way you decide to implement your driver (i.e., separately or as
> part of wingcc) is fine with me.
> 
> @Aaron:
> 
> Your question about the wingcc name got me curious so I looked at git
> log results for drivers/wingcc.c, and indeed the initial commit (back
> in 2004 by Andrew Roach was the originator of the plfreetype approach
> but who has since retired from PLplot development) was for "users of
> GCC windows compilers such as mingw".  But clearly the wingcc name is
> now a misnomer.  Of course, you don't want to break backwards
> compatibility on this name now unless there is a really good reason,
> but one possible good reason is if someone could think up a really
> compelling name for this device driver that we all immediately like.
> 
> Alan
> 

How about mswin. 
> __________________________
> 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-devel] Using Window's raw API for shapes and text

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

On 2015-05-13 20:56-0400 Jim Dishaw wrote:

>> On May 13, 2015, at 8:43 PM, Alan W. Irwin <ir...@be...> wrote:
[...] But clearly the wingcc name is
>> now a misnomer.  Of course, you don't want to break backwards
>> compatibility on this name now unless there is a really good reason,
>> but one possible good reason is if someone could think up a really
>> compelling name for this device driver that we all immediately like.
>
> How about mswin.

I don't particularly like that name because it emphasizes Microsoft.
That company clearly is in complete charge of the Windows platform,
but there are at least open enough about the specifications of that
platform that it is possible for others (such as the Wine developers)
to implement a Windows platform as well.  For example, the wingcc
driver works fine on the Wine version of Windows according to my
tests.

Before I posed my question, I was thinking of just "win", but I think
that is so bland it is almost meaningless.  Another possibility
is iwin (where the i stands for interactive), but again too bland.

Just free-associating here, but how about something like winwidget
to follow (roughly) the naming scheme for both our wxwidgets
and our qtwidget devices?

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-devel] Using Window's raw API for shapes and text

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

Would it be developed using the GDI?  If so, then maybe wingdi?

-----Original Message-----
From: Alan W. Irwin [mailto:ir...@be...] 
Sent: Wednesday, May 13, 2015 9:33 PM
To: Jim Dishaw
Cc: Aaron Hexamer; PLplot development list
Subject: Re: [Plplot-devel] Using Window's raw API for shapes and text

On 2015-05-13 20:56-0400 Jim Dishaw wrote:

>> On May 13, 2015, at 8:43 PM, Alan W. Irwin <ir...@be...>
wrote:
[...] But clearly the wingcc name is
>> now a misnomer.  Of course, you don't want to break backwards 
>> compatibility on this name now unless there is a really good reason, 
>> but one possible good reason is if someone could think up a really 
>> compelling name for this device driver that we all immediately like.
>
> How about mswin.

I don't particularly like that name because it emphasizes Microsoft.
That company clearly is in complete charge of the Windows platform, but
there are at least open enough about the specifications of that platform
that it is possible for others (such as the Wine developers) to implement a
Windows platform as well.  For example, the wingcc driver works fine on the
Wine version of Windows according to my tests.

Before I posed my question, I was thinking of just "win", but I think that
is so bland it is almost meaningless.  Another possibility is iwin (where
the i stands for interactive), but again too bland.

Just free-associating here, but how about something like winwidget to follow
(roughly) the naming scheme for both our wxwidgets and our qtwidget devices?

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-devel] Using Window's raw API for shapes and text

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

On 2015-05-13 21:44-0500 Aaron Hexamer wrote:

> Would it be developed using the GDI?  If so, then maybe wingdi?

Hi Aaron:

To respond to your first question even though I am not
that familiar with Windows, I did look up the article at
<http://en.wikipedia.org/wiki/Graphics_Device_Interface>, and it
appears that is an old API that has been replaced with the Direct2D
<http://en.wikipedia.org/wiki/Direct2D> API.
I then followed up with some google
searching to find
<https://msdn.microsoft.com/en-us/library/windows/desktop/ff729481(v=vs.85).aspx>
which implies you can use gdi for graphics and DirectWrite for text or
Direct2D for graphics and DirectWrite for text.  Furthermore,
<http://en.wikipedia.org/wiki/DirectWrite> says one of the features is

"Comprehensive support for Unicode, with over 20 scripts providing
layout and rendering of every language supported in Windows."

So that superficially sounds like exactly what we want to use (in
combination with either GDI or Direct2D), but remember my Windows
knowledge is quite limited and this research just took me a few
minutes of google searching so if our Windows developers have a
different preference for the Windows API that is used to render
unicode text, then we should adopt that preference.

@Aaron: I do think your general idea of the win+API name is better
than winwidgets. So I would be happy to use wingdi or windirect2d
depending on the decision about which one is used in conjunction with
DirectWrite.

@Jim, Arjen, and Phil:

As our most active Windows developers please chime in about what
Windows API you think we should use to render unicode text, and the
name you would like to see for what is currently called the wingcc
device driver.

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-devel] Using Window's raw API for shapes and text

[Jim D. <ji...@di...>]

> On May 14, 2015, at 12:11 AM, Alan W. Irwin <ir...@be...> wrote:
> 
> On 2015-05-13 21:44-0500 Aaron Hexamer wrote:
> 
>> Would it be developed using the GDI?  If so, then maybe wingdi?
> 
> Hi Aaron:
> 
> To respond to your first question even though I am not
> that familiar with Windows, I did look up the article at
> <http://en.wikipedia.org/wiki/Graphics_Device_Interface <http://en.wikipedia.org/wiki/Graphics_Device_Interface>>, and it
> appears that is an old API that has been replaced with the Direct2D
> <http://en.wikipedia.org/wiki/Direct2D <http://en.wikipedia.org/wiki/Direct2D>> API.
> I then followed up with some google
> searching to find
> <https://msdn.microsoft.com/en-us/library/windows/desktop/ff729481(v=vs.85).aspx <https://msdn.microsoft.com/en-us/library/windows/desktop/ff729481(v=vs.85).aspx>>
> which implies you can use gdi for graphics and DirectWrite for text or
> Direct2D for graphics and DirectWrite for text.  Furthermore,
> <http://en.wikipedia.org/wiki/DirectWrite <http://en.wikipedia.org/wiki/DirectWrite>> says one of the features is
> 
> "Comprehensive support for Unicode, with over 20 scripts providing
> layout and rendering of every language supported in Windows."
> 
> So that superficially sounds like exactly what we want to use (in
> combination with either GDI or Direct2D), but remember my Windows
> knowledge is quite limited and this research just took me a few
> minutes of google searching so if our Windows developers have a
> different preference for the Windows API that is used to render
> unicode text, then we should adopt that preference.
> 
> @Aaron: I do think your general idea of the win+API name is better
> than winwidgets. So I would be happy to use wingdi or windirect2d
> depending on the decision about which one is used in conjunction with
> DirectWrite.
> 
> @Jim, Arjen, and Phil:
> 
> As our most active Windows developers please chime in about what
> Windows API you think we should use to render unicode text, and the
> name you would like to see for what is currently called the wingcc
> device driver.
> 

Bad news.  I searched for my old Windows driver and I have lost it to the ether.  I found some remnants, but not enough to compile.  I can a recreate it without too much difficultly.  The driver version that I had created was a merge between the X11 and the Windows GDI drivers because I wanted to have a common UI between the two platforms (e.g. menubar, crosshairs for picking points, printing).

Is it worthwhile for me to recreate this driver?  It should not interfere with the plbuf/plmetafile cleanup; in fact, digging into a driver might help in clarifying some concepts.

Re: [Plplot-devel] Using Window's raw API for shapes and text

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

On 2015-05-21 22:14-0400 Jim Dishaw wrote:

> Bad news.  I searched for my old Windows driver and I have lost it
to the ether.  I found some remnants, but not enough to compile.  I
can a recreate it without too much difficultly.  The driver version
that I had created was a merge between the X11 and the Windows GDI
drivers because I wanted to have a common UI between the two platforms
(e.g. menubar, crosshairs for picking points, printing).

> Is it worthwhile for me to recreate this driver?  It should not
interfere with the plbuf/plmetafile cleanup; in fact, digging into a
driver might help in clarifying some concepts.

Hi Jim:

That is mostly up to you although I do take your point it could be a
useful exercise for you.  Also, I don't use the xwin menubar or
printer feature at all so cannot properly evaluate their capabilities,
but from comparisons I have done with the -locate feature of
examples/c/x01c.c, running the 3rd (interactive) page of example 20,
and running special octave interactive examples, the xwin crosshairs
capability is the best of all our devices, and it would well be worth
copying what it does to xcairo, qtwidget, wxwidgets, wingcc, and the
new windows device.

Just in case you decide it would be worthwhile to recreate that driver,
I searched through the plplot-devel archive I have collected over
the years and found some references to your Windows driver but
no specific attachments of your source code except possibly part
of your driver that was introduced by you back in 2008 in the following
discussion with Werner:

"
I think it is a good idea.  In fact, I have created an abstract driver
for an interactive device.  I wanted to create a consistent user
interface for both X11 and MS Windows.  In order to achieve that, I
split the driver into two parts:  The real device driver and the
abstract driver.

The real device driver defines the dispatch setup function (and the
dummy function if it is disabled).  The dispatch table is configured to
point to functions in the abstract driver.  The abstract driver calls
the underlying API calls through a function pointer.

I have attached my  abstract driver as a strawman.  It is a work in
progress, so it is a bit crude still.
"

I could send that attachment along to you if you lost the original
that you sent out.

You also might want to search the plplot-devel archives yourself
for any descriptions you gave of your device driver, if you
have not done that already.

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-devel] Using Window's raw API for shapes and text

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

Just seen this regarding raspberry pi and Windows 10. So Windows 10
will have an internet of thing (IoT) version to run on Arm processors.
I'm not sure if/how this is linked to Windows Phone. The aps have to
use either XAML, html5 or DirectX for the GUI. They can run "native
code" - i.e. compiled C++ rather than having to use a managed/scripted
language. Not sure if this influences how we might consider a windows
driver for the future?

Phil

Re: [Plplot-devel] Using Window's raw API for shapes and text

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

Sorry forgot to paste the link

http://www.theregister.co.uk/2015/04/30/windows_10_now_available_for_raspberry_pi_2_andminnowboard/

On 4 June 2015 at 13:10, Phil Rosenberg <p.d...@gm...> wrote:
> Just seen this regarding raspberry pi and Windows 10. So Windows 10
> will have an internet of thing (IoT) version to run on Arm processors.
> I'm not sure if/how this is linked to Windows Phone. The aps have to
> use either XAML, html5 or DirectX for the GUI. They can run "native
> code" - i.e. compiled C++ rather than having to use a managed/scripted
> language. Not sure if this influences how we might consider a windows
> driver for the future?
>
> Phil

Re: [Plplot-devel] Using Window's raw API for shapes and text

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

I have summarized the wingcc-relevant parts of this important thread at
https://sourceforge.net/p/plplot/feature-requests/20/.

@Jim:

I have made you the owner of this feature request on the assumption
(based on your historical interest in updating the wingcc device) that
it will be you that implements it.  N.B. my writeup of this feature
request was based on wikipedia-level information rather than
experience so feel free to edit this feature request if you find any
glaring errors in the current writeup.

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-devel] Using Window's raw API for shapes and text

[Jim D. <ji...@di...>]

Sounds good. I will take a look at it

> On Mar 9, 2016, at 6:07 PM, Alan W. Irwin <ir...@be...> wrote:
> 
> I have summarized the wingcc-relevant parts of this important thread at
> https://sourceforge.net/p/plplot/feature-requests/20/.
> 
> @Jim:
> 
> I have made you the owner of this feature request on the assumption
> (based on your historical interest in updating the wingcc device) that
> it will be you that implements it.  N.B. my writeup of this feature
> request was based on wikipedia-level information rather than
> experience so feel free to edit this feature request if you find any
> glaring errors in the current writeup.
> 
> 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
> __________________________