I am pretty sure you meant to send the remarks below (which you only
to me) to the list as well so I am going to CC the list in my reply so
everybody can see your remarks.
On 2014-02-28 17:54-0800 phil rosenberg wrote:
>> On Thursday, 27 February 2014, 17:04, Alan W. Irwin <firstname.lastname@example.org
>> From recent remarks made by Brad King at
appears the problem
>> had already been fixed in their development version but not released
>> So as the next step you should look at the dates of all the commits he
>> referred to and copy the blob from the latest of those into PLplot's
>> cmake/modules/FindFreeType.cmake (which will then supersede whatever
>> FindFreetype.cmake that comes with your CMake version) to see whether
>> that latest version works for you.
> I have just upgraded my CMake. Hopefully that will do the job.
I doubt even the latest released version of FindFreeType.cmake is
going to help you. Instead, please follow what I suggested above to
get the latest _unreleased_ version. And if it works, please commit
that new version of cmake/modules/FindFreeType.cmake so that all
PLplot users can benefit regardless of their CMake version.
> I've been working on plmap over the last few days and am
close to something worth submitting. I have rewritten the way it does
wrapping which should fix a few bugs and also be rather more
efficient. I also added support for shapelib polygons as well as
lines. Finally I added the ability to select which items of a
shapefile are drawn. I would like some advice on what people think is
the best interface for this. I added two new variables to plmap - an
array of PLINTs, and the number of points in that array. This array
provides the indices of the items the user wishes to plot. Passing
NULL for the array just plots everything. Does this sound like the
best way to do this to everyone? I could imagine doing it a few
different ways, such as passing an array of 0s and 1s to not plot or
plot each element or parsing
database requests to filter certain
features (but this is obviously more effort and actually a bit less
> I could also leave the old version in place as an overload, but I'm
not sure if all the languages we support have the ability to deal
Congratulations on being so close to an improved plmap API. I don't
feel that competent to answer the questions in the two paragraphs
above. For example, you talk about overloading which I know is
possible with f95, Python, and C++, and ???, but is that possible with
C which is the language you should use for all core library functions?
Anyhow, I am repeating your paragraphs above for the benefit of Andrew
and Hez who are likely more competent
to respond on these API issues.
A further remark is it is always a good idea to ask (as you have done)
when API changes are involved, but if Andrew and Hez are unable to
respond in a timely way just use your best judgement for the commit.
After all, they have a whole release cycle to modify your change if
they prefer something different than what you have chosen.
> Alan, I seem to remember when we first introduced shapefiles you
wanted a new example to show off the capabilities. I've attached a map
of the coastline to the north of Exmoor UK, generated with the
wxWidget driver, which I thought might do the job. It uses shapelib
files which are licenced such that we can redistribute them and
it makes use of all the shapelib features I have included
only thing I am struggling with is setting the white background on
just this plot within example 19, but leaving the others black. It
works on the wxWidgets interactive interface, but doesn't seem to work
on the colour ps driver or the saved wxWidget output.
I really like this example; a fully coloured contour map is a huge
improvement compared to the simple outline maps we currently have for
the rest of this example. However, I suggest you put some additional
decoration on the map (such as "Exmoor National Park, UK") to identify
it for our users. Also, Exmoor is a great start, but can you find
relevant shapefile map data featuring a spectacular crater or volcano
for an additional "coloured contour map" example page as well? I am
thinking along the lines of the big Island of Hawaii (with Mauna Loa
and Mauna Kea) or even Olympus Mons (on Mars). Teide in the Canary
Islands (see http://en.wikipedia.org/wiki/Teide
) looks fairly
spectacular as well, but, of course, for all of these suggestions the
big question for each of them is whether shapefile data exist in
publicly distributable form.
With regard to your question concerning changing the background from
one page to the next, the colour manipulations of example 16 (which
include a changed background colour in the colour palettes) are all
done just before the relevant pladv call, and I think the changed
backagrounds are consistent for all device drivers.
function used in example 19 contains an internal call to pladv. So my guess is
what you have to do is change the background just before the relevant
plenv call to get consistent changed background results for all device
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).