On Sat, Aug 22, 2009 at 4:05 PM, Alan W. Irwin<ir...@beluga.phys.uvic.ca> wrote:
> Here is some background to help all of us figure out what is going on. The
> -cmap0 option is parsed and executed when plparseopts is executed. If that
> command-line option is not specified, then plinit calls plstrm_init which
> calls plspal0("") (for the case when plsc->cmap0 == NULL) which has exactly
> the same effect (default colours) as the -cmap0 "" option. So we should
> never get bad cmap0 colours unless plsc->cmap0 != NULL, i.e., some other
> method has been used to initialize colours.
>
> The previous version of the cmap0 initialization used a call to plscmap0n(0)
> rather than plspal0(""). The old version of plscmap0n(0) would give you
> default cmap0 colours under certain circumstances, but now it just gives you
> red colours (see revision 10315 where I changed the identification but not
> the rgb values appropriately to "red" from "black") as a warning that the
> colours have not been initialized properly.
>
> Following up on this, I tried running plscmap0n(0) before plinit in example
> 10 for the python case (note not Ada), and sure enough I got an all-red
> result. Some of our users must be depending on the old cmap0 initialization
> behaviour (plscmap0n(0) setting default colours if called before plinit) so
> that past behaviour should be restored.
>
> Hez, could you do that by inserting a call to plspal0("") into plscmap0n for
> the case when it is appropriate (plsc->cmap0 == NULL)? N.B. I am
> specifically not asking for a return to all the specific hard-coded colours
> implemented with color_def that we had before since I far prefer the colour
> palette file approach. Of course, the tricky part of implementing this with
> colour palette files is to avoid infinitely recursive behaviour (since
> plscmap0n(0) is called by plspal0). That is why I am asking someone more
> expert than me in C to deal with this. :-)
>
> Once plscmap0n(0) gives default colours like before, I suspect all the Ada
> red issues will go away. Jerry has remarked before that our Ada bindings do
> their own colour initialization, and my guess is that boils down in the
> cmap0 case to a call to plscmap0n(0).
>
> Alan
Alan,
Thanks for tracking this down. I have attached a patch which should
properly implement what you want and, hopefully, fix the "all red"
problem. A brief code-review on the logic in this patch would be
appreciated as well - I don't think that there are any infinite loops
lurking in there, but it is possible. Example 2 runs cleanly with
this patch using the xwin driver, which is where I ran in to infinite
loop problems when implementing the palette file-based default color
map changes.
I do not have a working Ada install yet, so I can not test the effect
this has on the Ada bindings. I can, however, call plscmap0n with 0,
negative and positive values before calling plinit from OCaml,
resulting in a proper palette of colors.
If it works for you and/or others I will go ahead and commit the change.
Hez
--
Hezekiah M. Carty
Graduate Research Assistant
University of Maryland
Department of Atmospheric and Oceanic Science
diff --git a/src/plctrl.c b/src/plctrl.c
index a6e8a3f..8dcc367 100644
--- a/src/plctrl.c
+++ b/src/plctrl.c
@@ -81,6 +81,9 @@ plcmap1_def(void);
static PLFLT
value(double n1, double n2, double hue);
+PLBOOL
+is_cmap0_ready(PLINT ncol0);
+
/* An additional hardwired location for lib files. */
/* I have no plans to change these again, ever. */
@@ -749,6 +752,35 @@ plcmap1_calc(void)
}
/*--------------------------------------------------------------------------*\
+ * is_cmap0_ready()
+ *
+ * Check to see if color map 0 has enough color slots allocated to house the
+ * requested number of colors.
+\*--------------------------------------------------------------------------*/
+
+PLBOOL
+is_cmap0_ready(PLINT ncol0)
+{
+ PLBOOL ret_val;
+ ret_val = FALSE;
+
+ if (ncol0 > 0 && plsc->ncol0 == ncol0) {
+ /* OK because ncol0 = the number of currently allocated cmap0 colors */
+ ret_val = TRUE;
+ }
+ else if (plsc->ncol0 <= 0 || ncol0 <= 0) {
+ /* NOT OK because cmap0 is not ready to hold any colors */
+ ret_val = FALSE;
+ }
+ else {
+ /* NOT OK because more colors are requested than are currently
+ allocated */
+ ret_val = FALSE;
+ }
+ return ret_val;
+}
+
+/*--------------------------------------------------------------------------*\
* plscmap0n()
*
* Set number of colors in cmap 0, (re-)allocate cmap 0, and fill with
@@ -762,6 +794,10 @@ void
c_plscmap0n(PLINT ncol0)
{
int ncol, size, imin, imax;
+ PLBOOL first_allocation;
+
+ /* Is this the first initialization of cmap0? */
+ first_allocation = FALSE;
/* No change */
@@ -788,6 +824,7 @@ c_plscmap0n(PLINT ncol0)
plexit("c_plscmap0n: Insufficient memory");
}
imin = 0;
+ first_allocation = TRUE;
}
else {
if ((plsc->cmap0 = (PLColor *) realloc(plsc->cmap0, size))==NULL)
@@ -795,6 +832,7 @@ c_plscmap0n(PLINT ncol0)
plexit("c_plscmap0n: Insufficient memory");
}
imin = plsc->ncol0;
+ first_allocation = FALSE;
}
/* Fill in default entries */
@@ -802,6 +840,10 @@ c_plscmap0n(PLINT ncol0)
plsc->ncol0 = ncol;
plcmap0_def(imin, imax);
+ if (ncol0 == 0 || first_allocation == TRUE) {
+ plspal0("");
+ }
+
if (plsc->level > 0)
plP_state(PLSTATE_CMAP0);
}
@@ -1208,11 +1250,9 @@ c_plspal0(const char *filename)
return;
}
- /* Allocate default number of cmap0 colours if cmap0 allocation not
+ /* Allocate sufficient cmap0 colors to contain the data if it has not been
done already. */
- plscmap0n(0);
- /* Allocate sufficient cmap0 colours to contain present data. */
- if(number_colors > plsc->ncol0) {
+ if(!is_cmap0_ready(number_colors)) {
plscmap0n(number_colors);
}
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel
