luke@umnstat.uucp (Luke Tierney) (09/24/90)
We recently installed ULTRIX 4.0 and the 4.01 upgrade on a Decstation
3100 and ran into two problems that may be bugs in the X rervers. The
first problem involves the XDrawPoints routine and may only affect the
affects the Xmfb server. The second involves freeing colors allocated
with XAllocColor seems to affect both Xmfb and Xcfb servers.
I have put together two little programs that exercise the problems;
they are contained in the shar file attached below. I am not sure whom
to contact about this, so I am posting this note here in the hope that
someone at DEC may be able to relay this to the proper place.
Luke Tierney
luke@umnstat.stat.umn.edu
------------------------------(cur here)-----------------------------------
#! /bin/sh
# This is a shell archive, meaning:
# 1. Remove everything above the #! /bin/sh line.
# 2. Save the resulting text in a file.
# 3. Execute the file with /bin/sh (not csh) to create the files:
# dxbugs
# This archive created: Sat Sep 22 15:02:23 1990
export PATH; PATH=/bin:$PATH
if test ! -d 'dxbugs'
then
mkdir 'dxbugs'
fi
cd 'dxbugs'
if test -f 'Makefile'
then
echo shar: will not over-write existing file "'Makefile'"
else
cat << \SHAR_EOF > 'Makefile'
all: drawpointsbug freecolorbug
drawpointsbug: drawpointsbug.o
cc -o drawpointsbug drawpointsbug.o -lX11
freecolorbug: freecolorbug.o
cc -o freecolorbug freecolorbug.o -lX11
clean:
rm -f *~ *.o
cleanall:
make clean
rm -f drawpointsbug freecolorbug
SHAR_EOF
fi # end of overwriting check
if test -f 'README'
then
echo shar: will not over-write existing file "'README'"
else
cat << \SHAR_EOF > 'README'
The two little programs in this directory exercise two possible
bugs I have come accress.
BUG1
====
The XDrawPoints routine does not seem to work in the
CoordModePrevious. The program drawdpointsbug draws three points
using this routine alternating between CoordModeOrigin and
CoordModePrevious. Initially CoordModeOrigin is used, then each time
the draw() routine is called the mode is switched. draw() is called in
response to a click. When the Displaying on an MIT R4 server for a sun
3/50 clicking has no effect. On the ULTRIX 4.01 server the points are
not in a line when using the CoordModePrevious mode. I think the
ULTRIX 3.x servers may not have had this problem.
BUG2
====
Freeing a color allocated with XAllocColor or XAllocNamedColor
produces an error. The program freecolorbug exercises this. The
problem does not occur on an MIT R4 server running on a sun 3/50. I
think it also does not occur in ULTRIX 3.x servers.
SHAR_EOF
fi # end of overwriting check
if test -f 'drawpointsbug.c'
then
echo shar: will not over-write existing file "'drawpointsbug.c'"
else
cat << \SHAR_EOF > 'drawpointsbug.c'
#include <X11/Xlib.h>
#include <stdio.h>
#define TRUE 1
#define FALSE 0
/* Options */
#define OWN_DASHES FALSE
#define SOLID FALSE
#define FIX_GC FALSE
#define DASH_LENGTH 4
#define TOP 100
#define LEFT 100
#define WIDTH 400
#define HEIGHT 400
#define POINTS 3
#define YSTEP 3
#define START 100
Display *dpy;
int screen;
GC gc;
Window win;
XPoint rpts[POINTS], apts[POINTS];
main()
{
char *display_name = NULL;
unsigned long valuemask;
XGCValues values;
XEvent report;
int done = FALSE;
int i;
/* set up the point arrays */
rpts[0].x = START;
rpts[0].y = START;
for (i = 1; i < POINTS; i++) {
rpts[i].x = -1;
rpts[i].y = -YSTEP;
}
for (i = 0; i < POINTS; i++) {
apts[i].x = START - i;
apts[i].y = START - YSTEP * i;
}
/* connect to X server */
if ((dpy = XOpenDisplay(display_name)) == NULL) {
fprintf(stderr, "%s: can't connect to X server %s\n", "bug",
XDisplayName(display_name));
exit(-1);
}
/* get the default screen */
screen = DefaultScreen(dpy);
/* make GC */
valuemask = 0; /* ignore values */
gc = XCreateGC(dpy, RootWindow(dpy, screen), valuemask, &values);
XSetForeground(dpy, gc, BlackPixel(dpy, screen));
XSetBackground(dpy, gc, WhitePixel(dpy, screen));
/* create the window */
win = XCreateSimpleWindow(dpy, RootWindow(dpy, screen),
LEFT, TOP, WIDTH, HEIGHT, 1,
BlackPixel(dpy, screen),
WhitePixel(dpy, screen));
XSelectInput(dpy, win, ExposureMask | ButtonPressMask);
XStoreName(dpy, win, "Click In Me");
/* Display (map) the window */
XMapWindow(dpy, win);
while (! done) {
XNextEvent(dpy, &report);
if (win == report.xany.window) {
switch(report.type) {
case Expose: if (report.xexpose.count == 0) draw(); break;
case ButtonPress: draw(); break;
default: break;
}
}
}
XCloseDisplay(dpy);
}
/**************************************************************************/
/** **/
/** Drawing Functions **/
/** **/
/**************************************************************************/
draw()
{
static int relative = FALSE;
XClearWindow(dpy, win);
if (relative)
XDrawPoints(dpy, win, gc, rpts, POINTS, CoordModePrevious);
else
XDrawPoints(dpy, win, gc, apts, POINTS, CoordModeOrigin);
relative = ! relative;
}
SHAR_EOF
fi # end of overwriting check
if test -f 'freecolorbug.c'
then
echo shar: will not over-write existing file "'freecolorbug.c'"
else
cat << \SHAR_EOF > 'freecolorbug.c'
#include <X11/Xlib.h>
#include <stdio.h>
#define TRUE 1
#define FALSE 0
#define MULTIPLIER 62535
#define MYRED 0.8
#define MYGREEN 0.2
#define MYBLUE 0.2
main()
{
char *display_name = NULL;
Display *dpy;
int screen;
Colormap cmap;
XColor color;
/* connect to X server */
if ((dpy = XOpenDisplay(display_name)) == NULL) {
fprintf(stderr, "%s: can't connect to X server %s\n", "bug",
XDisplayName(display_name));
exit(-1);
}
/* get the default screen */
screen = DefaultScreen(dpy);
/* get the default color map */
cmap = DefaultColormap(dpy, screen);
/* allocate a color */
color.red = MULTIPLIER * MYRED;
color.green = MULTIPLIER * MYGREEN;
color.blue = MULTIPLIER * MYBLUE;
if (XAllocColor(dpy, cmap, &color) == 0) {
fprintf(stderr, "can't allocate color\n");
exit(-1);
}
/* free the color */
XFreeColors(dpy, cmap, &color.pixel, 1, 0);
XCloseDisplay(dpy);
fprintf(stderr, "No Error\n");
}
SHAR_EOF
fi # end of overwriting check
cd ..
# End of shell archive
exit 0hoyt@laura.alf.dec.com (Kurt Hoyt) (09/24/90)
In article <1990Sep23.171606.16370@cs.umn.edu>, luke@umnstat.uucp (Luke Tierney) writes: |> From: luke@umnstat.uucp (Luke Tierney) |> We recently installed ULTRIX 4.0 and the 4.01 upgrade on a Decstation |> 3100 and ran into two problems that may be bugs in the X rervers. The [stuff deleted] |> Luke Tierney |> luke@umnstat.stat.umn.edu [more stuff deleted] |> BUG2 |> ==== |> |> Freeing a color allocated with XAllocColor or XAllocNamedColor |> produces an error. The program freecolorbug exercises this. The |> problem does not occur on an MIT R4 server running on a sun 3/50. I |> think it also does not occur in ULTRIX 3.x servers. [code deleted] I ran this program on a DECstation 3100 with ULTRIX 4.0 (and the mandatory patches) using Xcfb. Repeatedly running freecolorbug resulted in the message "No Error" each time. Do we have differing environments of some kind? -- Kurt in Atlanta | "A Swiss court has ruled that cowbells are a hoyt@decatl.alf.dec.com | form of noise pollution and cows must be hoyt@laura.alf.dec.com | deprived of them immediately." National Review