From Newsgroup: comp.windows.x

First of all, a fair warning: the license for xbattle is confusing:

README> COPYRIGHT

README> Copied from the GNU Emacs 'copyleft' policy. Basically, you
README> are free to copy, modify and distribute this program as long
README> as you do not sell it or use it in a product that you sell.

The "do not sell" restriction would make this software fail the
FSF free software criteria, but at the same time, the license
text itself mentions no such restrictions. It goes on to state:

README> c) You may charge a distribution fee for the physical act of
README> transferring a copy, and you may at your option offer warranty
README> protection in exchange for a fee.

Now, unlike many free software enthusiasts, I do not consider
the /use/ of (possibly, in this case) non-free software to be
an ethical issue per se, but endorsing it is clearly questionable
in that department.

So there.

Now, there're two issues with xbattle 4.0 as shipped (e. g.,
http://ftp.gwdg.de/pub/x11/x.org/R5contrib/xbattle-4.0.tar.Z ):

a. main.c seems to misuse XSetIOErrorHandler and FD_SET in main,
and also XLookupString and XDrawImageString in process_event;
I have a straightforward patch from 2002 (below) fixing that;

b. it checks if the display supports a 8 bits per pixel PseudoColor
visual, and uses monochrome if that's not the case; this is where
I had to get creative.

What I've used for b. is: a. add new bw_p field to xwindow_type
in extern.h; b. make it used consistently in the code in place
of xwindow->depth tests; c. replaced the initial tests like:

- if(xwindow->depth != 8)
- {
- if (XMatchVisualInfo(xwindow->display, xwindow->screen, 8, PseudoColor, &vinfo))
- {
- visual = vinfo.visual;
- xwindow->depth = 8;
- }

+ if (xwindow->depth == 8) {
+ xwindow->bw_p = 0;
+ } else if (XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 8, PseudoColor, &vinfo)
+ || XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 15, TrueColor, &vinfo)
+ || /* ... and so on... */

So far I'm testing for PseudoColor at depth 8, and TrueColor
visuals at depths 15, 16, 24, 32.

And the question is: what's an easy way to check for a TrueColor
(PseudoColor) visual having /at least/ a given depth?

TIA. The diffs follow.

--- xbattle-4.0/main.c 1992-09-27 16:52:19 +0000
+++ xbattle-4.0/main.c 2002-03-26 09:06:18 +0000
@@ -155,7 +155,7 @@ main(argc,argv)

/************* main event loop *************/
#if UNIX
- XSetIOErrorHandler(1);
+ /* XSetIOErrorHandler(1); */
gettimeofday (&tp_old, &tzp);
oldtime = (tp_old.tv_sec%10000)*1000000 + tp_old.tv_usec;
targettime = oldtime + (delay/10000)*1000000 + (delay%10000)*100;
@@ -207,7 +207,8 @@ main(argc,argv)
{
if (winopen[i])
{
- if (FD_SET(fd = ConnectionNumber(xwindow[i]->display), &rfds))
+ fd = ConnectionNumber(xwindow[i]->display);
+ if (FD_ISSET(fd, &rfds))
{
while (winopen[i] &&
XEventsQueued(xwindow[i]->display, QueuedAfterReading) > 0)
@@ -543,7 +544,7 @@ process_event(event,board,colorarray,pla

/**** process key press events ****/
case KeyPress:
- textcount = XLookupString(&event, text, 10, &key, NULL);
+ textcount = XLookupString(&(event.xkey), text, 10, &key, NULL);
if (textcount != 0)
{
if ((enable_hex && (event.xbutton.y < boardsizey*2*hex_vert + hex_vert)) ||
@@ -1208,7 +1209,7 @@ replaygame(squaresize,board)
for (l=0; l<nsides; l++)
XDrawImageString(xwindow[k]->display,xwindow[k]->window,
xwindow[k]->hue[light],
- TEXTX,textyh[l],
+ TEXTX,*(textyh[l]),
messagestr,strlen(messagestr));
#else
XDrawImageString(xwindow[k]->display,xwindow[k]->window,
--- xbattle-4.0-build/draw.c.~1~ 1992-09-27 16:52:24 +0000
+++ xbattle-4.0-build/draw.c 2026-01-24 19:28:30 +0000
@@ -342,8 +342,8 @@ drawsquare(xwindow,square,windowcolor,sq
else
{
onhue = xwindow->hue[square->color];
- if ((xwindow->depth==1 && color2bw[sidemap[square->color]]<=GRAY3) ||
- (xwindow->depth==8 && sidemap[square->color]==BLACK))
+ if (xwindow->bw_p ? color2bw[sidemap[square->color]]<=GRAY3
+ : sidemap[square->color]==BLACK)
offhue = xwindow->hue[light];
else
offhue = xwindow->hue[dark];
--- xbattle-4.0-build/extern.h.~1~ 1992-09-27 16:52:24 +0000
+++ xbattle-5.0-build/extern.h 2026-01-24 19:17:38 +0000
@@ -103,7 +103,7 @@ typedef struct{
int xsize,ysize;
XSizeHints hint;
XWMHints xwmh;
- int depth,screen;
+ int bw_p,depth,screen;
GC hue[MAXSIDES+3], flip, gc_clear, gc_or;
Pixmap terrain[NHILLTONES];
XFontStruct *font_struct;
--- xbattle-4.0-build/utils.c.~1~ 1992-09-27 16:52:33 +0000
+++ xbattle-4.0-build/utils.c 2026-01-25 11:13:16 +0000
@@ -1,4 +1,6 @@
#include <stdio.h>
+#include <string.h>
+#include <time.h>

/**** x include files ****/
#include <X11/Xlib.h>
@@ -74,20 +76,27 @@ open_xwindow(xwindow,displayname,xpos,yp

visual = DefaultVisual(xwindow->display,xwindow->screen);

- if(xwindow->depth != 8)
- {
- if (XMatchVisualInfo(xwindow->display, xwindow->screen, 8, PseudoColor, &vinfo))
- {
- visual = vinfo.visual;
- xwindow->depth = 8;
- }
+ /** FIXME: add a command line option for black and white rendering */
+ if (xwindow->depth == 8) {
+ xwindow->bw_p = 0;
+ } else if (XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 8, PseudoColor, &vinfo)
+ /** FIXME: is this still a sensible approach? */
+ || XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 15, TrueColor, &vinfo)
+ || XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 16, TrueColor, &vinfo)
+ || XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 24, TrueColor, &vinfo)
+ || XMatchVisualInfo (xwindow->display, xwindow->screen,
+ 32, TrueColor, &vinfo)) {
+ visual = vinfo.visual;
+ xwindow->bw_p = 0;
+ } else {
+ xwindow->bw_p = 1;
}

- if (xwindow->depth != 8)
- xwindow->depth = 1;
-
- if(xwindow->depth == 8)
- {
+ if (! xwindow->bw_p) {
/**** create color map using visual ****/
#if NEWCOLORMAP
xwindow->cmap = XCreateColormap(xwindow->display,
@@ -123,11 +132,10 @@ open_xwindow(xwindow,displayname,xpos,yp
}

/**** get standard properties for window manager ****/
- if (xwindow->depth == 8)
+ if (! xwindow->bw_p)
XSetStandardProperties(xwindow->display, xwindow->window, htitle, htitle,
None, NULL, None, &xwindow->hint);
- else if (xwindow->depth == 1)
- {
+ else {
XSetStandardProperties(xwindow->display, xwindow->window, gtitle, gtitle,
None, NULL, None, &xwindow->hint);
}
@@ -142,8 +150,7 @@ open_xwindow(xwindow,displayname,xpos,yp
XSelectInput(xwindow->display, xwindow->window, eventmask);

/**** load color map ****/
- if(xwindow->depth == 8)
- {
+ if(! xwindow->bw_p) {
for (i=0 ; i<MAXCOLORS ; i++)
{
xwindow->xcolor[i].flags = 0;
@@ -307,8 +314,7 @@ open_xwindow(xwindow,displayname,xpos,yp
XSetFunction(xwindow->display, xwindow->flip, GXinvert);

/**** set drawing colors ****/
- if (xwindow->depth == 1)
- {
+ if (xwindow->bw_p) {
for (j=0; j<nsides; j++) {

if (color2bw[sidemap[j]] == WHITE)
@@ -350,7 +356,7 @@ open_xwindow(xwindow,displayname,xpos,yp
}

/**** set colors (link GC to color) ****/
- if (xwindow->depth == 8)
+ if (! xwindow->bw_p)
{
#if NEWCOLORMAP
for (j=0; j<nsides; j++)
@@ -387,8 +393,7 @@ open_xwindow(xwindow,displayname,xpos,yp
}
}

- if (xwindow->depth == 1)
- {
+ if (xwindow->bw_p) {
/**** set grey stipple color ****/

init_stipple (xwindow->display, xwindow->window, stipple);
@@ -426,7 +431,7 @@ open_xwindow(xwindow,displayname,xpos,yp
squaresize, squaresize, xwindow->depth);
}

- if (xwindow->depth == 8)
+ if (! xwindow->bw_p)
{
for (i=0; i<limit; i++)
{
@@ -438,7 +443,7 @@ open_xwindow(xwindow,displayname,xpos,yp
terrainhue[i], 0, 0, squaresize, squaresize);
}
}
- else if (xwindow->depth == 1)
+ else if (xwindow->bw_p)
{
if (enable_hex)
init_terrain_pixmaps (xwindow, limit, 2*hex_side);
@@ -544,7 +549,7 @@ open_xwindow(xwindow,displayname,xpos,yp
XSetFont(xwindow->display, xwindow->flip, xwindow->font_struct->fid);
}

- if (xwindow->depth == 1)
+ if (xwindow->bw_p)
{
for (i=0; i<nsides; i++)
{
@@ -567,8 +572,8 @@ open_xwindow(xwindow,displayname,xpos,yp
xwindow->drawletter[none] = FALSE;

/**** map window to make it visible ****/
- if (xwindow->depth == 8)
- XMapWindow(xwindow->display, xwindow->window);
+ /* FIXME: ? if (! xwindow->bw_p) */
+ XMapWindow(xwindow->display, xwindow->window);
}


--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.windows.x

On Wed, 11 Feb 2026 08:37:07 +0000, Ivan Shmakov wrote:

First of all, a fair warning: the license for xbattle is confusing:

README> COPYRIGHT

README> Copied from the GNU Emacs 'copyleft' policy. Basically, you
README> are free to copy, modify and distribute this program as
README> long as you do not sell it or use it in a product that you
README> sell.

The "do not sell" restriction would make this software fail the
FSF free software criteria, but at the same time, the license text
itself mentions no such restrictions.

Is the distinction between rCLcopyleftrCY and rCLnon-commercialrCY really that subtle, that some people are capable of being confused by it? The two
have nothing to do with each other, no matter what the xbattle
developer(s) may think.

In particular, rCLcopyleftrCY is a well-defined concept, while rCLnon-commercialrCY is not.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.windows.x

In comp.windows.x, Lawrence DOliveiro <ldo@nz.invalid> wrote:

On Wed, 11 Feb 2026 08:37:07 +0000, Ivan Shmakov wrote:

^^^^

Is the distinction between "copyleft" and "non-commercial" really that subtle, that some people are capable of being confused by it? The two
have nothing to do with each other, no matter what the xbattle
developer(s) may think.

In particular, "copyleft" is a well-defined concept, while
"non-commercial" is not.

That's very easy for you to say 34 years after the code was packaged.
The license included probably dates to the 1991 release (I found a date
of "10/1/91" in one file, the README just says "1991"), but the tar file referenced has all September 1992 files. The nuances of different open
source licenses were not well publicized at the time. The FSF probably
didn't have a website in September 1992.

After applying Ivan's patches, I also needed to add some includes to
main.c, and then changed the C standard in the Makefile. Note that
the Makefile is generated by xmkmf, and not included in the source.
I don't know enough imake to know if/how I should have modified that.

--- a/main.c
+++ b/main.c
@@ -1,4 +1,6 @@
#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>

/**** x include files ****/
#include <X11/Xlib.h>

--- a/Makefile
+++ b/Makefile
@@ -323,7 +323,7 @@ MODLDCOMBINEFLAGS = -r
PROJECTROOT = /usr

CDEBUGFLAGS = -g -O2 -fno-strict-aliasing
- CCOPTIONS =
+ CCOPTIONS = -std=c89

ALLINCLUDES = $(INCLUDES) $(EXTRA_INCLUDES) $(TOP_INCLUDES) $(INSTALLED_INCLUDES) $(STD_INCLUDES)
ALLDEFINES = $(ALLINCLUDES) $(STD_DEFINES) $(PROTO_DEFINES) $(THREADS_DEFINES) $(MODULE_DEFINES) $(DEFINES) $(EXTRA_DEFINES)


That got me an executible that runs, but all I can get it to do is spit
out usage.

$ /xbattle -hex -board 100 -horizon 20 -map -localmap
USAGE: xbattle
-<c1> <display> display name side uses color c1 with
default color to b/w mapping
-<c1><c2> <display> display name side uses color c1 with
b/w terminals using b/w shade c2
-hex Use hexagonal display
[... many lines ...]

So I'm not sure how to actually run it. Possibly a read of the code
would be helpful.

Elijah
------
might look at the start-up code tomorrow
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.windows.x

On Thu, 12 Feb 2026 06:53:49 -0000 (UTC), Eli the Bearded wrote:

In comp.windows.x, Lawrence DOliveiro <ldo@nz.invalid> wrote:

In particular, "copyleft" is a well-defined concept, while
"non-commercial" is not.

The nuances of different open source licenses were not well
publicized at the time [34 years ago].

rCLPublicizedrCY or not, it behooves a developer to at least read and understand the licence thay are considering using, does it not?
After all, they were going to include a copy with their code.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.windows.x

In comp.windows.x, Lawrence DOliveiro <ldo@nz.invalid> wrote:

"Publicized" or not, it behooves a developer to at least read and
understand the licence thay are considering using, does it not?
After all, they were going to include a copy with their code.

The "license" is not called that. It's called "COPYING POLICIES" and
it seems to be a modified, and perhaps shortened, crib from a license
from some version of GNU Emacs.

Oh, and I figured out how to get it run now. I had overlooked there's a
man page included. This is old-school X11 and multiplayer relies on
this one program being able to open windows on a remote system. So to
just see the game, the most basic command line would be:

xbattle -red me

(Which is a rather plain looking grid.)

Or any x11 color:

xbattle -$X11_COLOR_NAME me

Multiple player takes the hostname and $DISPLAY:

xbattle -red me -blue lawrence:0.0

Then the other options are optional, but you get more of a game board.

This died with a "Memory fault":

xbattle -red me -hex -board 100

But this gave me a complicated board:

xbattle -red me -hex -board 20 -horizon 4 -map -localmap \
-bases 10 -rbases 2 -armies 5 -militia 10 -hills 2 \
-forest 3 -sea 1 -farms 2 -erode 2 -decay 2 -towns 3 \
-guns 4 -para 4 -fight 1 -speed 1 -move 2 -digin 2

There's a noticable delay between starting the program and the board
getting drawn in the window. I could click around and things changed,
but not much, it wasn't clear what was happening. A more careful read
of the man page would probably help.

Elijah
------
wondering what other memory faults lurk in the code
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.windows.x

On 2026-02-12, Eli the Bearded wrote:
In comp.windows.x, Lawrence DOliveiro <ldo@nz.invalid> wrote:
On Thu, 12 Feb 2026 06:53:49 -0000 (UTC), Eli the Bearded wrote:

After applying Ivan's patches, I also needed to add some includes
to main.c, and then changed the C standard in the Makefile.

--- a/main.c
+++ b/main.c
@@ -1,4 +1,6 @@
#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>

ACK, thanks.

- CCOPTIONS =
+ CCOPTIONS = -std=c89

I will keep that in mind, but I'm going to guess that's needed
only when C23 is being the default. I've used GCC 10.5.0 myself,
and at a glance, it requires no -std= options for this code.

"Publicized" or not, it behooves a developer to at least read and
understand the licence thay are considering using, does it not?
After all, they were going to include a copy with their code.

The "license" is not called that. It's called "COPYING POLICIES"
and it seems to be a modified, and perhaps shortened, crib from a
license from some version of GNU Emacs.

I'm going to concur. One or both of the authors of this code
apparently misunderstood the point of copyleft - back when it
wasn't something widely known. So sue them.

Oh, and I figured out how to get it run now. I had overlooked
there's a man page included.

There're also tutorial1 and tutorial2 shell scripts that are
supposed to introduce new players to the game by starting it
up with different options, beginning with the most simple
arrangements.

This is old-school X11 and multiplayer relies on this one program
being able to open windows on a remote system.

I'd call it a multi-display multiplayer game. The display(s)
in question need not be remote: it's possible to explore game
mechanics with, say, -red me -blue me, and it's also possible
to play from local Xvnc instances (-red me -blue :29 -cyan :43)
- accessed through remote VNC clients.

When using X, my suggestion these days would be to use a nested
X server, such as Xephyr or Xnest, rather than granting access
to one's "primary" one with xauth or xhost.

There're at least two games like that currently in pkgsrc -
craft and spellcast - and I hope to add xbattle and xtank to
the list as well. (I've got xtank to build and run, but the AI
is not functional, and I'm yet to try to fix that.)

This died with a "Memory fault":

xbattle -red me -hex -board 100

I guess I can link it against Electric Fence and take a closer
look for possible out-of-bounds array accesses (assuming that's
what's happening here.)

But this gave me a complicated board:

xbattle -red me -hex -board 20 -horizon 4 -map -localmap \
-bases 10 -rbases 2 -armies 5 -militia 10 -hills 2 \
-forest 3 -sea 1 -farms 2 -erode 2 -decay 2 -towns 3 \
-guns 4 -para 4 -fight 1 -speed 1 -move 2 -digin 2

There's a noticable delay between starting the program and the board
getting drawn in the window.

Not something I've personally noticed, but I suppose it might
depend on the specific options used.

I could click around and things changed, but not much, it wasn't
clear what was happening. A more careful read of the man page would probably help.

I've actually thought of including a summary of the game in the
initial post, then decided against it - for the sake of keeping
to the point.

Basically, each grid cell (square or hexagon) may contain
armies of a player. That player can click near a side of that
cell to order armies to start crossing that side into adjacent
cell. More than one order can be active; left-click toggles
orders, middle-click overrides all orders with a single new one.

Encountering enemy armies results in fights. With -rbases,
-towns and -farms, it's possible to replentish armies, so it
becomes necessary to set orders in cells in a way to form
supply lines from towns to the line(s) of conflict. Overall,
it results in a gameplay somewhat reminiscent of Liquidwar.

FWIW, the options I've used to start games back in the day are:

-hex -board 12 -area -map -localmap -horizon -bound -march 4
-repeat -attack -reserve -scuttle -farms 7 -decay 2
-sea 8 -towns 3 -hills 4 -forest 4 -fill 4 -dig 4 -build 8
-guns 5 -para 5 -rbases 1 -militia 3
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.windows.x

On Thu, 12 Feb 2026 23:57:36 -0000 (UTC), Eli the Bearded wrote:

In comp.windows.x, Lawrence DOliveiro <ldo@nz.invalid> wrote:

"Publicized" or not, it behooves a developer to at least read and
understand the licence thay are considering using, does it not?
After all, they were going to include a copy with their code.

The "license" is not called that. It's called "COPYING POLICIES" and
it seems to be a modified, and perhaps shortened, crib from a
license from some version of GNU Emacs.

Nasty.
--- Synchronet 3.21b-Linux NewsLink 1.2