[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference bulova::decw_jan-89_to_nov-90

Title:	DECWINDOWS 26-JAN-89 to 29-NOV-90
Notice:	See 1639.0 for VMS V5.3 kit; 2043.0 for 5.4 IFT kit
Moderator:	STAR::VATNE

Created:	Mon Oct 30 1989
Last Modified:	Mon Dec 31 1990
Last Successful Update:	Fri Jun 06 1997
Number of topics:	3726
Total number of notes:	19516

754.0. "C++ & DECwindows HELP!!" by PRSIS4::STRAUSS () Wed May 10 1989 10:52

    Although this continues from the topic I started previously (no.744)
    I decided to start a new topic because my problem might lead on
    to further discussion of the subject.
    
    I am currently trying to carry out some simple (!) programming using
    DECwindows and C++. I decided not to use any of the features of
    the DECwindows or X toolkits, but to use the functions of Xlib.
    
    The program I am currently working on - or should I say having
    difficulty with, is supposed to just open a window in which the user
    can draw lines by moving the mouse. 
    
    On compilation, however, all the Xlib functions used are rejected
    with compilation warnings stating:
    
    	The declaration for (Xlib function) was not specified
    
    My problem is that I don't understand why. I have a working version
    of the program that was written using traditional C. The main difference
    in the C++ version is that I have created a class called
    graphics_window and then created an instance of this class. The
    previous functions to enable the drawing of lines in the window
    thus becoming member functions of the graphics_window class.
    
    I realise that my program may not be elegant or efficient, but any
    help from anyone who has had success in programming DECwindows with
    C++ would be greatly appreciated.
    
    
#include <stdio.h>    
#include <prseac$dub2:[strauss.cpp.cpplib]Xlib.h>
#include <prseac$dub2:[strauss.cpp.cpplib]Xutil.h>

Display *dpy;
Screen *scr;

class graphics_window {
	Window			id;
	int			x,y,width,height;
	XSizeHints 		xsh;
	GC			gc;
	XGCValues		xgcv;
	XSetWindowAttributes	xswa;
	struct 
	  LastPoint {
		int X,Y;
		    }		last_pt;

public:
	graphics_window();
	void do_handle_events();
	void do_motion_notify(XEvent*);
	void do_button_press(XEvent*);
	void do_set_last_point(int, int);
	void do_draw_from_last_point(int, int);
};

graphics_window::graphics_window()
{
	width  = 500;
	height = 500;
	x      = 50;
	y      = 50;

	xswa.event_mask = ExposureMask |
			  ButtonPressMask |
			  Button3MotionMask ;

	xswa.background_pixel = XBlackPixelOfScreen;

        id = XCreateWindow(dpy, XRootWindowOfScreen(scr), 
			    x, y, width, height,		   
		            0, XDefaultDepthOfScreen(scr), 
			    InputOutput,
			    XDefaultVisualOfScreen(scr),
			    CWEventMask | CWBackPixel, &xswa);

	XStoreName(dpy, id, "If this doesn't bloody work this time...");

	xgcv.foreground = XWhitePixelOfScreen;
	xgcv.background = XBlackPixelOfScreen;
	xgcv.line_width = 3;
        xgcv.line_style = LineSolid;
	xgcv.cap_style  = CapButt;
	xgcv.join_style = JoinMiter;

	gc = XCreateGC(dpy, id, GCForeground | GCBackground |
 			GCLineWidth | GCLineStyle | GCCapStyle |
			 GCJoinStyle, &xgcv);

	xsh.x = 50;		
	xsh.y = 50;	
	xsh.width = 500;
	xsh.height = 500;
	xsh.flags = PPosition | PSize;

	XSetNormalHints(dpy, id, &xsh);

	XMapWindow(dpy, id);
}

void graphics_window::do_set_last_point(int X, int Y)
{
	last_pt.X = X;
	last_pt.Y = Y;
}

void graphics_window::do_draw_from_last_point(int X, int Y)
{
	XDrawLine(dpy, id, gc, last_pt.X, last_pt.Y, X, Y);
}

void graphics_window::do_motion_notify(XEvent *eventP)
{
	if (eventP->xmotion.state = Button3Mask)
		{
			last_pt.X = eventP->xmotion.x;
			last_pt.Y = eventP->xmotion.y;

			do_draw_from_last_point(last_pt.X, last_pt.Y);
			do_set_last_point(last_pt.X, last_pt.Y);
		}
	return;
}

void grphics_window::do_button_press(XEvent *eventP)
{
	last_pt.X = (eventP->xbutton.x);
	last_pt.Y = (eventP->xbutton.y);

	unsigned int but = eventP->xbutton.button;

	if (but != Button1)
		{
		switch (but)
		{
			case Button2:
			     do_set_last_point(last_pt.X, last_pt.Y);	
			     return;

			case Button3:
                             do_draw_from_last_point(last_pt.X, last_pt.Y);
			     return;
		}
		}

	XUnmapWindow(dpy, id);
	XDestroyWindow(dpy, id);

	XCloseDisplay(dpy);

	exit(1);
}

void graphics_window::do_handle_events()
{
	XEvent event;

	for ( ; ; )
		{
			XNextEvent(dpy, &event);
			switch (event.type)
			{
				case MotionNotify:
					do_motion_notify(&event); break;

				case ButtonPress:
					do_button_press(&event); break;
			}
		}
}


static int main()
{
	dpy = XOpenDisplay(0);

	if (!dpy) { printf("Display not opened !\n"); }

	scr = XDefaultScreenOfDisplay(dpy);

	graphics_window gw;       /* declaration of gw as being of class
    				     graphics_window */

	gw.do_handle_events();

	delete gw;
}

    
Thanks in advance, any other helpful suggestions would also be appreciated.
    
    Marcus.

T.R	Title	User	Personal Name	Date	Lines
754.1	function protos?	PCINT::MERRILL	Brad Merrill DECtp SWE	`Wed May 10 1989 12:19`	11
	At a guess, the "declaration" error message seems to imply that it hasn't found a function prototype for the Xlib function in question. I did a quick search of DECW$INCLUDE:*.H and there are "extern func()" statement, but these may not be sufficient for c++. I would suggest providing explicit function prototypes for the Xlib routines that you reference. /Brad
754.2	Since some of us may be starting a new project...	IO::MCCARTNEY	James T. McCartney III - DTN 381-2244 ZK02-2/N24	`Thu May 18 1989 03:10`	5
	Would you mind sharing where and how you got a C++ compiler? James
754.3		60428::QUODLING	Just a Coupl'a days....	`Thu May 18 1989 11:39`	7
	re .2 Isn't GNU (aka Free Software Foundation) releasing one. q
754.4	C++ - where	37404::bonini	You are only coming through in waves...	`Thu May 18 1989 11:54`	13
	Check the C++ notes conference (cadsys::c_plus_plus) for locations and experiences. Lots of info there. Quickly, there are two you can get. cfront is a C++ preprocessor distributed by AT&T. Someone internally can get it for you but it costs money since it's not ours. The GNU C++ (g++) is free, on the net, and a compiler instead of a front-end. Again, info on both in the aforementioned notes conference.
754.5	Who maintains the compiler?	IO::MCCARTNEY	James T. McCartney III - DTN 381-2244 ZK02-2/N24	`Thu May 18 1989 18:26`	6
	But do you really want to build production software based on freeware? James
754.6	You're referring to X :-)	STAR::BRANDENBERG	Si vis pacem para bellum	`Thu May 18 1989 18:27`	2

754.7	Supported is not necessarily better...	7486::WITHROW	Robert Withrow	`Fri May 19 1989 15:06`	6
	Besides, the Gnu C compiler is known to be more reliable and in fact ``better'' in almost every way than the ``supported'' compiler. Also, should you discover a compiler bug, I'll bet you can fix it yourself much faster than to wait for a fix via the QAR method.
754.8	Unsupported is never better though...	IO::MCCARTNEY	James T. McCartney III - DTN 381-2244 ZK02-2/N24	`Sat May 20 1989 13:58`	25
	re: .-1 What happens when a customer using you code, compiled through a freeware compiler, finds a bug, induced by bad code in the compiler and because of the damage it causes sues Digital. Are you prepared to say: ``I thought that someone else's compiler is better than anything we can build.'' Exactly who did the tests of the Gnu compiler, where are the regression tests suites, and who has checked for conformance to Digital's quality standards? I'm just not ready to ``bet my business'' nor my customer's on someone else's unsupported mistakes. And as to fixing the compiler if it has a bug, that's not what I'm being paid for. My manager would not like it if I when to him and said, ``Sorry, we're slipping 3 weeks while we figure out what's wrong with our unsupported compiler.'' A discussion about the salary continuation plan would most certainly ensue. James
754.9	NIH at work again	30790::ROSE		`Sat May 20 1989 16:05`	26
	Re .8: The problems you mention hold for all software not developed by DEC, be it 'supported' or not. You are making a generic "not invented here" pitch to do everything in house. The problems you mention are also possible with the MIPS C compiler, or with the parts of Ultrix we got from U. C. Berkeley and AT&T and Sun, or with all the other software we buy... What happens when a customer using your hardware, whose gate arrays were designed using Valid Logic Systems' GED schematics editor, finds a bug, induced by bad code in the schematics editor, and because of the damage it causes sues Digital. Are you prepared to say: "Valid Logic Systems is a leader in electronic CAD technology and for this reason we relied on their schematics editor rather than using our own (SUDS)?" [For those of you unfamiliar with hardware design: DEC has about 1000 GED licenses from Valid. GED is to many DEC hardware engineers as LSE is to VMS software engineers. SUDS was an in-house predecessor to GED which ran on DECsystem-20s using VS-11 graphics terminals.] Your example shows a 3 week slip to fix a compiler bug. Does DEC respond to compiler SPR's with a fix in less than three weeks? Having the sources to fix bugs yourself can be a big advantage...
754.10	(standard disclaimer: I'm not a lawyer)	56733::MESSENGER	Bob Messenger	`Sat May 20 1989 18:24`	19
	Re: .9 > Re .8: The problems you mention hold for all software not developed by > DEC, be it 'supported' or not. No, they don't. There is a disclaimer on the GNU software that specifically states that the software is unsupported, i.e. the Free Software Foundation accepts no responsibility for its quality. > What happens when a customer using your hardware, whose gate arrays > were designed using Valid Logic Systems' GED schematics editor, finds a > bug, induced by bad code in the schematics editor, and because of the > damage it causes sues Digital. We sue Valid Logic Systems. We couldn't sue the Free Software Foundation because GNU is unsupported. -- Bob
754.11		PSW::WINALSKI	Paul S. Winalski	`Sun May 21 1989 17:24`	22
	RE: .10 If a customer finds a bug in shipped code, it doesn't matter whether it got there due to bad logic at the source level or a compiler bug. The fact that it slipped out the door in a product release indicates inadequate testing of the product before it was shipped. The fault is entirely the development team's. GNU C or other "freeware" is no different from a home-grown utility or software tool that a development team writes to get their job done. In both cases (freeware and the home-grown tool), the development team has the sources and has assumed the responsibility to fix problems in the tool, if they arise. It's a trade-off: if you use the DEC compiler, you don't have to fix problems that you find, but the fix might not be as timely as you'd like. If you use the freeware compiler, the bug fixes can be as timely as you like, but you must make them yourself. --PSW P.S.: Discussions of scenarios involving lawsuits are not appropriate for NOTES conferences. You wouldn't want your notes presented in court as evidence that DEC itself admits it is at fault in such circumstances, would you?
754.12	If you want to stake your career on a lottery ticket...	56733::MESSENGER	Bob Messenger	`Sun May 21 1989 17:31`	8
	Re: .11 In that case (i.e. if it's the development team's fault if something goes wrong) no project should use freeware unless they also budget the time to test it and maintain it. Otherwise they're taking a foolish gamble. -- Bob
754.13	I use GCC under Ultrix.	56564::thomas	The Code Warrior	`Sun May 21 1989 17:50`	19
	[This note was written from a program compiled by GCC] GCC allows me to make asure my code works in a ANSI C environment and GCC also catches many of the common mistakes that programmers make. It is simply another tool for maintaining/developing the software that I work on. I also use the MIPS C compiler, the standard Ultrix compiler, and the VAX C compiler in my work. The support of GCC is a non-issue. I use GCC for development and debugging but I use pcc/mcc to generate the production code. Just because I use freeware, does not mean that I am dependent on it. I can switch to another tool if I need to. But I do agree that becoming dependent on freeware with having someone officialy responsible to support it is stupid. It is my goal to make sure all my programs work correctly in a ANSI C/POSIX 1003.1 environment. The only tool that I can use to verify this is GCC until the time comes along that Digital supplies me a ANSI C compiler.
754.14	End of rathole (closeing statement)	7486::WITHROW	Robert Withrow	`Mon May 22 1989 00:19`	38
	re: <<< Note 754.8 by IO::MCCARTNEY "James T. McCartney III... Since I was the one to start this rathole, I feel I should get to post my rejoinder before terminating it: > What happens when a customer using you code, compiled through a > freeware compiler, finds a bug, induced by bad code in the compiler and > because of the damage it causes sues Digital. Are you prepared to say: > > ``I thought that someone else's compiler is better than anything we can > build.''" What happens when a customer using you code, compiled through the VAX C compiler, finds a bug, induced by bad code in the compiler and because of the damage it causes sues Digital. Are you prepared to say: ``I thought that our compiler is better than anything else in the world.'' " (These what-happens-when-you-get-sued retorts get so tiresome!) I have SPRd bugs in the C compiler (both inside and outside of digital) and have generally gotten "Thank you for your SPR. This bug will be fixed in a future release." I'm not carping at the compiler folks; They have their priorities and I have mine. To them, my bug is likely just one more thing that needs to be fixed; to me it likely the most importing thing. Who pays for the "extra three weeks" to work around the compiler bug? What if I could fix it myself in a couple of days? Does your manager really care if the three weeks is due to person A's bug or person B's bug? The choice of compiler is just one more issue that requires "engineering tradeoffs". If using a non-Dec or freeware product frightens you then dont use one. But don't kid yourself that you have automatically eliminated your cost and risk. ========== And with this I formally declare this rathole not merely dead, but really most sincerely dead!
754.15		8088::MERRILL	Brad Merrill DECtp SWE	`Thu May 25 1989 15:38`	10
	There is an internal C++ compiler available, see the VAX C++ conference. No, its probably not a good idea to do a product based on it "today", but there is no reason that advanced development couldn't happen, with actual product dates deffered until the supported version of C++ is released. We used to call this paralell development. It would probably help the C++ project if they knew there were groups willing to do active development with it. /Brad