sherouse@godot.radonc.unc.edu (George W. Sherouse) (10/21/87)
Section 3.2.1 of the X11 Xlib manual tells us, "The background pixmap and the window must have the same depth." God, I hate quoting manuals. As previously reported (by me) to xbugs, the V11R1 XMenu does not work on n>1-plane displays for this very reason. This week's release of X11 xfish (one of my favorite X applications, by the way) has the same problem. This suggests that most of you are using monochrome displays. At the risk of sounding elitist there is a real client portability problem. In its current state XSetWindowBackgroundPixmap is an accident waiting to happen. In my fix to XMenu I used a loop over planes to XCopyPlane from the bitmap to the pixmap. I am still not convinced that that is the best way to expand a bitmap, but that aside it seems clear to me that a new function or macro is needed to prevent ongoing transportability problems. I would propose either XSetWindowBackgroundBitmap (functionally identical to XSetWindowBackgroundPixmap but taking a bitmap as an argument and expanding it internally) and/or XMakePixmapFromBitmap (to standardize the expansion technique) be added to Xlib. The documentation of XSetWindowBackgroundPixmap should also reiterate the requirement that the Pixmap and Window be of the same depth. Meanwhile, please folks try to remember that XSetWindowBackgroundPixmap requires a pixmap of the same depth as the window and that that depth is not necessarily 1. - George
jkh@violet.berkeley.edu (Jordan K. Hubbard) (10/21/87)
In article <885@godot.radonc.unc.edu> sherouse@godot.radonc.unc.edu (George W. Sherouse) writes: | |In my fix to XMenu I used a loop over planes to XCopyPlane from the |bitmap to the pixmap. I am still not convinced that that is the best |way to expand a bitmap, but that aside it seems clear to me that a new |function or macro is needed to prevent ongoing transportability |problems. I would propose either XSetWindowBackgroundBitmap |(functionally identical to XSetWindowBackgroundPixmap but taking a |bitmap as an argument and expanding it internally) and/or |XMakePixmapFromBitmap (to standardize the expansion technique) be |added to Xlib. The documentation of XSetWindowBackgroundPixmap should Am I missing something? I thought that bitmaps finally went away in V11 (I haven't used them yet, so I haven't gotten to that section of the manual) and were replaced by depth 1 pixmaps. If this isn't the case, why? I liked the idea of only one data type. Jordan Hubbard
RWS@ZERMATT.LCS.MIT.EDU (Robert Scheifler) (10/21/87)
Date: 21 Oct 87 13:51:50 GMT
From: godot!sherouse@mcnc.org (George W. Sherouse)
In my fix to XMenu I used a loop over planes to XCopyPlane from the
bitmap to the pixmap. I am still not convinced that that is the best
way to expand a bitmap
It certainly isn't. One of the uses of CopyPlane is to fill a pixmap
from a bitmap in a single operation. Perhaps the "exactly one plane"
sentence in the Xlib manual confused you. The bit-plane you specify is
the source bit-plane, not the destination bit-plane. (For a bitmap
there is only one plane to choose from, but CopyPlane is more general in
allowing you to select an arbitrary bit-plane from a multi-bit source.)
If you have a bitmap, and you want to fill a pixmap with a two-color
image of it, set the foreground/background in a GC to those two colors,
set the plane-mask to all ones, the function to Copy, and do a CopyPlane
with a bit-plane of 1. No muss, no fuss. (Alternatively you chould
store the bitmap as the stipple of the GC and use fill-style
OpaqueStippled in a FillRectangle, but that's more work. Also
alternatively you could do a PutImage with format XYBitmap directly into
the pixmap, rather than creating the bitmap resource in between. It
probably would be nice if Xlib had an XCreatePixmapFromBitmapData to
simplify this.)black@masscomp.UUCP (Sam Black) (10/22/87)
Summary: Expires: Sender: Followup-To: Distribution: In article <885@godot.radonc.unc.edu> sherouse@godot.radonc.unc.edu (George W. Sherouse) writes: > >In my fix to XMenu I used a loop over planes to XCopyPlane from the >bitmap to the pixmap. I am still not convinced that that is the best >way to expand a bitmap, but that aside it seems clear to me that a new >function or macro is needed to prevent ongoing transportability >also reiterate the requirement that the Pixmap and Window be of the >same depth. > >... I would propose either XSetWindowBackgroundBitmap >(functionally identical to XSetWindowBackgroundPixmap but taking a >bitmap as an argument and expanding it internally) and/or >XMakePixmapFromBitmap (to standardize the expansion technique) be >added to Xlib. There is really no reason to iterate over planes as Mr. Sherouse describes. For what it's worth: Pixmap XMakePixmapFromBitmap(dpy, bitmap, width, height, depth, fg, bg) Disply *dpy; Pixmap bitmap; /* source pixmap (depth 1) */ int width, height; /* normally same as bitmap dimensione */ int depth; /* normally DefaultDepth(dpy, DefaultScreen(dpy)) */ int fg, bg; /* foreground and background colors */ { GC gc; XGCValues gcv; Pixmap pixmap = XCreatePixmap(dpy, width, height, depth); if (pixmap) { gcv.foreground = fg; gcv.background = bg; gcv.graphics_exposures = 0; gc = XCreateGC(dpy, pixmap, (GCForeground | GCBackground | GCGraphicsExposures), &gcv); XCopyPlane(dpy, bitmap, pixmap, gc, 0, 0, width, height, 0, 0, 1); XFreeGC(dpy, gc); } return(pixmap); } - sam black -------------------------------------- I'm pink, therefore I'm Spam. ...!{cca,decvax,seismo}!masscomp!black UUCP black%masscomp.uucp@<internet gateway> Internet --------------------------------------
hania@DECWRL.DEC.COM (10/23/87)
George Sherouse writes: I would propose either XSetWindowBackgroundBitmap (functionally identical to XSetWindowBackgroundPixmap but taking a bitmap as an argument and expanding it internally) and/or XMakePixmapFromBitmap (to standardize the expansion technique) be added to Xlib. And Robert Scheifler writes: It probably would be nice if Xlib had an XCreatePixmapFromBitmapData to simplify this. I, too, have proposed adding XCreatePixmapFromBitmapData -- in fact, I offered to do it a couple of days before the release, but was told, quite reasonably, that it was too late to put this in to the release. I will write the routine and investigate having it added to Xlib; I am saying this with my X contributor hat on, since I am getting a lot of flack about continual changes to "frozen software" from my employers... I will have to consult them first. Hania Gajewska DEC WSL