anon@dip.eecs.umich.edu (Omar S. Juma) (12/17/89)
Hello, DEC. Any word on when the tile-fill bug in the X server will be fixed? I'm refering to the bug with non-multiple-of-8-width pixmaps. As in: if you have a root pixmap whose width is not a multiple of 8, you'll end up with junk on your root. Several applications create or use such pixmaps. The effects, I must admit, are rather unpleasant. ...anon... Omar S. Juma EECS Department University of Michigan
jg@max.crl.dec.com (Jim Gettys) (12/18/89)
The X server? We only sell X servers on around a dozen different displays, on 3 different machine architectures, under 3 different operating systems. How about giving us a hint of what kind of display on what kind of machine, using whose X server (Digital's or MIT's), on which operating system? Or is that asking too much? - Jim
anon@dip.eecs.umich.edu (Omar S. Juma) (12/18/89)
In article <1439@crltrx.crl.dec.com> jg@max.crl.dec.com (Jim Gettys) writes: >The X server? We only sell X servers on around a dozen different >displays, on 3 different >machine architectures, under 3 different operating systems. > >How about giving us a hint of what kind of display on what kind of >machine, using >whose X server (Digital's or MIT's), on which operating system? Or is >that asking too much? > - Jim Sorry, the original message got a bit scrambled after a sleepless night. Here's the missing golden coins: Hardware: DECstation 3100, 16 MB RAM, 2 x RZ55 drives 19" color monitor standard DEC keyboard (LK201) & mouse (puffed-up hockey puck) Server: Xcfb from mid-spring (mod date on disk is May 1, which is fairly close to when my workstation was configured). Server came from DEC, since this was a demo machine (more precisely, from Palo Alto). OS: Ultrix 3.0 Symptoms: pixmaps of width not a multiple of 8 cause junk to appear in area where pixmap is repeatedly tiled, after first tile on left of area. When the root is tiled with such a pixmap, the root appears clean until you move a window or pop a menu, etc. Then you get window residuals on the root. A variety of similar alignment errors occur, all of the same flavor. I haven't had much time to test this on other DEC platforms on campus (no time anymore for such luxuries), but my guess is it's a MIPS-based bug. The bug may have been fixed by now, but I haven't heard anything official regarding a patch since I started using my DEC3100 in September. Other than this bug, it's one of the better servers I've worked with. I *believe* one of our labs has DECstations running Ultrix 3.1, and I've seen the problem repeated there as well. Of course, we're all anxiously awaiting the 4.0 release, when shared libraries finally make it into the OS. :-) ...anon... P.S: Here's a bitmap that definitely upsets my server: #define g0_width 51 #define g0_height 46 #define g0_x_hot 0 #define g0_y_hot 47 static char g0_bits[] = { 0x2e, 0x99, 0xb9, 0x64, 0xba, 0x64, 0x02, 0xc9, 0xa5, 0x27, 0x95, 0x25, 0x97, 0x06, 0xff, 0xff, 0xfe, 0xff, 0xfd, 0xff, 0x07, 0x22, 0x19, 0x89, 0x64, 0x8a, 0x64, 0x04, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07, 0x89, 0x49, 0x26, 0x93, 0x24, 0x25, 0x01, 0xff, 0xff, 0xff, 0xff, 0xfd, 0xff, 0x07, 0x2e, 0x99, 0x99, 0xac, 0xbf, 0x6c, 0x02, 0xc9, 0x66, 0x66, 0x93, 0x24, 0x93, 0x07, 0xff, 0xff, 0xfe, 0xff, 0xff, 0xff, 0x06, 0x22, 0x91, 0xa9, 0x64, 0xaa, 0x24, 0x01, 0xff, 0xff, 0xff, 0xff, 0xfb, 0xff, 0x07, 0xc9, 0x6c, 0x26, 0x97, 0x26, 0xb3, 0x01, 0x2e, 0x9b, 0xd9, 0x64, 0xd9, 0x6c, 0x06, 0xff, 0xfd, 0xbf, 0xff, 0xbf, 0xdf, 0x07, 0xc9, 0x24, 0x25, 0x91, 0x24, 0x93, 0x00, 0xff, 0xff, 0xff, 0xdf, 0xff, 0xff, 0x07, 0x26, 0x99, 0xa9, 0x64, 0x9a, 0x64, 0x06, 0xff, 0xff, 0xbe, 0xff, 0xfb, 0xfe, 0x07, 0xd9, 0x65, 0x6d, 0x57, 0x46, 0x97, 0x01, 0x26, 0x99, 0xa9, 0xa4, 0xa9, 0x64, 0x06, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07, 0xc9, 0x24, 0x25, 0x91, 0x24, 0x93, 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07, 0x2e, 0x9b, 0xba, 0x64, 0xda, 0x6c, 0x06, 0xc9, 0xe4, 0x26, 0x97, 0x25, 0x93, 0x02, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07, 0x22, 0x99, 0x88, 0x64, 0x8a, 0x24, 0x06, 0xff, 0xff, 0xff, 0xff, 0xf5, 0xf7, 0x03, 0xd9, 0x24, 0x33, 0x93, 0xce, 0x9c, 0x04, 0x2e, 0xdb, 0xcc, 0xac, 0x3b, 0x63, 0x05, 0xff, 0xff, 0xff, 0xff, 0xee, 0xfe, 0x07, 0xc9, 0x24, 0x25, 0x93, 0x24, 0x93, 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xdf, 0x07, 0x22, 0x99, 0x89, 0x64, 0xaa, 0x64, 0x02, 0xff, 0xff, 0xfe, 0xff, 0xff, 0xff, 0x07, 0xc9, 0x65, 0x25, 0x97, 0x45, 0x97, 0x01, 0x2e, 0x99, 0xdb, 0x64, 0xba, 0x64, 0x06, 0xff, 0xff, 0xfe, 0xff, 0xff, 0xff, 0x07, 0x49, 0x24, 0x25, 0x91, 0x24, 0x91, 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07, 0x2e, 0x59, 0xb9, 0x64, 0xba, 0x64, 0x07, 0xc9, 0x65, 0x27, 0x97, 0x25, 0x97, 0x01, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07, 0x22, 0x19, 0xc9, 0x64, 0x8a, 0x64, 0x06, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x07}; P.P.S: For a more vivid demonstration, run GWM with MWM emulation mode right off the distribution tape. The titlebars and window borders get striped for precisely the same reasons (for the longest time I thought it was a French twist on the Motif look, but after looking at the MWM emulation files, I saw "pixmap" all over the place, and my parrot started shrieking).
jg@max.crl.dec.com (Jim Gettys) (12/18/89)
You should currently be running Ultrix 3.1/UWS 2.1. Your bug is indeed lots of fun, and seems to exist both on UWS 2.0 and 2.1. I just tried your bitmap under UWS 2.2, and the server works correctly. UWS 2.2 should be out of the software distribution center sometime in the next few weeks; it was announced a month or two ago. - Jim