waterman@cory.Berkeley.EDU (T.S. Alan Waterman) (10/15/87)
Regarding this talk about multiple screens dying all over in a big way: About a month ago, I ran into a great display bug (feature ?-). I was running everything in non-overscan screens. Interlaced workbench, interlaced DBW VT100 (v2.2 -- anything higher tends do do nasty things to my system--haven't bothered to track it down yet), and two show (Fish disk 72 ??) screens, one full-blown HAM, the other sixteen color interlaced. Flipping through these, with the mouse and a little help from A-N and A-M, I got a display with the top half being the top half of the HAM picture, only it had the colors from the other show picture. This is where things start to get strange: the bottom half had EIGHT replicas, in miniature, of the hi-res picture, with the WORKBENCH colors!! Flipping up to another screen, it went away; so, I chased it. With the two show screens as the highest, sliding one down across the other would send the display into random hysterics, with little miniature images in the wrong colors everywhere. Once it went into that near_death syndrome called fireworks_display_mode in the manuals. Alright, so this is wierd, yeah, but who cares? Well, the thing is, try as I might, the machine would not crash from this. It was (is) infinitely repeatable, but it _never_ crashes the machine! Anybody have any clues? I wonder whether it's an Intuition bug that can happen even if you do everything `right', or if it's something peculiar (very) to Show. Anyone seen it quite like this? --T.S. =======the Truth, the (w)hole Truth, and nothing but the Truth====== =======I don't really care who's opinions these are. Do you? ======
ali@rocky.STANFORD.EDU (Ali Ozer) (10/15/87)
In article <4374@zen.berkeley.edu> T.S. Alan Waterman writes: > [Talking about multiple screens dying] > ... Alright, so this is wierd, yeah, but who cares? Well, the thing is, try >as I might, the machine would not crash from this. It was (is) infinitely >repeatable, but it _never_ crashes the machine! Anybody have any clues? > I wonder whether it's an Intuition bug that can happen even if you >do everything `right', or if it's something peculiar (very) to Show. Well, at least the "minimal" form of the bug (which Tom mentioned in an earlier msg: interlaced WB + interlaced screen + non interlaced screen + front-to-back + front-to-back => FIREWORKS) seems to be a system problem, and can crash the machine if you play with the mouse (instead of using A-N/A-M). Tom and I tried it on a second Amiga (512K, as opposed to Tom's 1.75M), and sure enough, same results. We tried the same thing with various other programs that open custom screens, such as Viewilbm (not Show), TeX previewer, Fed, and same results all the time. Then I tried it on a 2.5M machine, and once again, same results. So the programs aren't to blame, probably Intuition is... Seems like when you depth arrange screens of different view-modes the copper list fails to get assembled correctly (maybe?). (In all our tests the colors of the messed up screen seemed right --- Usually the view modes and the bit maps were screwed up.) Ali Ozer, ali@rocky.stanford.edu
ali@rocky.STANFORD.EDU (Ali Ozer) (10/15/87)
[] One more observation: From an interlaced WB, open up as many interlaced TeX Previews as you want. Things are fine --- You can front-to-back to your heart's content, no fireworks. But then simply "uninterlace" one of the preview screens (TeX previewer is so nice it lets you switch between non-interlaced and interlaced screen on the fly), front-to-back to WB, and front-to-back to an interlaced preview, and BLAM! The system can handle many interlaced screens, so the fact that it can't handle 2 interlaced and one not seems to be a silly bug (Dale? Jim?) rather than some SYSTEM LIMITATION. (Imagine the TV ads: You can open upto ONE noninterlaced and TWO interlaced screens at once! 6 simultaneous tasks! 4 windows on desktop! 40 directories per volume! Amiga, the machine with the unreachable limits!) [Aren't we glad this is not the case. 8-) ] Ali Ozer, ali@rocky.stanford.edu