pouët.net

Hannibal information 79 glöps

• general:
• level: user
• personal:
• first name: Soren
• last name: Hannibal

• demo Amiga OCS/ECS voyage by Razor 1911 [web]
• Loved the music and the code. Style seemed a bit inconsistent but had lots of good elements
• rulezadded on the 2016-01-03 18:55:06
• demo Amiga OCS/ECS State of the Art by Spaceballs [web]
• I didn't like it at the time because it didn't feel like a "real demo". In hindsight it was a great demo that expanded what a demo could be.
  (I also didn't like it because it beat my own demo :-)
• rulezadded on the 2016-01-03 18:53:15
• demo Amiga OCS/ECS Hardwired by Crionics & The Silents [web]
• This demo significantly raised the bar for demos at the time - consistent design and colors, impressive effects (at the time), stylish intro, efficient use of 1mb, lots of attitude. A shame it didn't win that competition
• rulezadded on the 2016-01-03 18:51:09
• demo Amiga OCS/ECS 9 Fingers by Spaceballs [web]
• I don't have anything against "animation player demos" (and did one myself :-). this one is technically brilliant due to the super impressive image processing tools used to generate the greyscale anims. The actual demo not too interesting to me
• rulezadded on the 2016-01-03 17:16:37
• demo Amiga OCS/ECS Extension by Pygmy Projects
• Great music, fast pace, solid timing and nice effects means this is still one of my favorites
• rulezadded on the 2016-01-03 17:10:56
• demo Atari STe We Were @ by Oxygene [web]
• Very impressive
• rulezadded on the 2015-12-24 06:34:57
• demotool Amiga AGA Amiga OCS/ECS WinUAE Demo Toolchain 5 by Lemon.
• I could repro the blank screen bug, but only if I
  1. used the plain blueberry startup code (not my example demo, which was fixed), which doesn't enable screen DMA but assumes it is enabled
  2. used only with kick 3.x, not with kick 1.x, because with kick 3.x it doesn't enable screen until something triggers a CLI window update, and
  2. used the "-debug" flag (which disabled the compression and therefore didn't print anything on screen).
  
  I don't think this is what you saw in your demo. Nevertheless, for next version, I will try to make it print something on screen at bootup, to not trigger this.
• isokadded on the 2015-11-28 06:15:31
• demo Amiga OCS/ECS Wildcat by DMA [web]
• KK, I tried to debug a bit further (this is kinda fun!) and saw that yes, your code actually does write to the proper registers
  However, at that point the copper DMA is still enabled, and the system copper is still set, so the next frame the system copper runs, and it overwrites those values. Even though your vblank interrupt sets a new copper list, the system copper has already been started.
  
  I imagine it works with some kickstarts because they set them to the same values you need. You could fix it by for example either
  
  1. instead of setting the registers with the CPU, you could put it in a "setup" copper list and load that into COP1LC. It would automatically be kicked the next frame you run. Optionally you could kick it right away
  2. turn off copper DMA before you set the registers, set your real copper, and then reenable the dma or
  3. just set them in the first "real" copper list you have. That's what I have traditionally done
  
  Side note: I see that after setting the copper (in dff080), you do clr.w dff088. Probably fine for you, but there are 2 details to know when you do this:
  1. writing to cop1jmp/cop2jmp from the CPU is dangerous if the blitter is running because it can mess up the blitter destination.
  2. The copper automatically starts at the beginning of each frame, so if you just set the copper pointer during the previous frame you don't need this
• isokadded on the 2015-11-26 07:52:17
• demo Amiga OCS/ECS Wildcat by DMA [web]
• Hey KK, I looked at this, and it seems that you are not setting DDFSTRT/DDTSTOP/DIWSTRT/DIWSTOP/BPLCON2/BPLCON3, and those are probably initialized differently in the various setups.
  I found this by launching 2 instances of WinUAE - one working and one not working, and pressing shift+F12 to see in the debugger - you can see that those registers are different. The values you need for a 320x256 screen are
  
  DIWSTRT($008e) = $2c81
  DIWSTOP($0090) = $2cc1
  DDFSTRT($0092) = $0038
  DIWSTOP($0094) = $00d0
  BPLCON2($0104) = 0
  BPLCON3($0106) = $0C20
  
  I'm not 100% sure the last two are needed, but these days I prefer to set all registers to cover myself. Try this and hopefully it helps (in the WInUAE debugger you could test it by directly writing to the registers with commands like this: W dff08e 2c81.w)
  
  Other than that, really good job on the demo. I will check if my example demo has the proper setup, too.
• rulezadded on the 2015-11-25 18:15:33
• demotool Amiga AGA Amiga OCS/ECS WinUAE Demo Toolchain 5 by Lemon.
• 7. I can't reproduce the problem, KK. I attempted these 2 things:
  runa4000_60.exe exampledemo\out\exampledemo.exe (both AmigaForever or regular roms)
  or
  toolchain\startwinuae.exe a500 kick31 exampledemo\out\exampledemo.exe (the override only works if you DONT have amigaforever roms)
  Both ran fine
  
  Which exact roms are you seeing this on? There are multiple different Kickstart 3.1 out there (amigaforever lists different ones for a600 vs a1200 vs a4000 etc)
  And what are your repro steps?
• isokadded on the 2015-11-25 17:53:33