Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Hourglass save-state issues are known on Windows 8, we don't know why it happens yet, but I am intending to make a Windows 8.1 machine later so that I can test it.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Dyshonest wrote:
What I'm trying to do is likely not a world record, but I know emulation can do some very weird things, and I remember from the Cheetahman 2 TAS that emulators always start at the same seed or something? Basically I'd just like to make sure things are still possible to do on a real console.
There is less worry about starting on the wrong seed with gameboy games as the bootloader sets all memory to 0 during boot. The only thing you may have to worry about regarding randomness is hardware clocks emulation and interrupt timings, both which at least Gambatte should get correctly enough.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
adelikat wrote:
IF you exploit Nintendo's VC bugs (which btw are not relevant here, this is an easy to emulate NES game that both emulators emulate flawlessly), then I see that as exploiting an emulator bug. Just because it is an official emulator doesn't mean it isn't an emulator bug. And those aren't (and shouldn't) be allowed here.
But if it is officially endorsed by Nintendo, means that technically the emulator is part of the game, and thus a game-glitch. That was the point I was trying to make.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Since I am the only one being extremely OCD about this I should try to explain my point a bit further I assume.
The way I see this thing is that Nintendo's emulator that is bundled with the ROM when it's in the WAD packaging is not the emulator used to record the TAS.
In my world this is the same principle as using a non-accepted emulator to make the TAS, and then try to have it published. Dolphin (or whatever the 3DS emulator with TAS tools will be called when that gets developed) would be the appropriate emulator here. The NES emulator in the WAD could theoretically have bugs in it that could theoretically produce a better time, and thus using this emulator would then be using the "buggy" emulator in this scenario, as it would then prevent glitches that would be possible using the bundled one and TAS it with Dolphin.
I suppose it could also be seen as an Unreleased submission, but in my world, it would require the game to actually be unreleased.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Am I the only one who's iffy about this submission based on how it was recorded?
This was never released for the NES, even if it's a completely legitimate NES ROM. We don't know if we can with this setup push the game beyond the limitations of the hardware it was intended for (Wii VC).
A comparison to using an emulator with emulation bugs and abusing those bugs to save time can in my opinion be made here.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
jlun2 wrote:
How similar is this to WINE?
The DLL uses kinda the same principle of replacing WinAPI calls.
After that how things are handled are completely different from WINE, specially since WINE enables Windows executables on Linux with no intention of being able to record things or have deterministic behavior.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Nach wrote:
Warepire wrote:
AMD Athlon64 4000+ ClawHammer core. (Yes, that old)
Unfortunately that lacks SSE3, but we can definitely optimize for it.
Are you running 64 bit Windows?
Yes, I am however getting a new PC in time for summer (need it for Hourglass development). So that would just be a lot of wasted effort to optimize for this outdated piece of tech.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Nach wrote:
Regarding DTC6, how does starting in March sound?
How do you guys feel about using lsnes for DTC6?
My current computer is not strong enough for SNES games. It ran Metal Force at 40 fps in NesHawk.
So if it's a SNES game I might not participate.
It's not like anyone would notice though.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
So much we never found ...
Un-hiding our LUA HUD: http://tasvideos.org/userfiles/info/19095439324807403
It does not work where the invulnerability timer doesn't change the health also. I never got around to fix that since Tompa doesn't like HUDs and I thought it was good enough.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
feos wrote:
Curious about the reasons.
GitHub is a bigger community, hopefully we can attract new developers this way.
We also felt like we needed to ditch SVN and move to Git, along with that Google Code just doesn't allow as much stuff as GitHub does.
It should improve the workflow for us who are currently developing as well.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
The music in this version makes my ears bleed, and the conveyor belts make my eyes ache. Which is why I pushed for the DOS version. Anyway, enough whine.
The progress you made is fantastic, and once muted it's not boring for one bit (pun intended).
Everyone thought I was expecting too much from a CC TAS when I thought it could improve some of the RTA times. 2 levels improved already! I do hope you manage to find an improvement in Blobnet, that level deserves it so hard!
Keep it up!
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
jonese1234 wrote:
Warepire wrote:
jonese1234 wrote:
Erm i was wondering if this was ever going to be fixed?
http://puu.sh/eqgbl/282ebfaff1.png
Or has a fix, or work around!
I have a way of doing it however its very inconsistant, and still crashes oftern especialy when trying to load save states!
That will be fixed before v1 is released, along with some of the save/load state crashes. The current save/load state code is the same as it was in original hourglass, so it will crash just the same as well. We have a partial fix for a few of them in a branch.
Any idea how long that will be? even just to get a stable version? AS the save state problem is the only one holding me back from really tasing heavy as it is always annoying when you lose work because of the crash.
When I have enough free time or when someone steps up and chooses to help out. Whichever comes first. Finding interested developers is really hard.
We're currently moving to GitHub, that shall improve the ability to get contributions though.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
jonese1234 wrote:
Erm i was wondering if this was ever going to be fixed?
http://puu.sh/eqgbl/282ebfaff1.png
Or has a fix, or work around!
I have a way of doing it however its very inconsistant, and still crashes oftern especialy when trying to load save states!
That will be fixed before v1 is released, along with some of the save/load state crashes. The current save/load state code is the same as it was in original hourglass, so it will crash just the same as well. We have a partial fix for a few of them in a branch.
fsvgm777 wrote:
This needs to be looked at.
(yes, I consider it a bug, because the text should look like in the former screenshot in the post I linked to)
I don't know what kind of issues that nitsuja saw, but I am not sure a full revert is the way to go. Will have to look at that code more in detail later. I cannot promise this will be fixed before v1.
