Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
What the actual hell?
It's a terrible idea to hire people to do work on a project you're not a part of, specially when those hired people will not communicate with the actual developers.
1) This is a very complex type of software which require deep understanding of Windows, process initialization, and concepts such as debuggers, emulators and code injection.
Why: Hourglass contain all the above concepts, and without the understanding of the above, you cannot understand the impact of the code you write. Hell, there are even more areas needed for some code, like low level graphics pipelining, sound processing, saving and restoring state of external hardware (not the CPU), etc.
2) The underlying problems as to why modern games don't work, are fundamental design issues, and will be very hard to get to work across computers.
Why: A lot of the problems that have to be solved (including clock, graphics, input, memory handling), all require a vast amount of time to get even remotely functional.
3) Any people capable of solving the above problems will either do it for free as a hobby / learning experience by finding me on their own, or they will charge a lot more than what they can earn through Fiverr (meaning, a full freelance software development contract, which provides a competitive salary for the field, easily over 3500 USD a month).
Why: The improvements I've done over the course of the last two years, took that amount of time, because solving those problems were very complicated. Some solutions I still can't add properly due to how the rest of Hourglass must change before the solution can cooperate with the rest of the code.
I highly doubt this will do anything but cause trouble and/or drama.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
Alyosha wrote:
Medium: Finding object tables: Usually enemy/sprite/player data used by a game is clumped into tables in RAM, if we could automatically get lists of object slots, hit boxes, attributes, and such that could save a lot of time.
Isn't the hitbox data usually packed in the ROM. Then values in RAM together with some look-up functions in the game code are used to determine the hitbox data for each entity in the game? I am not sure the extraction of hitbox data is a medium-level task in terms of complexity if you want it to work for a big set of games.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
If you log in to the main site using your forum details, and scroll down to the bottom, you'll find a link called "Upload a WIP...". Uploading your bk2 there will make sure it doesn't disappear.
Alternatively, if it's the same bk2 as you submitted, just link the submission, the submitted movie file will not be deleted even if the run is rejected.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
FearingSpoon wrote:
Warepire wrote:
If you found the values, why don't you go ahead and list the addresses and types so they are documented?
Here is the modified ram watch file that contain the values!
https://drive.google.com/open?id=0B79Jgh7LtuWbZEZiM1lJbUt2aGc
Here they are if you want to to look at the by hand
Kingsley's size: 2 Byte Unsigned 0x0E8520
Silver key: 2 byte signed 0x0E8590
purple key: 2 byte signed 0x0E8594
Orange key: 2 byte signed 0x0E8592
red key: 2 byte signed 0x0E8596
yellow key: 2 byte signed 0x0E8598
green key: 2 byte signed 0x0E859A
orange key 2: 2 byte signed 0x0E859C
You can use the userfiles storage. Log in to the main page, and then almost at the very bottom there's a link saying "Upload a WIP...".
It accepts RAM Watches and Lua scripts as well as movie files.
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
17:39 <kabuto> by the way, what would be the right place to write down info about glitches in games that where *not* used in a run?
17:39 <Warepire> kabuto: A Game Resources page
17:44 <kabuto> would this also be the right place to dump info about glitches in obscure though tas-able games that I found by accident at some point?
17:44 <Warepire> Yes
17:44 <Warepire> And maybe duplicate it in topics for the games in the forum
17:45 <Warepire> Alt. just link to the page in a topic
17:45 <kabuto> there are some serious glitches in the PC version of Captain Comic that I found at some point, though I'm pretty sure that game would end up in the vault
17:45 -!- Mothrayas [~mothrayas@ip-80-113-19-100.ip.prioritytelecom.net] has joined #tasvideos
17:45 <feos> do use game resources
17:46 <Warepire> Moth, kabuto wants some Game Resource pages, and possible editor status
17:46 <feos> or rather, a specific priv
17:46 <Mothrayas> game resources or editor, choose one
17:46 <Mothrayas> and tell them to contact me
17:47 <feos> game resources
17:47 <feos> Warepire: tell kabutoto contact Moth
17:47 <Warepire> kabuto: ^^^^
17:47 <feos> wait what
17:48 * feos missed a space
17:48 <Warepire> That's just my way to signal "read above"
17:48 <Warepire> He's here
17:48 <feos> you should have said
17:48 <feos> kabuto: contact to Moth
17:48 <Warepire> I figured the pings would bring them together
17:49 <feos> would be more funny
17:49 <feos> like 4 people around the table talking to each other only via the chain
17:49 <kabuto> okay thanks, will do
17:49 <kabuto> I will just need to edit/create game resources
17:49 <feos> Mothrayas: kabuto will contact you
17:49 <Warepire> feos: You are the weakest link
17:50 * feos gets thrown away
17:50 <feos> oh shit
17:50 <Mothrayas> alrighty then
17:50 <feos> Warepire: kabuto will contact Moth
17:51 <Mothrayas> tell him he now has game resource editing privileges ^^
17:51 <feos> Warepire: ^
17:51 <Warepire> This is too much for me, I can no longer keep the fabrics of my existence straight
17:51 * Warepire segfaults
17:52 <kabuto> oh no, don't let the message get lost!
17:52 <feos> Mothrayas: Warepire segfaulted
17:52 <Mothrayas> feos: cancel or abort?
17:52 <feos> Warepire: cancel or abort?
17:53 <feos> he's not ponging
17:53 * Mothrayas aborts Warepire from time
17:54 <feos> kabuto: what have you done
17:54 * Mothrayas cancels the abortion
17:54 * feos undoes the cancelation
17:54 <Mothrayas> retry?
17:54 <feos> ignore
17:55 <Mothrayas> yeah let's settle on that
17:55 <Mothrayas> ignore it is
17:55 * feos unexpectedly segfaults too
17:55 <KennyManZZZ> Mothrayas: Missed opportunity to say "aborts the abortion"
17:55 -!- KennyManZZZ is now known as KennyMan666
17:56 <Mothrayas> KennyMan666: it was intended to tie in on "cancel or abort" from just before
17:56 <KennyMan666> Yeah, yeah
17:56 <kabuto> this needs a system-level debugger to catch the segfault
17:56 <Mothrayas> clearly we need Warepire...oh, wait
17:57 * feos turns on Nach to enable Warepire to use Warepire to debug Warepire
17:58 * feos catches an exception
17:58 <feos> woosh
17:58 <feos> you guys are realy bugged
Joined: 3/2/2010
Posts: 2174
Location: A little to the left of nowhere (Sweden)
feos wrote:
GetBlock() is insanely slow, because for every action, there's a call from C# to lua core written in C, that generates a ton of overhead compared to the same function running in Gens.
For whatever reason, not wiping the offscreen data (and dropping all the bounds checks) slows me down a few fps.
The reason I keep mentioning the size of the resulting table is because it does slow the emu down if the level map is too wide, even if I don't even iterate through all the entries.
You could read 4 pixels at a time to reduce it from 16 reads to 4, then manually split the data.
Something like (c-like pseudo code, adapt as necessary):
Language: c
for (i = 0; i < 4; i++)
pixels = rl(a1+(i*4))
final.contour[i*4] = (pixels & (0xFF << (i * 8))) >> i * 8
//repeat for i+1, +2, +3
