It seems strange to me that the function registered with gui.register() gets called both:
1. between the two functions registered with emu.registerbefore() and emu.registerafter(), respectively.
2. whenever the screen is refreshed at frame boundaries.
I don't think this behavior is proper and would like to propose that the function registered with gui.register() should only be called at frame boundaries.
EDIT: with the proposed change, it would be possible to separate the output from emu loops like this:
function update_func()
-- updatenew data just once per frame
end
function display_func()
-- display with updated data
end
emu.register(update_func)
gui.register(display_func)
Under current state, this or similar cannot work correctly as the execution of update_func and display_func may be interlaced every frame depending on things such as whether Frame Advance is used.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
I figured out a way to save 2 frames in the last area of Act 1-1 circa Frame 2960-3030 by clipping into the top-right corner of the Z-shape passage earlier while still having the turret enemy shoot to the right side:
http://dehacked.2y.net/microstorage.php/info/1980497449/krionconquest-klmz-v13f.fm2.
No similar improvement applicable elsewhere.
I would like to request for replacing the submission with the updated fm2.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
Glad to see you enjoy it.
And thank you for reminding about the time goal as well as explaining the game rules. Good job!
The 2nd one looks good I think.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
I forgot about the case for the "slower" zipping to the left that works very stage by alternating walking Left and Right inside the wall. It's possible because the game checks the wall tiles from left to right, which lets the ejection by the tile on the right overwrite the ejection by the tile on the left on the same line.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
Thanks for voting! I'm glad that you enjoy this.
The "famous" horizontal zipping is actually constantly forward-ejecting onto 8-pixel grids plus slightly forward-sliding out of the grids every frame, which form a loop. If the Left or Right button on the D-pad is pressed in order to move Francesca out of the 8-pixel grids, the wall in front of her would only eject her backwards opposing her walking direction. That's why the horizontal zipping only works on Stage 2 AFAIK for it requires the D-pad-free sliding physics there.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
mtvf1 made one, but it seems to be reduced to 30fps by Nicovideo....
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
True that this game is Vault material mainly due to its level design. I said to myself "what still a cliimbing stage again?!" and laughed as I realized that I shouldn't have been surprised during watching the run weeks after it was finished. XD
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
For jumping on the normal thick stationary platforms in most stages, one frame can be saved every two jumps by manipulating the Y position. For the thin (eg. "lines") or moving platforms, it depends on the situation but yes the Y position affects the time. I compared various solutions of every section and chose the fastest one each time.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
The currently published "warpless" run as-is:
* Allows "intended" warps (using the warp rooms).
* Allows several "traditional" glitches such as Arbitrary DSS Activation, zipping on platforms, falling through moveable stairs in Underground Waterway and so on.
* Forgoes the "DSS-summoning spells + wall-jumps" trick to skip to Camilla (that is hard to call a "glitch").
* Forgoes any glitches to ascend through ceilings and room exits?
* Or just forgoes the Infinite Acsension glitch at all?
* Forgoes any DSS-summoning warps.
* Allows the Sliding Attack cooldown glitch or not? Maybe "Yes" as it doesn't look so glitchy?
* What about the DSS-Timefreeze glitch to skip fighting some bosses eg. the Iron Golem with dialogue events after the fights and still get the Magic Items behind them? I guess "No".
* What about the "in-bound" tricks as shown in this submission to skip dialogues with the Necromancer? I really can't be sure about this.
That's quite complicated and arbitrary.
Compared to a proposed Fighter Mode run:
* Just beat the game in Fighter Mode.
EIDT: replying to the second question.
Well, I think we can break down the issue into at least three parts:
1. Do we want this "alll cards" branch published?
2. Do we want a "warpless" branch (whatever definition fits) published?
3. Do we want some branches else published?
I don't think the resolution of Part 2 should interfere with the resolution of Part 1, given that the two branches are quite different visual-wise and technique-wise.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
An opinion is to maintain the status quo until such a new movie comes up. I use the term "deprecate" instead of "obsolete" in my suggestions to indicate that I consider it as a futural direction rather than an immediate change.
And we do have seen precedents such as [991] NES Bucky O'Hare by AnS in 22:13.60: The old normal mode "any%" run was obsoleted by a "1-hit-death Hard Mode" run that requires a password.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
That would be suboptimal as kicks are too weak for their time cost.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
I'll just share some basic learnings from a private debate regarding current rules on the said "endless games":
* A complete run is to complete the gameplay, not the game content; The run reaches the end of the gameplay, not the end of the game (credits, "the end" etc.).
* The player decides what the gameplay consists of; Whether that is acceptable by others is a different question.
* What we as the TASVideos community (not any RTA community) want to make clear as our common view (and set in stone as the Rules) is, what gameplay shall a TAS consist of to be approved and published here.
I am pessimistic about having everyone agree on all of the statements above.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
If (1) the criteria could be defined clearly well and if (2) then the expectable improvements allowed by the criteria could be incentive enough....
Such a run requires someone purposeful enough to work out at least a complete route to follow.
I am interested in seeing such a run anyways.
EDIT: Disambiguation.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
My "100%-ish" is just "all DSS cards" inspired by "all souls" in AoS and "all furnitures" in HoD. The "completionist" category is overdoing and reduces the need for clever skips and resource management IMO.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
You mention that a side-effect of the glitching is to corrupt the in-game timer. So I'm just wondering - what is the official in-game time for this TAS?
5281:46:20. If the in-game timer weren't corruptible, it would be 00:03:22.
Wow. Can the game timer even go that high naturally? If you left the game running for 5281 hours, would it actually hit that? Or can it only go that high because of the corrupting?
The in-game timer stops ticking as soon as it reaches 99:59:59.00 or higher. The only way to get past 99:59:59.00 is to corrupt the memory AFAIK.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
It's always great to see a 10 years old movie getting updated. Yes vote.
Do you plan to do other categories of this game? Looking back, the category of [1093] GBA Castlevania: Circle of the Moon by Cardboard in 21:40.22 is defined in a rather arbitrary way even back then. The only way to recreate a run taking the relatively "normal" path without making the category too arbitrary on glitch usage at this point is probably to use the Fighter mode so that you can't use magic at all.
I fuly agree that the "warpless" criteria of this game are arbitrary. Pike and I started a Fighter Mode "any%" earlier but the progress is stalled. On the other hand, I myself have made a "100%-ish" TAS, but the goal might feel arbitrary, too.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
Any input file that starts from power-on (for example, a previously submitted movie for that game) and creates the exact circumstances for your submission to sync will generally do.
Is it allowed to create the SRAM-verification movie itself from other verified SRAMs?
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
You mention that a side-effect of the glitching is to corrupt the in-game timer. So I'm just wondering - what is the official in-game time for this TAS?
5281:46:20. If the in-game timer weren't corruptible, it would be 00:03:22.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
One thing that I'm wondering though is whether an any%/100% category split should be done here, or maybe have the door open for a glitchless category instead. I'm honestly torn on this one.
I'm wondering the same thing, as saving every friend could be considered as a form of full completion. See this line from the Movie Rules:
If a progress counter is filled by collecting a set of items or fulfilling a set of flags, all individual components of this set must be collected or fulfilled. Collecting or fulfilling the same component multiple times to inflate the progress counter is not counted towards full completion.
I don't think the game was intended by the developers to be beaten without collecting all pre-spawned friends. Thus no "official any% vs official 100%": they would just be the same.
On the other hand, you can't save more than necessary to finish a level: the game checks the counter against exact zero for opening the goal gates. Once any glitch-spawned friends are saved, you can't collect all of the non-glitch-spawned ones in oreder to complete the level, unless the counter wraps around and matches with the remaining amount, which is very improbable if not impossible at all. Therefore no "duplication-glitched any% vs duplication-glitched 100%": the latter is just non-existing.
fmp wrote:
Without the glitch, aren't any% and 100% just the same thing here?
So there coud only be "duplication-glitched any% vs official (only in the sense of which friends to save) any%/100%". However, as the friend-duplication glitch is not applicable in most levels, an optimized "all friends" movie will vastly resemble this submission.
ThunderAxe31 wrote:
So the question is: should this submission obsolete the current movie or be published as a new branch?
The currently publication is unoptimized.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
What is it that you use the Iron Golem for? The submission text seems to be cut off there.
It's just used to perform the 0HP glitch. I've fixed up the text. Thanks.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
Enjoyed it. In-bounds movement is on the slower end of TASed castlevania runs but what can you do?
What's NKT mean? It shows up a lot in the niconico comments of the temp embed when the ending starts.
It's a meme that means (IIRC) "It was a long and challenging adventure." It's making fun of how short Castlevania TASes tend to be due to the glitches and movement tech.
The other one that commonly shows up is shouting IGAAAAAAAAAAA when IGA's name appears in the credits.
And "show time" (originated from Castlevania: Portrait of Ruin) when restart at the title screens to continue.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
When I was going to quote to reply a PM, I got somehow logouted and had to re-login. Things went normally until I got this:
No post ID was specified
DEBUG MODE
Backtrace:
includes/functions.php[743]: ErrorPrintBacktrace()
privmsg.php[1630]: message_die(202, string(24) "No post ID was specified"
)
EDIT: By the way, I've been seeing this frequently when I am trying to Preview or Submit an edited post or PM:
Invalid Session. Please resubmit the form.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do