Posts for Andypro
1 2 3 4
10 11
Experienced Forum User
Joined: 5/24/2004
Posts: 262
antd wrote:
current: Reno defeated next: Shinra HQ
Congrats! Sounds like you're making good progress.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
antd wrote:
i dont think anyone has replicated the glitch current: About to post a WIP so that all you guys have something to do this weekend while you're eating a meal in front of your computer.
Fixed for ya!
Experienced Forum User
Joined: 5/24/2004
Posts: 262
What a novel idea for a TAS. Big yes vote here. I'm impressed that you wrote a program for this. I would love to see the source code! Was it a lua script or something else? Edit: Bisqwit posted the source below. Thanks for the code, Bisqwit!
Experienced Forum User
Joined: 5/24/2004
Posts: 262
I voted yes for this submission. I think it would be wise to pretend for the moment as if there is no future OoT run that will be published. It takes away part of the debate, and each submission should be judged according to its own merits, not as a comparison to a future submission that doesn't yet exist. Given that, what are the remaining issues. The fact that this submission is not well-optimized and the fact that it uses the Japanese rom. To me, the former point is overridden by the fact that this obsoletes the currently published run by a huge margin. I believe even Guano knew his run was quite sub-optimal when he submitted it for publication. It still merited a yes vote and I believe this one does too. The second reason is much more valid in my opinion, but in my mind the vastness of the improvement justifies making an exception to the US rom rule. Having said that, I think those who are working on a US tas or planning the route have come to the realization that their TAS will now probably be about the same time as this one given that it will use the US rom with slower text, but include all the optimizations currently known to bring the time close to even with this submission. Thus whenever that submission comes it will not look as impressive as it otherwise would have on paper and will need to be accepted more for entertainment (which it will be). Unfortunately, whether this run is published or not, it remains an OoT tas which will be compared to whatever future submission arrives.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Looks great. Only thing I saw is that ryuka to silence tm after the top tower and then hydrofoil to the bottom tower would likely save some frames as opposed to the hydrofoil all the way method. I love that Chaz dealt the final blow to PD :)
Post subject: PAL detection patch
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Hi, I have finally found why FCEUX detects all of my NTSC roms as PAL. In iNES.cpp, it checks the opened ROM pathname for the typical PAL filename marks like (E), (F), etc. Here's an example pathname from my system: F:\Games\(e) NES\Blaster Master (U) [!].nes I think you can now see the problem. I've created a patch which I hereby submit to improve the pal detection process by only examining the name portion of the path. Please review and integrate into the latest svn, or let me know where I can submit the patch officially. Thanks!
Index: fceu/src/ines.cpp
===================================================================
--- fceu/src/ines.cpp	(revision 1644)
+++ fceu/src/ines.cpp	(working copy)
@@ -733,13 +733,46 @@
 	// TODO: MD5 check against a list of all known PAL games instead?
 	if(OverwriteVidMode)
 	{
-		if(strstr(name,"(E)") || strstr(name,"(e)")
-			|| strstr(name,"(F)") || strstr(name,"(f)")
-			|| strstr(name,"(G)") || strstr(name,"(g)")
-			|| strstr(name,"(I)") || strstr(name,"(i)"))
+		char *s, c, *p, *q;
+		int index = 0;
+
+		s = (char*)malloc((strlen(name)+1) * sizeof(char));
+		strcpy(s, name);
+
+		//  Reverse the path string "name" and put in s
+		if (strlen(s))
+		{
+			q = s;
+			while (*(++q)) ; // points q at '\0' terminator
+			for (p=s; p < --q; p++)  // ignores middle character when strlen is odd
+			{
+			  c = *p;
+			  *p = *q;
+			  *q = c;
+			}
+		}
+
+		char *nameRev = strstr(s,"\\");
+		if(nameRev != NULL) {
+			index = s - nameRev;  //  pointer arithmetic
+			index = index<0 ? (index = -index) : index;  //  absolute value
+		}
+
+		char *nameOnly = (char*)calloc(index+1, sizeof(char));
+		strncpy(nameOnly, s, index);
+
+		//  The parens are reversed because we're using a reversed string of
+		//  the filename.  No need to re-reverse it and allocate more memory
+		if(strstr(nameOnly,")E(") || strstr(nameOnly,")e(")
+			|| strstr(nameOnly,")F(") || strstr(nameOnly,")f(")
+			|| strstr(nameOnly,")G(") || strstr(nameOnly,")g(")
+			|| strstr(nameOnly,")I(") || strstr(nameOnly,")i("))
 			FCEUI_SetVidSystem(1);
 		else
 			FCEUI_SetVidSystem(0);
+
+		free(nameOnly);
+		free(s);
 	}
 	return 1;
 }
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Woo hoo! You rock!
Experienced Forum User
Joined: 5/24/2004
Posts: 262
FreshFeeling wrote:
EDIT: Hey, guess what I spent my evening doing!
Wow! That is a lot of work. Thanks, that's a great spreadsheet.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
The last bit with demi looks pretty good to me. What's the final count on frames saved between this latest WIP and your published run?
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Man, that DF3 battle is so boring and long. Are there any purchasable items or old weapons that could be thrown at him to speed it up at all? Probably a longshot. In the Lashiec battle, is it possible to manipulate that one round where Lashiec kills Chaz before he can cast tsu? If that were possible, he may only have to be revered once. It would be great if a way could be found to do that.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
janus wrote:
http://dehacked.2y.net/microstorage.php/info/11360821/Phantasy%20Star%20IV%20%284%29%20%5B%21%5D.gmv now almost 13k frames shorter after DF I. I did take the napalm shot, hoping it will create more damage in the future... btw, I was able to reach 600 damage with tsu
Wow, that's not the current WIP is it. That's actually a really old file isn't it. There's no shadowblade in that or anything. Or am I going crazy! Ah!
Experienced Forum User
Joined: 5/24/2004
Posts: 262
That is awesome. Will definitely help whoever decides to try the low level Chaz run.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Woo hoo, it is finally published! Congrats, janus.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
I just tried recording a GC movie on the latest: r4619. No dice, at least for me. I tried recording Ikaruga (PAL) chapter 1. Tried with both framelimit set to 0 and also framelimit set to auto. Neither file synced properly on playback. I haven't tried a Wii game yet, but I can do that later. I don't hold out much hope yet since even GC stuff doesn't seem quite ready.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
I've kept up with janus' progress and have seen the final movie - it seems well-optimized and a worthy candidate for the first published TAS of ps4. It's about time this game was done! Yes vote.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
It looks good. My only possible improvement question: After you beat DF3 and warp back to Kadary, wouldn't it be a bit faster to hop in the Hydrofoil and drive to the spaceport instead of walking? Hydrofoil is so fast.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
I watched your previous (v4) up to Rykros TAS and it looked very good. I especially liked how you redid the Lashiec fight by keeping the party alive and also that you didn't have to farm the towers on Rykros for experience.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Cool. Where did you make the gains? In battles?
Experienced Forum User
Joined: 5/24/2004
Posts: 262
It would be cool to see the game beaten with the least amount of leveling, weaponry, and skills possible imho. I think that would maximize the speed of the run - if you're able to just eek out a victory at every boss you encounter, you know you're at the minimum possible stats to do it. Your test runs have been pretty good, so I'll probably vote "yes" when you finally submit it if the quality of play is as high regardless of what you decide.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
I was initially thinking of giving this a "meh" vote because I noted several things that were suboptimal - mostly toward the beginning of the run in angel island and hydrocity. But, since this run doesn't really aim for best time, and we already have plenty of those runs, that it definitely gets a "yes" for entertainment value. The lava reef act 1 boss battle was one of the coolest things I've seen in a sonic TAS. Really good style overall.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Any updates, antd? Did the scripts from the poster above help in your battle testing?
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Nice progress. Why did you make that Brose macro while leveling but never use it? You subsequently overwrote that macro slot before the DF3 battle. Or, did I miss something?
Experienced Forum User
Joined: 5/24/2004
Posts: 262
I just watched your previous wip before the one you just posted up to the air castle basement - everything looks good to me.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
janus wrote:
i have 1500 reasons :p. No one will beat my run that easily
Cool. Well, good luck with v3. I'll look forward to more quality WIPs.
Experienced Forum User
Joined: 5/24/2004
Posts: 262
Oh, you restarted your run? Why'd you need to do that? Did you find out how many frames ahead you are after Zio in comparison to your v2 run (I think your current one is your third version)?
1 2 3 4
10 11