Belf

what's with the healing on the last turn (ch. 3)? are you trying to train ellen?

Here's the fixes for the tutorial and epilogue graphic problems: tutorial: Apply this to an untouched v0.98 rom. i don't know why sf won't allow me to attach it Epilogue: 0xA22001 - 00 80 > C0 7F Come on Gringe, let's get version 1.0 out now!

Okay that makes sense. I was editing some text with FEditor, and the pointer for that text had an 88 instead of 08 at the end. Makes sense since it was uncompressed in the data.

I know pointers generally end in 08 for little endian, but what about those that end in 88 or something else? where do those point to?

Found an error: the torch and torch staff descriptions are switched. Since they have the same name, i can see why this happened.

Kishuna WITHOUT RNG ABUSE

Holy shit. I'm on chapter 2 of AC on the original AW for the gba, and I cannot figure out a strategy to S-rank the chapter in 10 turns. I saw a solution on one of the FAQs on gamefaqs, but I haven't been able to find any other way to do it. Has anyone else figured out an alternative strategy for this chapter, or is there essentially only one way to do it?

That can all be changed through event hacking (it does not require ASM), however the translation patch team did not mess this up, it was just overlooked by the original developers (there was also a broken convo between thany and tate that never worked in chapter 19 sacae). I know this because you can see the event information in the original, unpatched rom.

jesus...

Tabachanker, can you specify how you found where the routine you had to change was located? This is always the hardest part for me when trying to ASM hack, and the only way I have managed to find the routines i was looking for was when I knew a location in the memory that was either read or written during the routine, and would just set a read/write break on that location. I don't imagine how I could have used that technique for this problem however.

"They want vengenance." - chapter 20x sacae opening error

I tried to apply your IPS RNG patch for FE6 in the thread you referenced but it ended up bricking the game where it won't even load up.

I went through the routine with a debugger. I hacked to go from ch. 11 HHM to ch. 23 HHM (since I didn't have a save), so I only had Matthew and Hector. Thus, I only used Hector, Matthew, and Hawkeye for the chapter. Hector gained 595 exp. points, Matthew 231, and Hawkeye 20. At the very end of the calculation routine the total was 1041, which it then subtracted 195 to get the correct total of 846, which it then checked to see if it was over 700. I wondered why it added an extra 195, and I later found out that the 195 was Hector's total accumulated experience before starting the chapter. In other words: Before the chapter: Hector lv. 2, 95 exp. Matthew lv. 2, 0 exp. After: Hector lv. 8, 90 exp. Matthew lv. 4, 31 exp. Hawkeye lv. 4, 20 exp. Matthew had not gained any experience prior to this chapter, and neither did Hawkeye since I just recruited him during the chapter. Hector, however, had 195 exp. gained before the chapter, so the routine adds the total accumulated exp. of all characters, then subtracts what they had before the chapter started. Conclusion: Dondon miscalculated, unless he provides proof he didn't.

I wasn't sure so I set breakpoints on both actually (subtracting one and the Thumb mode one). By setting breakpoints, I mean that I clicked on the line of code in the left hand main window of the debugger on no$gba. Yet, still the game never stopped from the end of the chapter to the next chapter, which is why I think I must have done something wrong.

So in layman's terms, those instructions happen regardless and are not conditional, and therefore irrelevant? What's weird is that I set breakpoints on all of those offsets of the ones I underlined, but the game never broke during the end scene. I have no idea what I'm doing wrong...

I've put the relevant information from ch. 23's events in the spoiler box. I don't know which is the exact routine though. Try the ones that are bolded and underlined.

Yes. I'm guessing I'm going to need to use a write break in the VRAM location where the portrait graphics are temporarily stored. After that there will hopefully be a simple branch operation only for Fa's specific unit slot (x03 or x04, don't remember).

The problem is, I think ONLY Fa's portrait is taller. Even the village kids' mugs are subject to the bottom cropping. Guess I gotta use ASM...

So it seems that most portraits in the gba series get their mugs shortened by 8 pixels. This is problematic for me for characters that have short mugs (Serra and some others) when I insert their mugs over the original characters ones (specifically for FE6). What's the easiest way to get around this? Also, does anyone know of an ASM solution as well? I understand ASM somewhat, but I wouldn't know where to find the routine for the cropping of the portrait.

AI priority is an algorithm, not a formula. Therefore, even if you try to follow the ASM operations, it's too complex to figure out. It's much easier and faster to just experiment with the game itself.

Yeah, that's what I was saying. Yodel's katakana is Yooderu. Yoder would not be Yooderu, but rather Yoodaa.

WTF, Yoder sounds nothing like the katana. It would be more like Yodaa. Not every name needs to have some symbolic or historic meaning behind it.

nvm.

I guess... If I had done what Gringe did though, I would have had my own version (my own name preferences) and public version(s) (unaltered names, majority supported names, etc.) rather than force everyone to use what I like or else they have to change it (which takes a LOT longer than you think).

Even so, every time you update the patch I have to redo the entire process, which is why I want to know a faster way or else it isn't worth it.