I'm pretty confident the Death+Circle bug will be fixed with my patch, I'm just overly cautious.
I'll add the regen/death bug to the list; it's probably in the same function.
Neither of the Chemist Item Swap bugs (Elemental Status, Power Drink) are fixable inline. I'm still debating what the best strategy will be here. There's an unused attack formula at C2/8261 which I might be able to use, but I'm hesitant to do that. I could do a long jump to another bank where there is free space, but that involves finding unused space (not as easy in FF5 as later games). The other option is to work on optimizing some of the code - using JSRs to consolidate duplicate code. That's not particularly ideal either. Once I decide on a course of action, my plan is to use $2044,X to store Power Drink's bonus Attack Power and $2045,X to store Mix's added elemental effects. I think I'm going to move on to some other bugs and come back to these with a clearer head.
So in my mind, the actual bug here is the whole function at C2/9A6F that recalculates stats when equipment has been changed. I've been working on this function to see if it can be optimized so I can free up space to make changes necessary for the Chemist Item Swap bugs. As it turns out, there are a whole class of bugs that are essentially due to this function. This covers both the Chemist Item Swap bugs, as well as various underflow and overflow bugs, most notably the Berserker Thornlet one. It turns out that this function performs no apparent overflow check, so while I didn't see any listed, there are probably numerous instances of that. This will be a lot of work, and while I'm still working on it, I've decided to move on to other bugs so I don't get burned out on it.
The Odin-Magic Lamp-Zantetsuken bug is interesting not because it was hard to debug or is overly hard to fix, but that there are actually a few things here that contribute to the problem. The first is simple - some bosses don't have the Heavy flag set, including Neo Exdeath Pt. 4
. So part of this fix is to go through and set all of those. That would (I believe) adequately fix the bug, since Zantetsuken's attack formula will "ignore" Heavy targets (they will animate being cleaved, but then come back). The second problem is that Magic Lamp, unlike Summon, doesn't check for Heavy targets when casting Odin. Therefore, Odin will never use Gungnir. So, a simple fix is to make Odin use Gungnir instead of Zantetsuken - change D0/EEE9 from 6A to 6F. This won't affect the actual summon. A better fix is to have Magic Lamp check for Heavy targets... which is a 24 byte inclusion there is no room for. Alternatively, Magic Lamp could just call Summon... but I'm not sure of the full implications of doing that. A cursory inspection shows they are similar but not exactly the same. So my plan is to audit the bosses, setting the Heavy flag that should be set on all of them. That should fix the issue, but if it doesn't then I'll look into routing Magic Lamp into Summon to see what happens.
So I think everyone knows the Scan Bug by now - C2/6C93 should be 10, not 08. Simple fix, not going to actually release a patch on it. What I will mention here for completeness is that Observe, Scan, and Analyze all work a little differently. Observe only returns level and HP, Scan and Analyze both return everything but Analyze works on bosses and Scan does not. That appears to be the intended behavior unlike some other bugs, so I won't be changing that.
I'm still alive, but I took a couple weeks off from coding, but I'm back in the game now!
SwdSlap, by all accounts, should have a chance to inflict paralysis on the target. It does not. All actions, whether commands or spells or whatever, have eight-byte data tables that roughly define how the ability works. For commands, bytes three and four indicate action flags and damage modifiers that get set in 201E,X and 201F,X respectively. For example, this is how Defend and Guard set their respective action flags. SwdSlap sets the damage modifier flag (201F,X) to #02. However, this is never used in the code; the function that parses these modifiers simply skips over it - there is no code to handle it. Presumably this flag was to indicate that the attack should paralyze.
One way to fix this would be to code handling for the flag. However, the
better way to do this is to emulate another command that combines a regular attack with a secondary effect - Mug/Capture. In fact, what I think happened is that SwdSlap was meant to be converted to this method since it is more elegant, but got overlooked during production. So this is what I'm working on, although it will require some work since C2 is rock solid as far as extra space.
While I'm here, a word on Capture. I've seen it mentioned on occasion that Mug/Capture is bugged since it doesn't use any other weapon effects. This is intended behavior. It deliberately calls a version of Fight (0816 versus 45FF) that ignores added abilities. I believe the intent is that you either get added effects OR you get the special effect, and I believe this is why FF6 simply overwrites and/or ignores those effects as well.
An easy one this time that I've actually made a patch for! What is this madness?! The Release-Banish bug was a simple one to identify and a simple one to fix. Moss Fungus and Gel Fish were calling spell index E9, which was a version of XZone that for whatever reason only targeted the caster. I pointed both of them to spell index 47, which is the Time Magic version of Banish/XZone other enemies use. I've attached the zip of the (very simple) patch.
Release doesn't parse the targeting flags completely, so targeting is limited to five options. It only looks at the "Target All" flag (#40) and the "Default Enemy" flag (#08). This produces four options: target all enemies (48), target all allies (40), target a random enemy (08), and target self (00). You can also get target a random ally by setting any flag other than those two. It might be an interesting project to expand this targeting so that there are more options for Release, but beyond the scope of this bug-fixing project.
As always, bugs can end up being pretty interesting and not always straight forward to fix. Fixing the various Water spells - Aqua Rake, Tsunami, and Big Tsunami - was still pretty straightforward, but Aqua Rake/Breath required a bit more ingenuity. The Tsunami spells use attack formula 06 - Magic Attack - and simply didn't have the element flag (parameter 3) set. It was a simple fix.
Aqua Breath (AB) on the other hand was a little more interesting. AB doesn't
always do Water damage throughout the series, so I debated whether it was intended behavior for it not to be water-elemental. It uses a separate attack formula (6C) that does a non-elemental magic attack with bonus damage against certain creature types. There are two possible solutions - one is to modify formula 6C to look at elemental damage, and the other is to modify formula 06 to look at creature type. Both formulas do not utilize parameter 1, so either approach is possible. Ultimately I decided on the second approach, since it involves only changing one other spell (Apollo Harp) and uses much less space. So this patch changes formula 06 to check for creature type, and changes AB to use the basic magic attack formula. This has a few consequences. AB now suffers multi-target penalties, and can now be absorbed, blocked, and otherwise defended against like any normal magic attack. However, AB now gains bonus damage from enemies weak against water in addition to damage against desert creatures. I consider this a fair trade-off.
