Forum Replies Created
-
Posts
-
Wow, you actually made more progress than I had made by the time I saw this post. Sorry about that :/ I’m afraid I don’t have much more to add.
I’ve never seen the water glitch myself, it’s weird that it would happen in the text at all…
Have you been able to tell if there is any difference between how the two tables are accessed? I was never able to figure that out.
Yeah, what happened in LYS’s case was it didn’t delete the extra pointers on loading a new rom. So when loading one rom, then another, it had double the amount of extra pointers (the extras seem to appear at random, so I save them to avoid the trouble of trying to figure out any pattern behind them).
The N!N!N errors are actually a format I used while debugging the rom. It’s to indicate strange pointer values.
Pointer1!Pointer2!Officer No.
So 0!211521!162 indicates there was no pointer before officer 162, and the value of the pointer after was 211521. There are all kinds of weird issues with officer 161, because that’s the "Pirate" officer type with 5 extra values for no apparent reason.
In 0.6, the pointer duplication issue should be fixed (I released that one yesterday evening).
Sorry about that error! I’ll take another look at this issue tomorrow.
Okay, the soldier progressions are not in any standard format.
They’d have to be stored in 3 bytes each (since otherwise the 999999 wouldn’t fit)… which is awfully awkward in itself, plus I couldn’t find any. :/ It looks like they’re described by some kind of function, but I’m not sure how I would go about finding such a function in the hex…
Updated the sprite and tactics tables, they are available here.
One of the changes in the sprite table is that I differentiate between Zhang Bao (Shu) and Zhang Bao (Turbz).
/edit: Bonus table update! ‘ has been added to the reverse table too.
You may need the .Net framework. I didn’t actually think to mention it because it’s pretty common nowadays.
On name-wrapping:
I noticed when looking through some of the characters, some people had values other than FF right before their names! (this was one of the first problems I ran into when parsing – there’s an exception to every rule, including this one :D ) I have no idea what the numbers there actually mean, though, so while I don’t think there will be side-effects… there may well be.
Such people were the tri-name people (Cheng Yuan Zhi, Zhu Ge Liang).
I don’t have support for this in the tool (as it’s not confirmed to work yet), but you could try changing the 10 right before Zhuge’s name to FF and see if that fixes it. If it works as I suspect it does, this will also fix Cheng Yuanzhi, although Pei Yuanshao will never be displayable properly (as Yuanshao is 8 characters).
To add apostrophes, you should be able to just add the ‘ to the list of characters in both the alpha-tables: doaetable.tbl and revdoaetable.tbl, just as with a normal table file. As long as you keep the formatting (in the first, 2 characters = 1 char, in the second the reverse) it should load them fine and have no problems.
The APs, ACs, and Level progression I’m not sure about. They’ve got to be stored somewhere and are probably in an order…
I did notice, though, that prices are stored with the weapon names. Ex. Dagger has 32 right before it -> 50, Flail has 64 before it -> 100. All item names are stored starting at 2FAF0.
The An Sha flag…
I don’t see anything Zhou Yu, Sun Quan, Cao Pi and Sima Yi have in common aside from a white portrait background, sadly. Any idea if Yuan Shu and Yuan Shao are also immune to An Sha? Who else is known to be immune?
Have you tried setting Sima Yi’s int to 0 and An Sha-ing him?
I came to realize today what I should have looked for a long time ago.
You know how every time you get hit, you’re forced to press a button?
9 of Liu Bei’s
soldiers
were
(press button)
defeated.
What if you didn’t have to press it anymore?
Change the value at:
28731 from FD to FB to make this into a simple line break, and you won’t have to do this when you get hit ever again.
I haven’t found the level-ups yet, but I *have* found where regular soldier values are stored.
Look in 33EE0 (just past all the Officer stuff).
You’ll see a mysterious sequence
0000640068006E
…
If you convert 0064 to decimal, you get 100. 0068 is 104. These are encoded as 01 and 02 in the table we made…
These continue up to A000 which is 40960 :)
Unfortunately, there it looks like it gets more messy. I don’t know if I could decipher the level-up amounts. I’ll look into it a bit more.
Any idea how much EXP it takes to go up each level for, say, even the first 5 levels? It would help a lot to figure this out, if it’s stored in anything near normal formatting.
Okay, got a final burst of energy for the night :)
I’ve updated the program somewhat and now it changes both kinds of pointers as needed.
I had a brief issue with townspeople again (this time, only Song Yong, Song Ren, and the guy who complains about Cui Zhouping, in testing)… but it has been resolved (I update their pointers too).
As a test, I converted DoaE 2.0 beta (or the version I have anyway) into the new pointer system. After some initial hurdles (which thankfully helped fix some tricky pointer-bugs), I have ported the old system (using FFs to fill empty values) to the new system (updating pointers instead). The new version loads fine in the editor (which displays that now there are 108 extra characters of space left! The original had 70)… and I tested it as well; no issues that I saw.
So… this is still a test build, but it’s far beyond what it was before. It’s a bit rough around the edges, and it may still have some bugs – but that’s what beta testing is for. Please take a look and let me know what I can/should work on.
Operating instructions:
1) Load the ROM.
2) Click on an officer’s name.
3) Change stuff about that officer.
4) Click "Save" to save changes to that officer.
5) Repeat 2-4 as desired. There is no error accounting for using *too much* space – generally I would say it would lead to very bad things. However, it does tell you how much space is left.
6) When you’re ready to save the new ROM, click file->Export to ROM. I recommend saving it in a new file and testing, just in case it still isn’t all working.
Porting older changes:
The extra FF values show up as underscores "_" in the editor. To get rid of them, simply delete, save the officers as you go, then export.
–I’ll look into the level-ups tomorrow if I get a chance. Not sure where to begin looking, but I may be able to ask someone more knowledgeable for advice.
Okay, looks like said beta-release may be delayed a bit.
When I posted about how pointers were calculated, I figured they just used a simple diff +256 if it was negative. Turns out there’s another set of values indicating what page they refer to… but thankfully they’re easy to find.
I’ll figure out how these work, then hopefully have something good to show. Probably won’t be till tomorrow, though – feeling rather fuzzy tonight and fuzziness + pointers = bad.
1) I’m not sure how rebel soldiers work. I haven’t had a chance to look into it – I only know they’re mysteeeeerious :P
2) I’m not aware how the experience-per-level is determined, either, I’m afraid – too busy with the officer editing stuff to look into much of anything else.
and… victory!
Unfortunately it only worked the first time around. I was reading townsperson pointers wrong, but writing them correctly… so loading what I changed then re-saving caused…
Attack of the clones!
Fixed that problem, so now we have…
Even more thorough victory!
Guan Yu’s name is actually "reallylongnamenosrslythistime". Unfortunately, the game only displays 14 characters in-battle (curiously, 15 in general-screen).
James: I’m not sure how feasible it would be to add officers or to change which ones are in which battle. I’m really hoping that the pointers there are done by number of officer (i.e. officer number 160 is Pang De, or some such, so "officer 160" is in set battles). If they are, that shouldn’t be too bad.
I have no idea what the best remedy for Han Zhong is. I would guess finding where he’s being referenced, then fixing up a dictionary reference for the region would be the best solution… but then we’d have to do more dictionary stuff.
I will look over and do more thorough testing tomorrow morning (re: afternoon as I like sleep). Then it may be beta-release time :)
Well, that was fast. Turns out the pointers after the officers all point to townspeople, and I had forgotten to update them! Whoops.
By the way: as far as I can tell there are 70 spaces we can fit officer names in. If we fix all the Si Mas and Zhu Ges, that gives us a goodly amount of stuff to fill in with other things. This should make pretty much any officer mod no problem when I finish it :)
I’ve taken a look at auto-generating it and I’m not sure there’s an easy way to go about doing it. I thought it was integer-indexed (i.e. if "Bob" is right after "Fred", Fred’s index is Bob’s + 1), but I tried generating it like that and it was off by quite a bit. When I worked on mine (the one in the table), I mostly looked for words I knew – I went off of the first speeches in the game for many of them.
Also, since I’ve been working on it lately, an update on pointers and such!
I believe I am updating the pointer refs properly… but apparently NPC sprite refs are right after officers are!
I tried changing "Cheng Yuan Zhi" to "Cheng Yuanzhi", then updating everything. I essentially shifted everything back one value, updated the pointers… the generals seem to have mostly come out okay but…
Whoa!
The mysterious ghost Song brothers…
Somewhat bizarrely, nearly everyone in Xuzhou has changed into one of three sprites:
1) Blue Pirates
2) Blue Zhang Bao (like on the icon for DoaEd)
3) Blue Liu Bei-lord-sprite
I’m looking into what might’ve caused this now.
Yeah, I love the pointer thing. It makes things a lot easier.
Although it has some unintentional effects, like Tao Qian mysteriously changing :)
I haven’t had a chance to take a look at this yet, but I will probably do so over the weekend.
Well, I’ve more or less figured it out…
It took some number crunching in Excel-equivalent (too cheap to buy Office), though.
It has a starting point (206881 dec, don’t have the hex on hand), stored by 11 hex. The next officer location is the in the byte difference between that, added to that point. I’ll do all my math in decimal to keep it simple.
51 – 17 = 34, so our next officer is expected to start at 206881+34 = 206915.
For our next, do the same:
84 – 51 = 33, so our next officer is expected at 206915+33 = 206948. When the diff is negative, we take the complement (add 256 to it).
The one thing I don’t get is why the duplicates are there (the 3 extra 74s at the beginning, for example). I did this in Excel-equiv to figure out what they did; turns out they are just junk (i.e. everything matches up without them).
So I think what I’ll be doing is storing the pointer values with each officer, then generating that section afterward. Take the diff between "original name length" and "current name length", add that to the value, make sure we aren’t going to spill over into anything important, then write the officers and their pointers.
