Re: (Incomplete) Telefang English Translation

[kmeisthax]

Patch v70 is out. This is primarily an antipatch for v40, removing the workarounds that fixed extraneous VRAM clears because making the VWF more conservative in terms of tile usage also caused it to not clear enough VRAM. The combination of v40 and v67 introduced a bug on status screens where garbage tiles from previous screens would be present on moves that were not long enough to cover up the last tile allocated. This patch fixes this.

[andwhyisit]

A helpful Kochia put the beatdown on me and showed me a bug :<



EDIT: Tested v68 and it showed Crypto's name correctly when he was needled to death by a Monstera.

v71:
http://www.mediafire.com/?58lale7mv8hojgk

This fixes the issue and adds the new denjuu names to the patch.

---

Alright, now to explain..

v69:
When you get a Denjuu's number the game automatically stores an auto-generated 6-bye truncation of your Denjuu's name. I decided that trying to break this limit in sram was a bad idea. So I made each new Denjuu nickname "E6 E0 E0 E0 E0 E0" then modified the function that read these values to check for this value at the start of the name, should it be found it looks for the id of the Denjuu (stored elsewhere in sram) and uses that to read the name directly from bank 34, else it simply reads the nickname from sram as per normal. So E6h at the start of a nickname means "use default instead".

tl;dr: This fixes the auto-nickname limit.

Note: A number of E5 codes were changed in the script.

v71:
The bug RacieB pointed out is because the Denjuu's info in sram (including id and with the exception of nickname) is deleted before the nickname is read, so the id will always return 00, and Denjuu 0 is Tsunonasu. To fix this I found the function that deletes the info and made it store the id in CCBF before clearing the data. I then made the "use default" function check if all 16 bytes of the Denjuu's data in sram are empty, if so then it pulls the id from CCBF, if not it pulls the data from sram as per normal.

tl;dr: Bug fixed.

---

v72:
http://www.mediafire.com/?1cqkeh5b9xivtc9

Bug fix for v71. Denjuu names other than Crypto were behaving stangely due to the code looking for the sram Denjuu data in 6-byte blocks rather than 16-byte blocks. Not sure how I missed that one.

---

v73:
http://www.mediafire.com/?h00bkea2ujncal9

A bugfix for my.. bugfix? :blink: Moved some code and forgot to change a jump.

[Kimbles(imported)]

Note: A number of E5 codes were changed in the script.

Uuuugh, you know that means we have to update the script that's on the wiki if we ever want to update something in that section of text. You could at least tell us what you changed.

Also, please please please don't do this again, just wait until you have time to write up a changelog before releasing it. Otherwise it causes a lot of confusion and nobody else can work on it until you explain the changes you made. :/ If you're not going to document what you're doing at all, you should at least make yourself available on some kind of instant communication so we can ask you questions. I know you hate IRC for some reason, but it's causing a lot of inconvenience for the other people working on the patch... Nobody's asking you to actively take part in the conversation or anything, but just sticking around there would help us a lot.

[andwhyisit]

Uuuugh, you know that means we have to update the script that's on the wiki if we ever want to update something in that section of text. You could at least tell us what you changed.

All E5 E1 C9 were changed to E5 A3 CC.

And I'm sorry but in order to expand the names stored in ram I need to move them where there is space, and in turn that requires that I modify E5 codes, else the dialogue will look at the wrong address and start spewing machine code until it hits an E0. There is no other way around it.

I'll make sure I specify the E5 codes changed in the future but I WILL need to modify E5 codes again.

---

If you're not going to document what you're doing at all,

Woah! Hold on there...

The last time (and the first time) I didn't changelog it was because someone was trying to sleep in the same room as the computer and my dad said "get off that computer NOW" in a voice so acidic I thought he was going to pull the plug.

This time it was because someone caught me still on the computer at 2am in the morning.

What makes you think I make a regular habit of this?

---

just wait until you have time to write up a changelog before releasing it.

Well v69 took ~14 hours to produce. I wasn't going to wait until someone released a patch that conflicted with mine so that I had to reapply everything again manually, or reapply their hack manually.

Though I did work on v71 first thing in the morning and completely forgot to post the changelog there and then, of which I take full responsibility for.

[Kimbles(imported)]

I don't have any issue with you changing stuff, you just really have to keep us updated. XD

Changelogs aside, you don't really explain what you're working on or planning on changing or give details on what parts you've changed, which are the sort of things the other people hacking the game at the same time need to know. :O

Sorry for being snappy, but it's kind of frustrating to watch. I'd just prefer to see people working together more. ):

[andwhyisit]

On an unrelated note, would be more advantageous to have a list of all known bugs in the patch added to Wikifang?

[Sanqui]

Patch v74: http://wikifang.meowcorp.us/etc/TPatch74.ips

* Changed one ld a,09 to ld a,10 so attack names don't get cut off in-battle
* Changed "Cultured Denjuu" to "Cultivated Denjuu"
* Changed "Dark Void Denjuu" to "Dark Space Denjuu"
* Inserted Umekichi "dialogue"
* Incremented the version number by one

[andwhyisit]

* Changed one ld a,09 to ld a,10 so attack names don't get cut off in-battle

Shouldn't that be a ld a,11 (16 characters and a trailing E0)? Otherwise the original would be ld a,08.

---

* Incremented the version number by one

lol.

[Sanqui]

* Changed one ld a,09 to ld a,10 so attack names don't get cut off in-battle

Shouldn't that be a ld a,11 (16 characters and a trailing E0)? Otherwise the original would be ld a,08.

If I recall correctly, no move is long 16 characters. If one is, though, it may get merged with the previous phrase... I can't make it simply be ld a,11 because the 0x11th byte isn't a e0; I'd have to add code that writes it.

[andwhyisit]

* Changed one ld a,09 to ld a,10 so attack names don't get cut off in-battle

Shouldn't that be a ld a,11 (16 characters and a trailing E0)? Otherwise the original would be ld a,08.

If I recall correctly, no move is long 16 characters. If one is, though, it may get merged with the previous phrase... I can't make it simply be ld a,11 because the 0x11th byte isn't a e0; I'd have to add code that writes it.

Unless you're pulling it from the attack name list directly you'll most likely find that the trailing E0 is there already, in the byte sequence that you are reading from.

I'd like to take a look though to confirm though. What's the address of the ld you changed?

---

v75:
http://www.mediafire.com/?3i363mxfhicnecb

Moved the name data in D460 to D6A0 to prevent data in ram from being overwritten or misread.

Note: All E5 60 D4 codes in the script are now E5 A0 D6. Unfortunately there are too many to list.

v76:
http://www.mediafire.com/?gginzeict2onohw

Moved the name data in D440 to D68D for the same reason as stated above.

Note: That one E5 40 D4 code in the script is now E5 8D D6. It's from the line that says "[name] appeared!" in bank 45.