Shadow of the Erdtree Compatibility #931
Conversation
vawser
commented
Jun 30, 2024
vawser
commented
- Added new Paramdex stuff for ER
- Added new Paramdex stuff for DS3
- Added new Paramdex stuff for AC6
- Added MSB changes for ER
- Added MSB changes for DS3
- Added MSB changes of AC6
- Added "Param Validation", "Paramdef Validation" and "Map Validation" tests for checking the param padding, paramdef accuracy and MSB accuracy.
- Fixed bug with eventParam save in ER where it would fail due to the condition checking for sysParam != null instead of eventParam
- Added new Paramdex stuff for ER - Added new Paramdex stuff for DS3 - Added new Paramdex stuff for AC6 - Added MSB changes for ER - Added MSB changes for DS3 - Added MSB changes of AC6 - Added "Param Validation", "Paramdef Validation" and "Map Validation" tests for checking the param padding, paramdef accuracy and MSB accuracy.
| @@ -172,6 +180,9 @@ internal override Event ReadEntry(BinaryReaderEx br) | |||
| case EventType.RetryPoint: | |||
| return RetryPoints.EchoAdd(new Event.RetryPoint(br)); | |||
|
|
|||
| case EventType.AreaTeam: | |||
| return AreaTeams.EchoAdd(new Event.AreaTeam(br)); | |||
|
|
|||
There was a problem hiding this comment.
Happy to see updates here. Not sure if this what tk was bemoaning but it'll help.
| @@ -782,17 +784,17 @@ public class UnkStruct1 | |||
| /// <summary> | |||
| /// Unknown. | |||
| /// </summary> | |||
| public uint[] DisplayGroups { get; private set; } | |||
| public uint[] DisplayGroups { get; set; } | |||
There was a problem hiding this comment.
with all the choices to private indexes for sanity, why are we making these setters public?
being an array, it's already an immutable reference to a correctly-size mutable array? maybe that's not how c# handles it idk
There was a problem hiding this comment.
Main use case is XML serialization as it would fail on this
If this is handled by some kind of setter method, it should be overriding this default setter, and the propertyshould be backed by a private field
|
|
||
| /// <summary> | ||
| /// The map to load when on this collision. | ||
| /// </summary> | ||
| public byte[] MapID { get; private set; } | ||
| public byte[] MapID { get; set; } |
There was a problem hiding this comment.
again privacy thing.
I'm guessing this is a convenience for setting to a mapid.
Perhaps we can write in the length-check in the setter method?
There was a problem hiding this comment.
Main use case is XML serialization as it would fail on this
If this is handled by some kind of setter method, it should be overriding this default setter, and the propertyshould be backed by a private field
| /// Unknown. | ||
| /// </summary> | ||
| [IgnoreProperty] | ||
| public short UnkT00 { get; set; } |
There was a problem hiding this comment.
ideally this is placed at the top of the class, in regular field order
| @@ -15,6 +15,7 @@ | |||
| <PackageReference Include="Google.Protobuf" Version="3.25.2" /> | |||
| <PackageReference Include="Grpc.AspNetCore" Version="2.60.0" /> | |||
| <PackageReference Include="Grpc.Net.Client" Version="2.60.0" /> | |||
| <PackageReference Include="ZstdNet" Version="1.4.5" /> | |||
There was a problem hiding this comment.
soapstone is requiring it?
There was a problem hiding this comment.
It's the new compression for regulation.bin
| bw.WriteInt32(8); | ||
| */ | ||
| } | ||
|
|
There was a problem hiding this comment.
I'm still hopeful we'll figure this out, though it seems more and more likely we'll need to inject some hardcore ghidra user like chain with a few shots of caffeine
| @@ -45,7 +45,8 @@ | |||
| <Field Def="f32 emissiveValueEnd"></Field> | |||
| <Field Def="f32 emissiveTime"></Field> | |||
| <Field Def="u8 bIntpEnable"></Field> | |||
| <Field Def="dummy8 pad_01[3]"></Field> | |||
| <Field Def="u8 unknown_0x59"></Field> | |||
| <Field Def="dummy8 pad_01[2]"></Field> | |||
There was a problem hiding this comment.
one of these days we'll do due diligence and check it's a byte and not a short the first time. But I'm happy with this for now.
| @@ -27,6 +27,7 @@ | |||
| <Field Def="s32 priority = 3" /> | |||
| <Field Def="f32 execInvalidTime" /> | |||
| <Field Def="u8 execButtonCircle = 1" /> | |||
| <Field Def="dummy8 pad4[3]" /> | |||
| <Field Def="u8 unknown_0x45" /> | |||
| <Field Def="dummy8 pad4[2]" /> | |||
There was a problem hiding this comment.
Good finds here in ds3, gracias
| @@ -1397,7 +1397,12 @@ | |||
| </Field> | |||
| <Field Def="s32 finalDamageRateId"> | |||
| </Field> | |||
| <Field Def="dummy8 pad7[12]"> | |||
| </Field> | |||
There was a problem hiding this comment.
This dummy I'm fairly sure came after finaldamagerate, so these could just be renamings.
Also it may be worth checking if finaldamagerate was this as well, since it showed up in like 1.06 or 1.07 and wasn't in original paramdefs
| <Unicode>True</Unicode> | ||
| <FormatVersion>203</FormatVersion> | ||
| <Fields> | ||
| <Field Def="s32 Unknown0" /> |
There was a problem hiding this comment.
prefer lowercase unknown, and byte offset than number
| @@ -6,15 +6,15 @@ | |||
| <Unicode>True</Unicode> | |||
| <FormatVersion>203</FormatVersion> | |||
| <Fields> | |||
| <Field Def="u8 enableRaytraceAO"></Field> | |||
| <Field Def="u8 enableRaytraceShadows"></Field> | |||
There was a problem hiding this comment.
this was confirmed as community gaslighting wasn't it?
| <Unicode>True</Unicode> | ||
| <FormatVersion>203</FormatVersion> | ||
| <Fields> | ||
| <Field Def="s32 Unknown0" /> |
There was a problem hiding this comment.
lowercase+offset, many such comments apply in places other than where they first appear
src/StudioCore/Assets/Paramdex/ER/Defs/MapGridCreateHeightDetailLimitInfo.xml
Outdated
Show resolved
Hide resolved
| @@ -264,9 +264,14 @@ | |||
| <Maximum>1</Maximum> | |||
| <SortID>2100</SortID> | |||
| </Field> | |||
| <Field Def="dummy8 pad1:7"> | |||
|
|
|||
| <Field Def="dummy8 pad1_old:7" RemovedVersion="11210015" /> | |||
There was a problem hiding this comment.
Is there particular reason to rename old field? Curious, since it's also a relic of original paramdef. Some util I'm guessing is expecting distinct names across versions?
My preference is for pad_new, though I don't think this change is a stopper lol
| <Field Def="u8 swordArtsTypeNew" FirstVersion="10701000" RemovedVersion="11210015" /> | ||
| <Field Def="dummy8 pad[1]" RemovedVersion="11210015" /> | ||
|
|
||
| <Field Def="u16 swordArtsTypeDlc" FirstVersion="11210015" /> |
There was a problem hiding this comment.
Nord confirmed that swordArtsTypeNew should've been u16 since 10701000
I fucked that one up
| <pad AltName=""/> | ||
|
|
||
| <swordArtsTypeNew AltName="Sword Arts Type" Wiki="Offset for TAE animation. 0 is a600. To pass to the behavior script to determine which swashbuckler" /> | ||
| <swordArtsTypeDlc AltName="Sword Arts Type" Wiki="Offset for TAE animation. 0 is a600. To pass to the behavior script to determine which swashbuckler" /> |
| 100116315 Blade of Calling | ||
| 100116316 Blade of Calling | ||
| 100116317 Blade of Calling | ||
| 100101100 Fire Knight's Shortsword / Blade Knife - |
There was a problem hiding this comment.
I'm assuming these dashes are placeholder for labelling individual moves?
src/StudioCore/Assets/Paramdex/ER/Names/Gconfig_LightingQuality_scarlett.txt
Show resolved
Hide resolved
src/StudioCore/Assets/Paramdex/ER/Names/MapGdRegionInfoParam.txt
Outdated
Show resolved
Hide resolved
| @@ -179,7 +179,6 @@ | |||
| 21910000 Radagon - Crucified (Unscaled) | |||
| 22000000 Elden Beast | |||
| 22000078 Elden Beast | |||
| 22000178 Elden Beast | |||
There was a problem hiding this comment.
I'm guessing removals in this file are because they don't appear, than that they don't match chr-id conventions?
| @@ -3536,7 +3536,6 @@ | |||
| 30917003 Cough up your coin, all of it! | |||
| 30917100 Hm? Wait, don't tell me, is that you? | |||
| 30917101 Oh, cripes, please, wait! | |||
| 30917102 I surrender! I surrender, I swear! | |||
| param WepAbsorpPosParam: modified && !added: reserve: = vanillafield reserve | ||
| 11210015L | ||
| 1.12.1 - (SwordArtsParam) Set swordArtsTypeDlc to Vanilla v1.12.1 values |
| 11210015L | ||
| 1.12.1 - (WorldMapPointParam) Set unknown_0xb7 to Vanilla v1.12.1 values | ||
| param WorldMapPointParam: modified && !added: unknown_0xb7: = vanillafield unknown_0xb7 |
There was a problem hiding this comment.
we can coalesce as unknown_0xb[789] if we use = vanilla; (I think previous commands may have been written before that shortcut)
| 11210015L | ||
| 1.12.1 - (WepAbsorpPosParam) Set unknown_0x54 to Vanilla v1.12.1 values | ||
| param WepAbsorpPosParam: modified && !added: unknown_0x54: = vanillafield unknown_0x54 |
| 11210015L | ||
| 1.12.1 - (SpEffectVfxParam) Set unknown_0x96 to Vanilla v1.12.1 values | ||
| param SpEffectVfxParam: modified && !added: unknown_0x96: = vanillafield unknown_0x96 |
| 11210015L | ||
| 1.12.1 - (SpEffectParam) Set unknown_0x352_6 to Vanilla v1.12.1 values | ||
| param SpEffectParam: modified && !added: unknown_0x352_6: = vanillafield unknown_0x352_6 |
There was a problem hiding this comment.
unknown_0x35(2_[67]|3_[012345])
| param ToughnessParam: modified && !added: unk2: = vanillafield unk2 |
There was a problem hiding this comment.
param SwordArtsParam: modified: swordArtsTypeNew: = field swordArtsType will stop functioning if the name is changed again (should re-check if other unks were identified too)
| 11210015L | ||
| 1.12.1 - (AssetGeometryParam) Set Reserve_0 to Vanilla v1.12.1 values | ||
| param AssetGeometryParam: modified && !added: Reserve_0: = vanillafield Reserve_0 |
There was a problem hiding this comment.
Nice to see we don't have a heap of fields being actively moved.
Are there any additions where we should modify modified && added to = 1 for example, like damagerate fields that are otherwise going to be initialised as 0?
I don't see any new refs like we had for finaldamagerate, but same story.
- Changed DecompressDCPZSTD function name to DecompressDCXZSTD - Re-privatized some MSBE.Part properties - Added in proper paramdef names for the new 1.12.1 params - Added in some ER Names that were missed - Restored some ER Names that were accidently removed
- Actually included the restored names
- Updated Paramdex to latest changes.
