overflow (int32) decode an nbt (schematic file) with very large number of blocks (3192890994=3501*194*4701)

[CMA2401PT]

overflow when decode an nbt (schematic file) with very large number of blocks (3192890994=35011944701)

at nbt/decode.go ,line 190:

case TagByteArray:
aryLen, err := d.readInt32() <- overflow here, get -1102076302= 0xbe4fa672(int32), but expected to be 3192890994 = 0xbe4fa672(uint32)

    if err != nil {
        return err
    }
    if aryLen < 0 { <- then error
        return errors.New("byte array len less than 0")
    }

version is go-mc@1.7.0, but it seems current version has same error since it still reads an int32

could you please consider change it to an uint32?
something like:
aryLenInt32, err := d.readInt32()
aryLenInt:=uint32(aryLenInt32)

[Tnze]

It could be changed. But are you sure it can solve the problem you meet? Such a large array?

[Tnze]

According to https://wiki.vg/NBT, the TAG_Byte_Array prefixes a signed integer.

[CMA2401PT]

yes, already solved
if it is necessary, I can send the schematic file to you

here is the commit in the relevant project:

https://github.com/LNSSPsd/PhoenixBuilder/commit/73b40cb45f3dd0ae1f3f7648153cc12a3b289c0c
bouldev/phoenixbuilder-depends#2

[Tnze]

Can other nbt implement handle your file?

[CMA2401PT]

you are right, it seems that the author of this schematic file omit the standard you mentioned to save such large area… sorry for this issue