Take for instance this array from my PDB:
0xc6fb : Length = 14, Leaf = 0x1503 LF_ARRAY
Element type = 0x13B9
Index type = T_ULONG(0022)
length = 48
Name =
48 means the array is 48 bytes long, but the current PDB parser interprets it as an array of 48 ulongs (which are 4 bytes each, so that array would take up 4x more space).
I currently have this fix implemented, I'll PR it soon.