pnguncorrupt.jl

#!/usr/bin/env julia

using CRC

# crc computing function
const handler = crc(CRC_32)
# offsets of IDAT chunks (points to 'I' byte)
const offsets = [114,131197,262278,393361,524445,655526,786609,917691,1048775,1179858,1188429]
# correct crcs. No 0x0a in those so we can assume they're correct ;)
const crcs = [0xf55a745d, 0x06e15ebf, 0xa0bdc18a, 0xadefb326, 0x09c557f7,
              0x0848d2c4, 0x295c44c9, 0xa4afc965, 0x2fad0fb3, 0x7ec18f9e,]
# Number of missing bytes / chunk
const missing_bytes = [1, 3, 1, 0, 3, 1, 2, 0, 1, 0,]

# bytes I don't care about
const start_bytes = [0x89,0x50,0x4e,0x47,0x0d,0x0a,0x1a,0x0a,0x00,0x00,0x00,0x0d,0x49,0x48,0x44,0x52,0x00,0x00,0x03,0xc0,0x00,0x00,0x02,0x58,0x08,0x02,0x00,0x00,0x00,0x22,0xf6,0x18,0x56,0x00,0x00,0x00,0x03,0x73,0x42,0x49,0x54,0x08,0x08,0x08,0xdb,0xe1,0x4f,0xe0,0x00,0x00,0x00,0x09,0x70,0x48,0x59,0x73,0x00,0x00,0x15,0xfd,0x00,0x00,0x15,0xfd,0x01,0xcd,0x70,0x51,0x61,0x00,0x00,0x00,0x1c,0x74,0x45,0x58,0x74,0x53,0x6f,0x66,0x74,0x77,0x61,0x72,0x65,0x00,0x41,0x64,0x6f,0x62,0x65,0x20,0x46,0x69,0x72,0x65,0x77,0x6f,0x72,0x6b,0x73,0x20,0x43,0x53,0x36,0xe8,0xbc,0xb2,0x8c,] # already fixed
const end_bytes = [0x00, 0x00, 0x00, 0x00, 0x49, 0x45, 0x4e, 0x44, 0xae, 0x42, 0x60, 0x82,]

function handle_chunk(chunk, checksum, missing, starti=1)
    if missing == 0
        checksum_correct = (handler(chunk) == checksum)
        checksum_correct && println("Found matching checksum!")
        return checksum_correct, chunk
    end
    found, i = find_next_0a(chunk, starti)
    while found
        # insert 0x0d
        insert!(chunk, i, 0x0d)
        is_correct, chunk = handle_chunk(chunk, checksum, missing-1, i+2) # skip 0x0d0a
        if is_correct
            print("Found location of 0x0d #$missing: at $i, around (")
            show(chunk[i-4:i+4])
            println(')')
            # chunk found, return
            return true, chunk
        end
        # delete 0x0d
        chunk = copy(chunk)
        deleteat!(chunk, i)
        found, i = find_next_0a(chunk, i+1) # skip current 0x0a
    end # while found
    # nothing found, return
    false, chunk
end

function main()
    # read chunks
    bs = open(readbytes, "Downloads/corrupt.png")
    chunks = Vector{Uint8}[]
    for i in 1:10
        chunk = bs[offsets[i]:offsets[i+1]-9]
        push!(chunks, chunk)
    end
    # fix chunks
    for (chunki, chunk) in enumerate(chunks)
        println("Fixing chunk $(chunki)...")
        is_correct, chunk = handle_chunk(chunk, crcs[chunki], missing_bytes[chunki], 1)
        @assert is_correct
        println(length(chunk))
        chunks[chunki] = chunk
    end
    # write chunks
    write_to_file(chunks)
end

# utils
function find_next_0a(bs, i)
    while i <= length(bs)
        bs[i] == 0x0a && return (true, i)
        i += 1
    end
    (false, i) # not found
end
function write_to_file(chunks)
    f = open("Downloads/corrupt_fixed.png", "w")
    map(b -> write(f, b), start_bytes)
    for (chunki, chunk) in enumerate(chunks)
        # write chunk length
        chunklength = (chunki == 10) ? [0x00, 0x00, 0x21, 0x6f] : [0x00, 0x02, 0x00, 0x00]
        map(b -> write(f, b), chunklength)
        # Write chunk type and chunk data
        map(b -> write(f, b), chunk)
        # write checksum
        map(b -> write(f, b), ntoh(crcs[chunki])) # ntoh to change endianness
    end
    map(b -> write(f, b), end_bytes)
    flush(f)
    close(f)
end

main()