Re: Concat Occasionally Drops a Character or Two

[Jonathan Shaltz]

Additional update: Creating a file populated with the values A through Z (0x41 through 0x5A), it loads just fine, even with a large-ish file (1.3 MB).

I repeated this exercise with a file that consisted of the characters 0x00 through 0xFF, repeated through the end of a 1300000-byte file. The very first chunk lost a byte on its way into MySQL. Chunks #6 and #9 also lost a byte each. I'm still bewildered. It would make *some* sense if the program collapsed Windows-style linebreaks (0x0D, 0x0A) to UNIX (0x0A), though the BINARY flag should prevent that. But this test file does not have any such combinations. Because there are only 256 possible values, each 10,000-byte chunk includes about 400 repetitions of each combination of (byte X) || (byte X+1).

I suppose I'll sift through the decryption routines and see if I can wrap the data in AES and get it loaded correctly. Doubtful, since it would just shuffle one binary file into another, but I'm grasping at straws here.

- Jonathan