Comment 77 for bug 581925

Martin,

the regressions snug in someway

- the patch to tell dwd-rw in sequential mode from restricted overwrite mode posted in comment #63 is correct and is exactly the same as your proposed fix, but I can take the blame for not having revised the committed code thoroughly

- I don't know, where the number 17 in ioctl that reads first 32 blocks comes from, I assume that it's the least sufficient portion to read but I guess that reading a portion of data that is not power of two may give unpredictable results

- the committed variable renaming scheme in this part of code is somewhat confusing (though, of course has no impact on operation) - the "block" is really the offset and the "offset" is the byte currently being examined

The flaws that may cause incorrect detection of DVD media never got into my ppa code - I didn't sync it with git and I had no issues, but I admit, I should have done this