21:12:06froggeypathnames are currently hashed using their namestrings. does (namestring path) return consistent values?
21:13:06fittestbitsOh, good question - probably not.
21:13:18froggeyusing the namestring isn't quite right, but it's been good enough so far
21:16:25fittestbitsThey look the same now - running namestring on the keys in the hash and namestring on the path. But, I think when I set the value in the hash, the namestring of the keys was different.
21:16:46fittestbitsThe namestring depends on the default value of the device.
21:17:09froggeya quick hack to make it work would be to modify sxhash so it returns 0 (or some other consistent value) for pathnames. that'd make lookup O(n), but it'd work at least
21:18:48fittestbitsThat's the problem - I changed the default device to be the same when I created the hash table and now I'm getting hits in the hash table.
21:18:58fittestbitsI'll have to think about how I want to fix this.
21:20:37fittestbitsBTW, for the dual boot system, I ended up with a logical partition in an extended partition. So, I updated detect-mbr-partition-table to handle extended partitions.