Bloom filter problem when number of elements are low and FPRate is high

[MBakhshi96]

The length of bloom filter as determined in this line

python-bitcoinlib/bitcoin/bloom.py

Line 116 in b5540e8

self.vData = bytearray(int(min(-1 / LN2SQUARED * nElements * math.log(nFPRate), self.MAX_BLOOM_FILTER_SIZE * 8) / 8))

becomes 0 when number of elements nElements is 1 and nFPRate is more than 0.03. This causes every pubkeyhash to return True when checked using contains method.

The way the length of the filter is determined using bytearray also causes some problems, for example for some values, say 0.0005 < nFPRate < 0.01 and nElements = 1 when you change nFPRate the length of the filter doesn't change. This means that regardless of the false positive rate, filter length and its construction is the same. This leads to a wrong false positive rate, different than the parameter passed for filter construction.

A possible fix could be to use bitarray instead of bytearray.