Hi everyone,
I'm was wondering: would it be a good idea to have the FileInput class
support a read() method? I found myself running into a small problem
while using xml.sax.parse in combination with fileinput. Here's a
snippet that demonstrates the problem:
###
import xml.sax
import fileinput
xml.sax.parse(f ileinput.input( ), xml.sax.Content Handler())
###
This doesn't work, because fileinput.input () isn't a file-like object:
it doesn't have read(), so it can't be converted into a nice InputSource.
But I thought it might be nice to have one, so I cooked up a quick
class to make it work:
###
class ByteStream:
def __init__(self, f):
self._buffer = []
self._f = f
def read(self, bytes):
if not self._buffer: self._prime()
next_block = self._buffer[:bytes]
del self._buffer[:bytes]
return ''.join(next_bl ock)
def _prime(self):
self._buffer.ex tend(list(self. _f.readline()))
###
And now I can get:
###
xml.sax.parse(B yteStream(filei nput.input()), xml.sax.Content Handler())
###
to work. But I'd rather make this solution a bit more permanent.
*grin* Would it be a good idea to add a read() method to FileInput to
make it more file-like?
Thanks!
