#/usr/bin/python -i
"""Demonstrates a bug or quirk in urllib
To run the test, run this script. When it tries to download the two
files in `doGet()` for a second time, you may see a message like

Retrieval of 'ftp://ftp.gnome.org/pub/debian/dists/stable/main/source/Sources.bz2' failed with error: [Errno ftp error] 200 Switching to Binary mode.

More remarks by bug submitter:

The bug is not seen if only 1 file is transferred
and allowed to finish before the next one starts.

Thus if the files transfer so quickly that one xfer finishes
before the 2nd begins then you may not see the problem,
in which case pick bigger files or use a computer
with a slower connection.
"""

import gc
import threading
import time
import urllib


class GetFile:

    def __init__(self, fromURL):
        self.fromURL = fromURL

        self._fromFile = None
        self._readBytes = 0
        self._totBytes = None

        self._getThread = threading.Thread(name="get", target=self._getTask)
        self._getThread.setDaemon(True)
        self._getThread.start()

    def _cleanup(self, exception=None):
        """Clean up everything. Must only be called from the _getTask thread.
        """
        if self._fromFile:
            self._fromFile.close()
        global transfersInProgress
        transfersInProgress -= 1

        if exception:
            print "Retrieval of %r failed with error: %s" % (
                    self.fromURL, exception)
        else:
            print "Done retrieving %r; read %s bytes" % (
                    self.fromURL, self._readBytes)

    def _getTask(self):
        """Retrieve the file in a background thread.
        Do not call directly; use self._getThread.start() instead.
        """
        try:
            print "Start retrieving %r" % self.fromURL
            global transfersInProgress
            transfersInProgress += 1
            self._fromFile = urllib.urlopen(self.fromURL)

            self._totBytes = self._fromFile.info().getheader("Content-Length")
            if self._totBytes:
                self._totBytes = int(self._totBytes)
                print "Total bytes:", self._totBytes

            while True:
                nextData = self._fromFile.read(8192)
                if not nextData:
                    break
                self._readBytes += len(nextData)
                # Seems to be needed to avoid long freezing after the first
                # run of `doGet`
                gc.collect()
                try:
                    ratio = 100.0 * self._readBytes / self._totBytes
                except AttributeError:
                    # `_totBytes` not set
                    pass
                else:
                    print int(ratio) * "#", int(ratio), "%"

            self._cleanup()
        except Exception, e:
            self._cleanup(e)


# prefix = 'ftp://ftp.astro.washington.edu/pub/users/rowen/python/'
# names = ('RO%202004-10-13.zip', 'RO%202004-05-21.zip')
prefix = 'ftp://ftp.gnome.org/pub/debian/dists/stable/main/source/'
names = ('Sources.bz2', 'Sources.bz2')

transfersInProgress = 0

def doGet():
    for name in names:
        GetFile(prefix + name)
    time.sleep(0.1)
    while transfersInProgress:
        time.sleep(0.1)

doGet()
time.sleep(3)
doGet()