This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

classification
Title: ftplib.retrbinary fails when called from retrlines callback
Type: behavior Stage:
Components: Library (Lib) Versions: Python 2.4
process
Status: closed Resolution: rejected
Dependencies: Superseder:
Assigned To: Nosy List: christianmlong, facundobatista, giampaolo.rodola, sjones
Priority: normal Keywords:

Created on 2003-06-10 06:51 by christianmlong, last changed 2022-04-10 16:09 by admin. This issue is now closed.

Files
File name Uploaded Description Edit
ftplib_bug_report_one_file.py christianmlong, 2003-06-10 06:51 Broken version, with working version appended
Messages (5)
msg16314 - (view) Author: Christian Long (christianmlong) Date: 2003-06-10 06:51
Subject: ftplib.retrbinary() fails when called from
inside retrlines() callback function

I'm using ftplib to backup files from a Linux server to
a Windows 2000 worksation. 


Client:  
Windows 2000 Pro
ActivePython 2.2.2 Build 224 (ActiveState Corp.) based on
Python 2.2.2 (#37, Nov 26 2002, 10:24:37) [MSC 32 bit
(Intel)] on win32
Komodo IDE

Server:
ProFTP server (ProFTPd version 1.25) 
Mandrake Linux 9.0


Summary:

When I use it like this it works fine.

# Build a list of files that are on the remote server
f.retrlines('NLST', makeListOfFiles)

--then--
# Iterate over the list, retrieving each file
for remoteFileName in listOfFiles:
    --snip--
    f.retrbinary('RETR %s' % remoteFileName,
localFile.write)
    --snip--


But it fails if I try to do the retrieve directly in my
callback function.

def transferFile(listLine):
    --snip--
    f.retrbinary('RETR %s' % remoteFileName,
localFile.write)  <--fails here on first time through
    --snip--

# get list of files from server, adn transfer each file
as it gets listed
f.retrlines('LIST', transferFile)

--snip--
  File "D:\My Documents\Computer World\Python
Resources\My Utilities\backup_remote_files.py", line
45, in ?
    f.retrlines('LIST', transferFile)
  File "C:\Python22\lib\ftplib.py", line 413, in retrlines
    callback(line)
  File "D:\My Documents\Computer World\Python
Resources\My Utilities\backup_remote_files.py", line
36, in transferFile
    f.retrbinary('RETR mra.py', localFile.write)
--snip--
  File "C:\Python22\lib\ftplib.py", line 300, in makepasv
    host, port = parse227(self.sendcmd('PASV'))
  File "C:\Python22\lib\ftplib.py", line 572, in parse227
    raise error_reply, resp
error_reply: 200 Type set to I.


It looks like the server is returning a 200 instead of
a 227 when retrbinary() is called inside a callback
function for retrlines().



Files:

2 Files are included:  a broken version and a version
that works

This One Is Broken -  retrbinary() called from inside a
callback function for retrlines().
===================================================
import ftplib
import os
import time

REMOTE_DIR = "/home/mydir"
LOCAL_DIR = "C:\My Documents"
TIME_FORMAT = "%y%m%d"    # YYMMDD, like 030522

def transferFile(listLine):
    # Strips the file name from a line of a
    # directory listing, and gets it from the
    # server.  Depends on filenames
    # with no embedded spaces or extra dots.
    if listLine.endswith('.py'):
        #Split file name on the dot
        splitFileName=remoteFileName.split('.')
        # Add a timestamp
        localFileName="%s_%s.%s" % (splitFileName[0],
                                   
time.strftime(TIME_FORMAT),
                                    splitFileName[1])
        # Open a local file for (over)writing, in
binary mode.
        # print os.path.join(LOCAL_DIR,localFileName)
       
localFile=file(os.path.join(LOCAL_DIR,localFileName), 'wb')
        print remoteFileName
        print localFile
        # Execute the FTP retrieve command, calling
        # the write() function of the local file
        # for each block retrieved from the FTP server
        # BUG: This should work, but I get the
following traceback
        f.retrbinary('RETR %s' % remoteFileName,
localFile.write)    #<--- Fails
                                                      
              #     Here

#        mra.py
#<open file 'D:\My
Documents\Work\IA\Miller\MRA\Dev\Backup of remote
files\mra_030610.py', mode 'wb' at 0x00886B70>
#Traceback (most recent call last):
#  File "C:\Program Files\ActiveState Komodo
2.3\callkomodo\kdb.py", line 430, in _do_start
#    self.kdb.run(code_ob, locals, locals)
#  File "C:\Python22\lib\bdb.py", line 349, in run
#    exec cmd in globals, locals
#  File "D:\My Documents\Computer World\Python
Resources\My Utilities\backup_remote_files.py", line
45, in ?
#    f.retrlines('LIST', transferFile)
#  File "C:\Python22\lib\ftplib.py", line 413, in retrlines
#    callback(line)
#  File "D:\My Documents\Computer World\Python
Resources\My Utilities\backup_remote_files.py", line
36, in transferFile
#    f.retrbinary('RETR mra.py', localFile.write)
#  File "C:\Python22\lib\ftplib.py", line 385, in
retrbinary
#    conn = self.transfercmd(cmd, rest)
#  File "C:\Python22\lib\ftplib.py", line 346, in
transfercmd
#    return self.ntransfercmd(cmd, rest)[0]
#  File "C:\Python22\lib\ftplib.py", line 322, in
ntransfercmd
#    host, port = self.makepasv()
#  File "C:\Python22\lib\ftplib.py", line 300, in makepasv
#    host, port = parse227(self.sendcmd('PASV'))
#  File "C:\Python22\lib\ftplib.py", line 572, in parse227
#    raise error_reply, resp
#error_reply: 200 Type set to I.

# The problem is that the self.sendcmd('PASV') call is
not getting a 227
# reply from the server.  Rather, it is getting a 200
reply, confirming
# that the type was set to I (Image).


        localFile.flush()
        localFile.close()

f=ftplib.FTP('server', 'user', 'password')
f.cwd(REMOTE_DIR)
# List directory contents, and call the transferFile
# function on each line in the listing
f.retrlines('LIST', transferFile)

f.close()




===================================================



This One Works - retlines() builds a list, and then
files are transferred by iterating over
that list and calling retrbinary() for each.
===================================================
import ftplib
import os
import time

REMOTE_DIR = "/home/mydir"
LOCAL_DIR = "C:\My Documents"
TIME_FORMAT = "%y%m%d"    # YYMMDD, like 030522

listOfFiles = []

def makeListOfFiles(remoteFileName):
    # Strips the file name from a line of a
    # directory listing, and gets file from the
    # server.  Depends on filenames
    # with no embedded spaces or extra dots.
    if remoteFileName.endswith('.py'):
        listOfFiles.append(remoteFileName)

f=ftplib.FTP('server', 'user', 'password')
f.cwd(REMOTE_DIR)
# List directory contents, and call the transferFile
# function on each line in the listing
f.retrlines('NLST', makeListOfFiles)
print listOfFiles

for remoteFileName in listOfFiles:
    #Split file name on the dot
    splitFileName=remoteFileName.split('.')
    # Add a timestamp
    localFileName="%s_%s.%s" % (splitFileName[0],
                                time.strftime(TIME_FORMAT),
                                splitFileName[1])
    # Open a local file for (over)writing, in binary mode.
    # print os.path.join(LOCAL_DIR,localFileName)
   
localFile=file(os.path.join(LOCAL_DIR,localFileName), 'wb')

    # Execute the FTP retrieve command, calling
    # the write() function of the local file
    # for each block retrieved from the FTP server
    f.retrbinary('RETR %s' % remoteFileName,
localFile.write)
    localFile.flush()
    localFile.close()

f.close()





===================================================
msg16315 - (view) Author: Shannon Jones (sjones) Date: 2003-06-15 00:45
Logged In: YES 
user_id=589306

The problem seems to happen when you use a callback within a
function that was called as a callback. Here is a much
simpler case that demonstrates the problem:

-----------------------------
import ftplib
 
def transferFile(listLine):
    filename = listLine.split()[-1]
    if filename == 'README':
        # Note that retrlines uses a default
        # callback that just prints the file
        f.retrlines('RETR README') # <-- Fails
 
f=ftplib.FTP('ftp.python.org', 'ftp', 'anon@')
f.cwd('/pub/python')
f.retrlines('LIST', transferFile)
f.close()
-----------------------------

This fails with the following:

Traceback (most recent call last):
  File "ftptest.py", line 10, in ?
    f.retrlines('LIST', transferFile)
  File "/home/sjones/src/python/dist/src/Lib/ftplib.py",
line 407, in retrlines
    callback(line)
  File "ftptest.py", line 6, in transferFile
    f.retrlines('RETR README') # <-- Fails
  File "/home/sjones/src/python/dist/src/Lib/ftplib.py",
line 396, in retrlines
    conn = self.transfercmd(cmd)
  File "/home/sjones/src/python/dist/src/Lib/ftplib.py",
line 345, in transfercmd
    return self.ntransfercmd(cmd, rest)[0]
  File "/home/sjones/src/python/dist/src/Lib/ftplib.py",
line 321, in ntransfercmd
    host, port = self.makepasv()
  File "/home/sjones/src/python/dist/src/Lib/ftplib.py",
line 299, in makepasv
    host, port = parse227(self.sendcmd('PASV'))
  File "/home/sjones/src/python/dist/src/Lib/ftplib.py",
line 566, in parse227
    raise error_reply, resp
ftplib.error_reply: 200 Type set to A.


Note this is with the current CVS version on Redhat 9.
msg16316 - (view) Author: Facundo Batista (facundobatista) * (Python committer) Date: 2005-01-15 18:50
Logged In: YES 
user_id=752496

Same behaviour, Py2.4 on Win2k, sp2.
msg73552 - (view) Author: Giampaolo Rodola' (giampaolo.rodola) * (Python committer) Date: 2008-09-22 02:21
In FTP every data channel is supposed to be used for a unique transfer
(RFC-1123, chapter 4.1.2.6) and every client should open only one data
connection at time. Actually there isn't any official RFC which
explicitly states my second sentence but the common practices for
servers receiving a PORT or PASV request while another transfer is in
progress usually are:

- processing the request when the transfer is finished
- creating a new data channel closing the old one
- returning a 4xx temporarily failure response code

IMHO it's your use case which, even if not "officially" declared
incorrect, is usually discouraged and hence should not be covered by
base ftplib module.


My2cents
msg104051 - (view) Author: Giampaolo Rodola' (giampaolo.rodola) * (Python committer) Date: 2010-04-23 20:48
Closing this out as rejected.
History
Date User Action Args
2022-04-10 16:09:07adminsetgithub: 38618
2010-04-23 20:48:36giampaolo.rodolasetstatus: open -> closed
type: behavior
resolution: rejected
messages: + msg104051
2008-09-22 02:21:59giampaolo.rodolasetmessages: + msg73552
2008-03-19 21:59:53giampaolo.rodolasetnosy: + giampaolo.rodola
2003-06-10 06:51:09christianmlongcreate