File upload duplicate filename check mechanism fail

[fadlytabrani]

The file upload duplicate filename check mechanism will fail if the upload directory contains a dot and when the filename does not.

django.core.files.storage:Storage

def get_available_name(self, name):
        """
        Returns a filename that's free on the target storage system, and
        available for new content to be written to.
        """
        # If the filename already exists, keep adding an underscore to the name
        # of the file until the filename doesn't exist.
        while self.exists(name):
            try:
                dot_index = name.rindex('.')
            except ValueError: # filename has no dot
                name += '_'
            else:
                name = name[:dot_index] + '_' + name[dot_index:]
        return name

# Failure example, uploading duplicate file "testlog" will create new directory upload_.10122008
# Correct result would be "testlog" would be uploaded with the filename testlog_
get_available_name("c:\\django\\media\\upload.10122008\\testlog")

Using os.path would solve the problem

def get_available_name(self, name):
        while self.exists(name):
            dname, fname = os.path.split(name)
            try:
                dot_index = fname.rindex('.')
            except ValueError:
                fname += '_'
            else:
                fname = fname[:dot_index] + '_' + fname[dot_index:]
                
            return os.path.join(dname, fname)

[fadlytabrani]

Initial submitted fix doesn't work, this does

def get_available_namex(self, name):
        while self.exists(name):
            dname, fname = os.path.split(name)
            try:
                dot_index = fname.rindex('.')
            except ValueError:
                fname += '_'
            else:
                fname = fname[:dot_index] + '_' + fname[dot_index:]
            name = os.path.join(dname, fname)
        return name

[dc]

Little improvement:

def get_available_namex(self, name):
    dir_name, file_name = os.path.split(name) 
    while self.exists(name):
        try:
            dot_index = file_name.rindex('.')
        except ValueError:
            file_name += '_'
        else:
            file_name = file_name[:dot_index] + '_' + file_name[dot_index:]
        name = os.path.join(dir_name, file_name)
    return name

[fadlytabrani]

Patch attached

The line :

dir_name, file_name = os.path.split(name)

should be within the while loop to acommodate checking more than once

def get_available_name(self, name): 
    while self.exists(name):
		dir_name, file_name = os.path.split(name)
        try:
            dot_index = file_name.rindex('.')
        except ValueError:
            file_name += '_'
        else:
            file_name = file_name[:dot_index] + '_' + file_name[dot_index:]
        name = os.path.join(dir_name, file_name)
    return name

[dc]

Replying to fadlytabrani:

The line should be within the while loop to acommodate checking more than once

Nope. dir_name never changes and it should be calculated once. But your version calls unneeded split every iteration. Just check in debugger (or something like that) if you unsure.

[fadlytabrani]

Replying to dc:

Replying to fadlytabrani:

The line should be within the while loop to acommodate checking more than once

Nope. dir_name never changes and it should be calculated once. But your version calls unneeded split every iteration. Just check in debugger (or something like that) if you unsure.

Thanks dc, you're correct :) I missed that one

Have uploaded correct path

[dc]

Still a bug which must be fixed.

[(none)]

Milestone post-1.0 deleted

[Andy Durdin]

Added tests, and updated the patch to unified diff format.

[Gary Wilson]

(In [10702]) [1.0.X]: Fixed #9610 -- Fixed duplicate uploaded file name mangling when directory contained a dot and file didn't. Based on patches from fadlytabrani and adurdin.

Backport of r10701 from trunk.

[Jacob]

Milestone 1.1 deleted