The DB() constructor will error if it's passed a str object as the path on 
Python 3.

Seems like you should call path.encode("utf-8") before passing it to the 
underlying library -- this should work on Python 2 and Python 3.

# python inserter_maker.py _dummydb _dummy-items < big-list
Loaded 71 new_encoded_urls from db
10000   read at 10170   URLs/sec, 0     inserted/queued, 10000  already in db
20000   read at 10042   URLs/sec, 0     inserted/queued, 10000  already in db
30000   read at 10309   URLs/sec, 0     inserted/queued, 10000  already in db
40000   read at 10070   URLs/sec, 0     inserted/queued, 10000  already in db
^CException KeyboardInterrupt in <bound method _PointerRef.close of 
<leveldb._PointerRef object at 0x7f1651bac190>> ignored
50000   read at 10143   URLs/sec, 0     inserted/queued, 10000  already in db
60000   read at 10260   URLs/sec, 0     inserted/queued, 10000  already in db

# python --version
Python 2.6.6

It was almost certainly running this when I hit Ctrl-C:

    packed_next_item_id = db.get("$next_item_id$", fill_cache=False)
    item_id, = unpack("<i", packed_next_item_id)

    batch = db.newBatch()
    for u in new_encoded_urls:
        db.putTo(batch, reversed_encoded_url(u), packed_next_item_id)
    db.putTo(batch, "$next_item_id$", pack("<i", item_id + 1))
    db.write(batch)

My latest code is available at 
https://github.com/ludios/greader-item-maker/blob/master/inserter_maker.py

Three of the four LevelDB bindings I know of use the same module name 
("leveldb"). This causes havoc when I'm trying to test the various bindings, 
even if I try to be careful to not have them installed at the same time.

I know it's not only your fault, but please consider choosing a distinguishable 
module name (e.g. "leveldb_py") so at least you won't be part of the problem.

leveldb-py does not complain when a unicode object is used as a key (or when 
creating a scope) but produces some strange invalid behavior : 

* put still insert the value but the unicode key is not respected : the value 
is inserted with only the beginning of the key
* DBInterface created with scope(unicode) do not iterate correctly on the key 
corresponding to the given prefix.

If I use str(key) (where key is an unicode object) everything works as expected.

I think it would be nice to specify this in the documentation and maybe enforce 
it in the code by calling str(key) as it's rather easy to inadvertently use a 
unicode object thinking it's a string (I've been bitten by this while using the 
json module : I did not realize that the dict it was generating contained only 
unicode objects and not strings).


I can provide an example if needed.

Branch name:
iteration

Purpose of code changes on this branch:
Add iteration support to leveldb-py

When reviewing my code changes, please focus on:
General awesomeness.

After the review, I'll merge this branch into:
master

What steps will reproduce the problem?

db = leveldb.DB('mydb')
batch = leveldb.WriteBatch()
for x in xrange(10000):
    batch.put(str(x), str(x))
db.write(batch)

What is the expected output? What do you see instead?
When repeating the above snippets in a short time, it results in a segmentation 
fault.

What version of the product are you using? On what operating system?

Working on Linux Mint, using the git repository master branch.

The "Iterator" object is not recognized as a valid iterator in Python 3. Trying 
to use it gives this error:

  File "/Users/matt/Dropbox/workspace/env/py33whoosh/lib/python3.3/site-packages/leveldb.py", line 361, in range
    for row in self:
TypeError: iter() returned non-iterator of type 'Iterator'

You can fix this by adding a __next__() method to the object, e.g.:

    def __iter__(self):
        return self

    def __next__(self):
        return self.next()

leveldb-py has no setup.py, so I cannot install it with distutils (python 
setup.py install --user) or pip (pip install --user .)

I couldn't run the tests on Python 2.6 (on my CentOS 6.4 machine):

# python --version
Python 2.6.6

# python test_leveldb.py
Traceback (most recent call last):
  File "test_leveldb.py", line 762, in <module>
    main()
  File "test_leveldb.py", line 758, in main
    unittest.main(argv=sys.argv[:1], exit=False)
TypeError: __init__() got an unexpected keyword argument 'exit'

After removing the exit= argument:

# python test_leveldb.py  
...........................E........................E..
======================================================================
ERROR: test__contains__ (__main__.LevelDBTestCases)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "test_leveldb.py", line 342, in test__contains__
    with self.assertRaises(KeyError):
TypeError: failUnlessRaises() takes at least 3 arguments (2 given)

======================================================================
ERROR: test__contains__ (__main__.MemLevelDBTestCases)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "test_leveldb.py", line 342, in test__contains__
    with self.assertRaises(KeyError):
TypeError: failUnlessRaises() takes at least 3 arguments (2 given)

----------------------------------------------------------------------
Ran 55 tests in 75.682s

FAILED (errors=2)

plyvel.DB(db_dir, create_if_missing=False, compression=None)

pip install leveldb-py

ERROR: Could not find a version that satisfies the requirement leveldb-py (from versions: none)
ERROR: No matching distribution found for leveldb-py