summaryrefslogtreecommitdiff
path: root/eventlet/db_pool.py
Commit message (Collapse)AuthorAgeFilesLines
* Reformat with autopep8Jakub Stasiak2014-10-111-35/+69
|
* PEP-8 fixesSergey Shepelev2014-08-271-2/+40
|
* db_pool: customizable connection cleanup function; Thanks to Avery FaySergey Shepelev2014-07-101-73/+102
| | | | | | | | | https://github.com/eventlet/eventlet/pull/64 Also: - PEP8 - except Exception - .put() must not catch SystemExit
* Python3: __bool__, im_code, im_funcVictor Sergeyev2014-04-241-2/+5
|
* python3 compat: print() function syntaxSergey Shepelev2013-12-031-2/+4
|
* db_pool: Fix pool.put() TypeError with min_size > 1; Thanks to Jessica QiSergey Shepelev2012-12-131-9/+7
| | | | https://bitbucket.org/which_linden/eventlet/issue/125
* Merged in pjscott/eventlet (pull request #1)Sergey Shepelev2012-02-061-1/+3
|\
| * Fix db_pool .clear() when min_size > 0Sergey Shepelev2011-10-121-1/+3
| |
* | Make db_pool close connections after time-out as documented,Peter Scott2011-08-181-5/+9
| | | | | | | | | | and fix get-after-close race condition that occurs when using TpooledConnectionPool.
* | Add MySQL's insert_id() method to connection pool's proxy class.Peter Scott2011-07-201-0/+1
|/
* Removed cancelled attribute from timer class, it's not used by anything and ↵Ryan Williams2010-04-181-2/+1
| | | | is just deadweight at this point.
* Fixed a few small things.Ryan Williams2010-02-211-1/+0
|
* Version-specific test changes, plus name aliasing in db_pool.Ryan Williams2010-02-161-4/+4
|
* Imported Timeout class from gevent, deprecated exc_after, replaced ↵Ryan Williams2010-02-151-4/+5
| | | | TimeoutError with Timeout (should be seamless).
* Thanks tjd for catching this poor except clause.Ryan Williams2010-02-111-1/+1
|
* Extended tpool's autowrap functionality to use the name of the attribute, ↵Ryan Williams2010-02-061-7/+2
| | | | which is a lot cleaner than type-checking (which is still supported). Added psycopg tests to db_pool tests.
* Oops.Ryan Williams2010-01-031-1/+1
|
* Fix for rtyler's issue with psycopg2. Turns out violations of DB-API 2.0 ↵Ryan Williams2010-01-031-6/+6
| | | | are difficult to spot in various libraries.
* Removed saranwrap as an option for making db connections nonblocking in ↵Ryan Williams2009-12-151-19/+0
| | | | db_pool. Trying to keep changelog updated with my changes.
* Corrected PEP8 behavior, thanks SergeyRyan Williams2009-11-041-9/+6
|
* Removed copyright headers from individual files, added LICENSE and AUTHORS ↵Ryan Williams2009-10-031-22/+0
| | | | files to give credit instead. Linked authors list into documentation.
* Ammended eventlet Sphinx documentation. This removed some of the ↵Huin Linden2009-09-171-2/+3
| | | | Doxygen-isms, and introduced some more complete linking between parts of the documents.
* De-doxygenified various modules.Ryan Williams2009-08-061-5/+7
|
* Moved db_pool docs from the .py to the .rstRyan Williams2009-07-141-68/+0
|
* Disabled __del__ doing a put back to connection pool (best practice has ↵Ryan Williams2009-06-171-1/+3
| | | | always been to do put in a finally block). Disabled two_simultaneous_connections test since it's timing-sensitive. Reenabled these tests in runall.py.
* db_pool: fix the doctestsDenis Bilenko2009-06-151-9/+8
|
* kill trailing whitespaceDenis Bilenko2009-06-121-41/+41
|
* move copytright and license out of docstring in the commentsDenis Bilenko2009-06-121-24/+22
|
* db_pool: remove duplicate copy pasted methodsDenis Bilenko2009-06-081-5/+0
|
* remove a number of unused importsDenis Bilenko2009-06-081-1/+0
|
* import db_pool.py from https://bitbucket.org/which_linden/eventlet/Denis Bilenko2009-05-191-26/+55
|
* Some introductory/overview docs added to db_pool, moved DatabaseConnector ↵rdw2008-09-241-39/+111
| | | | class towards bottom of file.
* Refactored to create connect() classmethods that return the appropriate ↵rdw2008-09-071-13/+42
| | | | database connection. These will hopefully be independently useful.
* Tweaked expiry so it's microscopically faster, small changes, making tests ↵rdw2008-09-071-5/+7
| | | | easier to read.
* Changed ConnectionTimeout to ConnectTimeout, added unit test for connection ↵rdw2008-09-061-7/+6
| | | | timeout, and a benchmark unit test so we can see how much overhead the various pools add. Added testing of raw connections to the mix.
* Added connection timeout argument to db_pool.rdw2008-09-051-19/+30
|
* Get the connector in on the max_idle and max_age party.rdw2008-08-131-1/+7
|
* Fixed all the db_pool tests.rdw2008-08-131-11/+25
|
* Time-based expiry for db_pool. This adds the ability to expire connections ↵rdw2008-08-131-33/+151
| | | | by idleness and also to cap them to a finite lifespan (which will come in handy for e.g. mysql, which holds some locks on a per-connection basis).
* [svn r134] Fixes for some interesting issues discovered only after our count ↵which.linden2008-07-031-0/+16
| | | | of unit tests using database connections crested 100.
* [svn r93] "nonblocking" support in tpool, to be exploited by dbmgr, so thatseeping.blister2008-02-191-1/+1
| | | | nonblocking DB calls don't need to context-switch to a native thread.
* [svn r88] Refactored the body-parsing logic that used to be in read_body() ↵which.linden2008-02-131-1/+2
| | | | into parsed_body() and added some docs. This means that if you were calling read_body expecting to get anything but a string out, you should change to parsed_body() instead. Refactored tests.py a little bit.
* [svn r81] Added a RawConnectionPool, which can be used to get -pooled- DB connsseeping.blister2008-01-251-0/+6
| | | | with neither saranwrap nor tpool wrappering.
* [svn r74] make a subclass of db_pool that uses tpool.Proxy instead of ↵which.linden2008-01-041-20/+42
| | | | saranwrap. Unit tests for all possible combinations thereof.
* [svn r63] See what happens when you commit without testing? This is improved.which.linden2007-12-101-2/+9
|
* [svn r62] Cleaned up the two different paths to closing a ↵which.linden2007-12-101-8/+14
| | | | PooledConnectionWrapper, so that they are the same: the put method of ConnectionPool.
* [svn r61] Finally solved the bool(saranwrapped) problem. Moved the tests to ↵which.linden2007-12-071-1/+2
| | | | use the time module since uuid.getnode() was hanging forever on this machine. Cleared out the pool pointer on close in db_pool.
* [svn r60] A wrapper for the pooled connections so that clients can call ↵which.linden2007-12-061-5/+88
| | | | close() with impunity. Also it resolves the annoying bool() limitation of saranwrapped objects.
* [svn r59] Moved saranwrap to eventlet, moved mysql_pool over and abstracted ↵which.linden2007-12-051-0/+94
it a bit to db_pool. It *should* be compatible with other DB-API implementations but no testing has been done.