Ryan Williams 488f4abd42 Added cool little debug mode to Hub which makes it easier to see why a particular fileno is in the hub by recording the traceback of when the fileno was added. Enable this mode by doing api.get_hub().debug = True. Tweaked tpool so that it only starts up its thread pool when you actually call execute(). I think this is not a performance hit and it will make it somewhat nicer to use. Some other randomc changes like more explicit package name in test_named so it passes when called from within nose.

2009-08-24 22:35:25 -07:00

..

green

Some slight 2.6 compatibility changes

2009-08-13 17:59:33 -07:00

hubs

Added cool little debug mode to Hub which makes it easier to see why a particular fileno is in the hub by recording the traceback of when the fileno was added. Enable this mode by doing api.get_hub().debug = True. Tweaked tpool so that it only starts up its thread pool when you actually call execute(). I think this is not a performance hit and it will make it somewhat nicer to use. Some other randomc changes like more explicit package name in test_named so it passes when called from within nose.

2009-08-24 22:35:25 -07:00

support

remove eventlet.support.pycurls

2009-06-15 11:33:25 +07:00

twistedutil

fix twistedutil.protocol.ValueQueue in regard to new Queue class

2009-06-22 15:56:34 +07:00

__init__.py

Merge from eventlet 0.8 branch

2009-07-13 17:54:20 -07:00

api.py

Refactored the hubs again, now we support multiple readers/writers on a single socket, not sure if that's useful or not but it's sure better than silently unscheduling the original reader/writer, which was the original behavior.

2009-08-13 17:25:28 -07:00

backdoor.py

add __main__ section to backdoor.py

2009-06-22 15:32:57 +07:00

corolocal.py

Slight doc shuffling

2009-07-14 00:21:10 -07:00

coros.py

Minor doc change.

2009-07-14 00:36:55 -07:00

db_pool.py

De-doxygenified various modules.

2009-08-06 08:56:11 -07:00

greenio.py

Brought api_test up to date with code changes, the daggy fix approach appears to have not worked so well just now. I've finally decided that masking the ZeroReturnError is not correct (naked SSLConnection objects raise it), and therefore expect it in the ssl test.

2009-08-23 17:54:50 -07:00

pool.py

Added waiting method back to pool.Pool that was lost, plus a unit test so it doesn't get lost again.

2009-08-14 17:18:56 -07:00

pools.py

Removed ConnectionPool which didn't seem usable, and also Pool.fan which has been supplanted by better abstractions.

2009-07-14 00:20:57 -07:00

proc.py

proc: add deprecation warning to trap_errors()

2009-06-20 18:13:12 +07:00

processes.py

Fixed saranwrap bug on 2.6; turns out the warning messed with the communication.

2009-08-24 02:21:57 -07:00

saranwrap.py

Fixed saranwrap tests which have appear to have always been broken. Removed support for raising string exceptions, because who does that anymore, and it was getting annoying to see the warning every time.

2009-08-23 18:42:02 -07:00

timer.py

kill trailing whitespace

2009-06-12 15:31:40 +07:00

tpool.py

Added cool little debug mode to Hub which makes it easier to see why a particular fileno is in the hub by recording the traceback of when the fileno was added. Enable this mode by doing api.get_hub().debug = True. Tweaked tpool so that it only starts up its thread pool when you actually call execute(). I think this is not a performance hit and it will make it somewhat nicer to use. Some other randomc changes like more explicit package name in test_named so it passes when called from within nose.

2009-08-24 22:35:25 -07:00

util.py

Typo

2009-08-14 17:34:21 -07:00

wsgi.py

Finished up SSL.Connection compatibility. Backwards compatibility change: it used to be that you could just call close() on an ssl socket, but now you have to call shutdown() and then close(). This is in line with the normal OpenSSL behavior and means that we no longer penalize correctly-written clients. Fixed wsgi and api_test to reflect this changed assumption.

2009-07-19 03:55:54 -07:00