Memory leaks in cprocessors DecimalResultProcessor

[Former user]

There are two memory leaks in the cprocessors DecimalResultProcessor.

One is a missing Py_DECREF in the process method. Also, the DecimalResultProcessor itself lacked a dealloc method to deref the provided type and formatter string.

Tests are included; obviously you have to compile and use the cextensions to really exercise the code. :)

Pullable from https://bitbucket.org/taavi_burns/sqlalchemy .

[Former user]

Pullable specifically from https://bitbucket.org/taavi_burns/sqlalchemy/changeset/f19d7e5b972e60c843acb3389e8ae07eb6043818

[Mike Bayer]

nice !

[Former user]

Shortened tests with https://bitbucket.org/taavi_burns/sqlalchemy/changeset/df1d3cb9d575dfa38c7cb81f5486a80ddf114b3f

[Former user]

(original author: taavi) Pushed http://www.sqlalchemy.org/trac/changeset/6935%3A2406615164f0?old=6930%3A3b0167978a8e

[Former user]

(original author: ged) Doh! Seems like I didn't check for memleaks after implementing the scale argument... Thanks for the fix.

Also, the UnicodeResultProcessor suffers from the same small (once per processor) memleak: it needs a dealloc method -- I guess I thought they were optional when I wrote that code. I'll open a ticket for that.

[Mike Bayer]

this one's fixed right ?

[Former user]

(original author: taavi) Yup, it's fixed. Wasn't sure what the ticket closing procedure was at the tie, and just got around to the "what do I do about it" to-do now. ;)

[Mike Bayer]

Removing milestone: 0.6.6 (automated comment)