lp:~niemeyer/storm/timedelta-improvements

Created by Gustavo Niemeyer and last modified
Get this branch:
bzr branch lp:~niemeyer/storm/timedelta-improvements
Only Gustavo Niemeyer can upload to this branch. If you are Gustavo Niemeyer please log in for upload directions.

Branch merges

Related bugs

Related blueprints

Branch information

Owner:
Gustavo Niemeyer
Project:
Storm
Status:
Merged

Recent revisions

196. By Gustavo Niemeyer

Merged from trunk.

195. By Gustavo Niemeyer

Test _parse_interval() through TimeDeltaVariable instead of
getting it directly.

194. By Gustavo Niemeyer

- Rewrote interval parser for TimeDeltaVariable(). Many more formats
  are accepted now for all backends. The supported format is based
  on PostgreSQL intervals, which are pretty flexible.
- Added support for TimeDeltas in SQLite and MySQL. Both don't have
  internal support for intervals, so they should use a string column
  for that.
- Fixed TimeDelta precision issue in PostgreSQL by sending a string
  representation of the TimeDelta instead of relying on what psycopg2
  does.

193. By Gustavo Niemeyer

Updating NEWS file.

192. By Gustavo Niemeyer

Added TimeDelta to locals. [trivial]

191. By Gustavo Niemeyer

Merged insert-returning branch [r=radix,jkakar]

This branch implements three important changes:

- It changes the flushing mechanism to not load values inserted
  in the database at flush time. Instead, undefined variables are
  set to AutoReload, and resolved once first accessed.

- It implements support in the postgres backend to use the RETURNING
  extension of the INSERT statement to retrieve the primary key on
  inserts for object identity mapping (for versions >= 8.2)

- It prevents a pathological case which happens when a statement like
  "SELECT ... WHERE table.id = currval(...)" is executed in postgres.
  In these cases, the index of "table.id" isn't used (now used in
  versions < 8.2 only).

190. By Gustavo Niemeyer

Fixed compilation of currval() in the postgres backend when using
escaped schemas, tables, and/or column names, and also added one
more test using the public schema to verify that it basically works
end-to-end. [r=radix,bigkevmcd]

189. By Gustavo Niemeyer

A couple of fixes: [r=jkakar,oubiwann] [f=146580]

- Schema names are accepted in __storm_table__ when using
  PostgreSQL as the database (e.g. "schema.table"). (#146580,
  reported by James Mayfield)
- Tables named with reserved keywords are properly escaped.

188. By Gustavo Niemeyer

Changed PostgresDisconnectionTest to use the STORM_POSTGRES_HOST_URI
environment variable, if set, and use STORM_POSTGRES_URI as a fallback,
if it actually has a host set. STORM_POSTGRES_URI without the host
will disable the test class. STORM_POSTGRES_HOST_URI without a host
is an error. [trivial]

187. By Jamu Kakar

- Replaced iteritems with items in ZStorm.iterstores to avoid the
  problem where (because a WeakKeyDictionary is used) a store is
  de-allocated during iteration resulting in a RuntimeError [trivial].

This branch contains Public information 
Everyone can see this information.

Subscribers

No subscribers.