lp://staging/~free.ekanayaka/storm/fix-order-by-expressions-in-ref-set

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp://staging/storm at revision 455

Thomas Herve (community): Approve on 2013-01-03

Storm Developers: Pending requested 2013-01-03

Diff: 0 lines api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Bug #620369: RuntimeError when using Desc in ReferenceSet order_by

Medium

Fix Released

Link a bug report

Related blueprints

457. By Free Ekanayaka on 2013-01-21

Updated NEWS file

456. By Free Ekanayaka on 2013-01-21

Lints

455. By Free Ekanayaka on 2013-01-03

Handle SuffixExpr-derived instances in PropertyResolver

454. By Thomas Herve on 2012-12-19

Merge count-limited-union [r=therve] [a=radix]

Remove the order_by when creating aggregates on ResultSet, so that
replace_columns doesn't complain.

453. By Free Ekanayaka on 2012-09-28

Use 'IF EXISTS' when running the 'DROP TABLE patch' statement in Schema.drop, in case the user-supplied drop statements have already dropped the patch table [trivial] [r=therve]

452. By Free Ekanayaka on 2012-08-27

In ZStormResourceManager set the schema store URI all the times, to let clean() run the schema delete statements if needed [trivial] [r=therve]

451. By Free Ekanayaka on 2012-08-22

Merge fast-schema-check-in-resource-manager [f=1036158] [r=lifeless,therve]

This branch makes ZStormResourceManager's setup a bit faster between subsequent
test runs, by saving timestamps of schema's patch packages and using them to
decide whether to upgrade the schema or not. For example one can set
ZStormResourceManager.schema_stamp_dir to something like /tmp/my-project-zstorm-stamp
to achieve faster and yet safe setups across all project's branches.

Furthermore, it changes ZStormResourceManager.make() to not create all stores
upfront, but lazily, when they are actually used by tests or code under test.

450. By Guilherme Salgado on 2012-06-28

Add connection commit/rollback tracer hooks and use them in the TimeoutTracer
to reset connection._timeout_tracer_remaining_time. [r=radix,allenap]

This is to fix a bug in which the first query of a transaction could run
with no timeout set because of the _timeout_tracer_remaining_time left
in the connection during the previous transaction.

449. By Free Ekanayaka on 2012-06-12

Merge tpc-support [f=132485] [r=stub,therve]

This branch adds support for two-phase commits, using the DB API
version 2.0 (only implemented by psycopg2 at the moment). An example
usage is:

xid = Xid(0, "my-txn", "my-branch") # This models a XA-compliant transaction ID
store.begin(xid)
store.execute(...)
store.prepare()
store.commit()

Or using ZStorm, just call zstorm.set_default_tpc("my-store", True) and ZStorm
will automatically use two-phase commit for that store when running
transaction.commit().

In order to transparently use store.commit() and store.rollback(), without
introducing two new ad-hoc APIs (e.g. Store.tpc_commit/Store.tpc_rollback),
the code of the Connection class keeps track of whether we are in two-phase
transaction or in a regular one, and uses the raw connection's DB API
tpc_commit/tpc_rollback or commit/rollback methods accordingly.

448. By Sidnei da Silva on 2012-06-04

Fix wsgi.make_app() to not make assumptions about how to consume data from whatever is returned by its inner app() call [r=sidnei, therve]