lp:~jameinel/bzr/1.16-chk-direct

Created by John A Meinel on 2009-06-08 and last modified on 2009-06-18
Get this branch:
bzr branch lp:~jameinel/bzr/1.16-chk-direct
Only John A Meinel can upload to this branch. If you are John A Meinel please log in for upload directions.

Branch merges

Related bugs

Related blueprints

Branch information

Owner:
John A Meinel
Project:
Bazaar
Status:
Merged

Recent revisions

4428. By John A Meinel on 2009-06-18

Merge bzr.dev 4460 resolving NEWS

4427. By John A Meinel on 2009-06-18

The new add_inventory_by_delta is returning a CHKInventory when mapping from NULL
Which is completely valid, but 'broke' one of the tests.
So to fix it, changed the test to use CHKInventories on both sides, and add an __eq__
member. The nice thing is that CHKInventory.__eq__ is fairly cheap, since it only
has to check the root keys.

4426. By John A Meinel on 2009-06-17

NEWS entry about --2a improvement.

4425. By John A Meinel on 2009-06-17

Merge bzr.dev 4454 in prep for NEWS

4424. By John A Meinel on 2009-06-17

Pull out the common 'populate this CHKInventory' code out into a helper
and share it between the creators.

4423. By John A Meinel on 2009-06-17

Clean upt the test_inv tests that assumed _root_node was real and not just a key.

4422. By John A Meinel on 2009-06-17

Some cleanup. Move the check that from_dict works into test_chk_map.

4421. By John A Meinel on 2009-06-08

Change some asserts into raise: calls.

4420. By John A Meinel on 2009-06-08

Switch to using a single code path for from_dict().
Remove an extra pdb.set_trace() statement.

4419. By John A Meinel on 2009-06-08

Clean up the calls for '_create_inv_from_null' so they use the apis correctly.

In the end this shaves off as much as 2s (15.5s => 13.5s) for an initial commit of
a mysql tree. Some of that is potentially the InternalNode.map() fix, some of it is
not going through a regular Inventory and then back into an apply_delta loop, etc.

Stuff like InternalNode.map() has 2 node._current_size() calls, so that it can see
if the size changed so it knows to check for remap, etc, is wasted on an initial
build. And optimizing the 'build-from-scratch' is somewhat reasonable, since it is
the only time that we should be dealing with that many objects.

Branch metadata

Branch format:
Branch format 7
Repository format:
Bazaar RepositoryFormatKnitPack6 (bzr 1.9)
Stacked on:
lp:~bzr/bzr/trunk-old
This branch contains Public information 
Everyone can see this information.

Subscribers