- Committer: John Arbash Meinel
- Date: 2010-08-10 20:03:44 UTC
- Revision ID: john@arbash-meinel.com-20100810200344-6muerwvkafqu7w47
Rework things a bit so the logic can be shared.
It turns out that some of the peak memory is actually during the inventory
to string to bundle translations. So re-use the refcount logic there.
This actually does show a decrease in peak memory.
Specifically 'cd bzr.dev; bzr send ../2.2' drops from 221MB peak to 156MB.
We don't speed anything up (16.5s both ways), but peak memory is quite
a bit better.
It turns out that some of the peak memory is actually during the inventory
to string to bundle translations. So re-use the refcount logic there.
This actually does show a decrease in peak memory.
Specifically 'cd bzr.dev; bzr send ../2.2' drops from 221MB peak to 156MB.
We don't speed anything up (16.5s both ways), but peak memory is quite
a bit better.
| Filename | Latest Rev | Last Changed | Committer | Comment | Size | ||
|---|---|---|---|---|---|---|---|
 ..
|
|||||||
__init__.py |
5346.1.2 | 14 years ago | Vincent Ladeuil | Deleted BundleSerializer.write() which was never p | 7 KB |
|
|
|
v08.py |
4415.2.2 | 15 years ago | Martin Pool | Remove one use of DummyProgress | 20.2 KB |
|
|
|
v09.py |
4183.7.1 | 15 years ago | Sabin Iacob | update FSF mailing address | 2.4 KB |
|
|
|
v4.py |
5374.2.5 | 14 years ago | John Arbash Meinel | Rework things a bit so the logic can be shared. I | 30.3 KB |
|
|