Smart server leaks memory each commit
Bug #494406 reported by
Gareth White
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Bazaar |
Fix Released
|
Medium
|
John A Meinel | ||
2.0 |
Fix Released
|
Undecided
|
John A Meinel |
Bug Description
When running a smart server, bzr.exe leaks about 200-400 KB per commit. To reproduce:
* start bzr serve --allow-writes
* bzr init bzr://localhost
* bzr checkout --lightweight bzr://localhost
* cd testcheckout
* bzr commit -m "test" --unchanged
* (repeat previous command)
Every time you repeat the "bzr commit", the memory usage of bzr.exe increases by about 200-400 KB. I've tested this with both Bazaar 2.0.1 and 2.1.0b3 with Python 2.5 on Windows XP SP3.
Related branches
lp:~jameinel/bzr/2.0.4-494406-serve-memory-leak
Merged
into
lp:bzr/2.0
- Andrew Bennetts: Approve
-
Diff: 38 lines (+8/-2)2 files modifiedNEWS (+7/-1)
bzrlib/_bencode_pyx.pyx (+1/-1)
- bzr-core: Pending requested
-
Diff: 107 lines (+40/-15)3 files modifiedNEWS (+12/-1)
bzrlib/_bencode_pyx.pyx (+1/-1)
bzrlib/export/dir_exporter.py (+27/-13)
Changed in bzr: | |
assignee: | nobody → John A Meinel (jameinel) |
milestone: | none → 2.1.0b4 |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Gareth White wrote: /testbranch /testbranch testcheckout
> Public bug reported:
>
> When running a smart server, bzr.exe leaks about 200-400 KB per commit. To reproduce:
> * start bzr serve --allow-writes
> * bzr init bzr://localhost
> * bzr checkout --lightweight bzr://localhost
> * cd testcheckout
> * bzr commit -m "test" --unchanged
> * (repeat previous command)
>
> Every time you repeat the "bzr commit", the memory usage of bzr.exe
> increases by about 200-400 KB. I've tested this with both Bazaar 2.0.1
> and 2.1.0b3 with Python 2.5 on Windows XP SP3.
>
> ** Affects: bzr
> Importance: Undecided
> Status: New
>
200-400KB could easily be 'noise' that will be cleaned up when the
python garbage collector runs. It depends more on whether the leak is
sustained after say 100 commits (then equaling 20-40MB or so). Also note
that while you are committing repeatedly, the size of the repository is
*also* increasing. So it isn't like the first run is dealing with the
same content that the last run is dealing with.
importance: medium
status: incomplete
tags: memory
John enigmail. mozdev. org/
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://
iEYEARECAAYFAks f8c8ACgkQJdeBCY SNAAP7jgCgyhHzm JNeOjxbLIbNvdun EzwN jx/0Bs+ pnGSmmzDGy6SM12 L8s
/hQAoJ/
=higI
-----END PGP SIGNATURE-----