[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Duplicity-talk] Re: Macports & Duplicity 0.4.10
From: |
Darik Horn |
Subject: |
Re: [Duplicity-talk] Re: Macports & Duplicity 0.4.10 |
Date: |
Mon, 31 Mar 2008 10:26:47 -0500 |
User-agent: |
Thunderbird 2.0.0.12 (X11/20080227) |
> and it was causing regression to break.
On which regression test was it breaking?
I am interested in properly fixing this glitch, but run-all-tests.sh
currently reports "OK" for me.
> If you need the prefix to list() the file, then you
> need to add the prefix when you put()/get() the file
> so it will be there when you list it.
The +'d' in list() is used as a filter, so it does not compose a key
name. The behavior of file_naming.py should always result in the 'd'
appearing in the remote_filename parameter to put() and get().
I added this code primarily to reduce 500 and 503 errors from Amazon S3
on large buckets, and secondarily as a work-around for an early boto
bug. The duplicity retry mechanism gets choked when a bucket goes past
past 100,000 keys (ish).
Relying on the stable behavior of file_naming.py is unwholesome, but my
environment is large enough to need the optimization. (I should have
commented it, so my bad.)