root/tags/1.05/TODO

Revision 153, 2.2 kB (checked in by bradfitz, 2 years ago)

brackup-restore's verbose flag is more verbose now, showing files
as they're restored.

brackup-restore can restore from an encrypted *.brackup file now,
firing up gpg for user to decrypt to a tempfile

Line 
1-- in Restore.pm:
2      # TODO: inefficient!  we don't want to download the chunk from the
3      # target multiple times.  better to cache it locally, or at least
4      # only fetch a region from the target (but that's still kinda inefficient
5      # and pushes complexity into the Target interface)
6
7-- only use tempfiles in a 0700 directory under /tmp/brackup-$USER/ or
8   wherever.  or option, when using encryption, to only doing
9   in-memory tempfiles, to avoid hitting disk?
10
11-- <lj user=grahams>: if a file which existed when Brackup was
12  "discovering files" subsequently goes away while Brackup is working
13  it's magic (like a vim .swp file) Brackup tosses an error and bails:
14
15-- shouldn't backup the digest file (.brackup-digest.db)
16
17-- don't skip files ending in whitespace.  figure out why GPG barfs.  would
18   the metafile also barf, having a trailing \r or \n?
19
20-- "brackup-target <target> gc": find/clean orphan chunks on a target
21   from metafiles
22
23-- figure out how files with intenal \r or \n in filename get written
24   to metafile.  need to be escaped?
25
26-- FUSE script to mount a *.brackup file
27
28-- tool to clean digestcache, based on prefixes, looking for files that no longer
29   exist or have new mtimes, etc?  or keep track of "last used" date
30   field in the digestcache and just delete things that are too old?
31   but then no longer just a dictionary.  SQLite would work, but we'd ideally
32   like lots of dumber cache mechanisms.  maybe a ->clean method is optional?
33   then a memcached backend/etc doesn't have to use it.
34
35-- Tools to rebuild your inventory database from the target's enumeration
36   of its chunks and the target's *.brackup metafiles isn't yet done, but
37   would be pretty easy. 
38
39-- ionice stuff.  network nice stuff.
40
41-- make tests pass without 'noatime' mount option
42
43-- --ignore-debian-files option (if managed by a package management
44   system (and not an unmodified conffile), don't back it up)
45
46-- reuse tempfiles in Chunk.pm as Restore.pm does
47
48-- restoring from existing config file.
49
50-- better test coverage.  currently at 84%.  should ignore
51   test coverage of test modules.  move Brackup::Test to
52   t/lib/ probably
53
54-- should do TODOs in code.  :)
Note: See TracBrowser for help on using the browser.