/stdhome

To get this branch, use:
bzr branch http://bzr.ed.am/stdhome
68 by Tim Marston
added tools
1
* perform update with conflict
2
* performing an "add", when the repo is not up-to-date (i.e., there are updates
3
   pending) just prints "stdhome add: bzr commit", with no other explaination as
4
   to what's going on
5
* add --all-files to `stdhome st` to show status of all files in repo?  or
6
   should we add `stdhome ls`?
7
* symlink substitutes don't acept a symlink in the repo and a directory on the
8
   FS (as well as the other way around).  Should they?  It means ~/src (which is
9
   a symlink in my home repo) shows as modified when compared to the symlink of
10
   the same name on oak.
11
* skip unchanged directories
12
   now that update, revert, resolve and init report on home dir changes, resolve
13
   and revert show all directories as being modified (because we always copy
14
   them out without checking for changes)
15
* add ~/bin/rebuild-tags!
16
* look at what need to be done before v0.1
17
* make the copy walker accept symlinks in place of files (as well as
18
   directories) if they are on the symlink accept list
19
* check that bin/quicktile.py on tims-laptop is an acceptable symlink
20
* !!! TAG 0.1 !!!
21
* check directory permissions before updating them
22
* during copy-out (not copy-in), show what is being changed with lines like
23
   "<SP><SP><OP><SP><FILE>" where OP is A (add), D (delete), M (modified) and K
24
   (kind change)
25
* currently, verbose print lines for bzr stuff are in the commands -- should
26
   they be in the bzr module?
27
* make the copy walker accept symlinks that differ if they are on the symlink
28
   accept list (so that symlinks can be pointed elsewhere in ~/)
29
* add code that can determine a relative file from a full file and a basepoint
30
* when ignoring files in the walker, populate the list of file more
31
   inteligently.  ".bzr" (and ".bzrignore") need to come from the VCS backend.
32
   ".stdhome", ".stdhomerc" and/or ".stdhomeignore" need to be converted to full
33
   filenames and then their relative file names derived from the configured home
34
   directory
35
* KNOWN ISSUE: handle renames as renames (rather than delete + add) because a
36
   directory renamed in the repo could contain other stuff that isn't versioned
37
   in ~/ which is lost, at the moment
38
* KNOWN ISSUE: decide on whether to implement remote --keep flag (see plan:483);
39
   currently, on the remote end, stuff deleted in repo during an update deletes
40
   files in ~/