summaryrefslogtreecommitdiff
path: root/doc/devblog/day_463-465__back_buggy_external_special_remotes.mdwn
blob: f374a51cb6754883e2c1e93ef3da7e455ad6ec73 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
I'm back working on git-annex after a month away! It's good to be back, and
I've made some decent improvements this week. 

New features include a `GIT_ANNEX_VECTOR_CLOCK` environment variable that
may be useful for those using git-annex in a HIPPA compliance setting,
where timestamps are verboten (but verifying full compliance is up to
you!), and remote.name.annex-ignore-command and
remote.name.annex-sync-command configurations that let shell commands be
run to vary what remotes are used depending on eg, what network it's on.

Also, I took a look at the external special remote protocol, and noticed
two problems with it. First, keys with spaces in their names can't be used
with it. This only affects the WORM backend, and it seems no one has ever
run into the problem. Rather than complicate the implementation of external
special remotes, I decided to deprecate having spaces in key names. Which
is just asking for trouble anyway. So now there's a nice error message, and
a migration path.

The other problem was that the external special remote documentation
incorrectly said that a filename parameter never contained spaces. But in
fact, there are situations where it can. This was not a problem with the
protocol, but only with its documentation, and potentially with the
implementation of some special remotes. So, I've spent some time today
auditing every git-annex special remote that I know about. A few scripts
that are bundled with git-annex were buggy and got fixed, and I
filed bugs on 9 other external special remotes. A few did already get it
right!

Today's work was sponsored by Trenton Cronholm on 
[Patreon](https://patreon.com/joeyh/)