From c1affb549de74ec13d8933732fed8a9859bf50b6 Mon Sep 17 00:00:00 2001 From: Joey Hess Date: Wed, 19 Mar 2014 16:55:22 -0400 Subject: devblog --- .../day_137-138__bug_triage_and_too_much_windows.mdwn | 15 +++++++++++++++ 1 file changed, 15 insertions(+) create mode 100644 doc/devblog/day_137-138__bug_triage_and_too_much_windows.mdwn (limited to 'doc/devblog/day_137-138__bug_triage_and_too_much_windows.mdwn') diff --git a/doc/devblog/day_137-138__bug_triage_and_too_much_windows.mdwn b/doc/devblog/day_137-138__bug_triage_and_too_much_windows.mdwn new file mode 100644 index 000000000..feff2ba8a --- /dev/null +++ b/doc/devblog/day_137-138__bug_triage_and_too_much_windows.mdwn @@ -0,0 +1,15 @@ +Yesterday, worked on cleaning up the todo list. Fixed Windows slash problem +with rsync remotes. Today, more Windows work; it turns out to have been +quite buggy in its handling of non-ascii characters in filenames. Encoding +stuff is never easy for me, but I eventually managed to find a way to fix +that, although I think there are other filename encoding problems lurking +in git-annex on Windows still to be dealt with. + +Implemented an interesting metadata feature yesterday. It turns out that +metadata can have metadata. Particularly, it can be useful to know when a +field was last set. That was already beeing tracked, internally (to make +union merging work), so I was able to quite cheaply expose it as +"$field-lastchanged" metadata that can be used like any other metadata. + +I've been thinking about how to implement [[required_content]] expressions, +and think I have a reasonably good handle on it. -- cgit v1.2.3