From 2f117ec7b75a5a3b5e486e115de58645c901d2fc Mon Sep 17 00:00:00 2001 From: Joey Hess Date: Sat, 9 Feb 2019 14:03:59 -0400 Subject: [PATCH] thought --- doc/todo/import_tree.mdwn | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/doc/todo/import_tree.mdwn b/doc/todo/import_tree.mdwn index dfa816b9c3..76fabae1b1 100644 --- a/doc/todo/import_tree.mdwn +++ b/doc/todo/import_tree.mdwn @@ -129,6 +129,15 @@ to a given remote. This reduces the use cases a lot though, and perhaps so far that the import tree feature is not worth building. The adb special remote needs both. +Is this race really a significant problem? One way to look at it is +analagous to a git pull overwriting a locally modified file, which should +not happen. Git may in some condition not notice the modification before +overwrite, but if so that would be considered a surprising bug; it's +certianly *possible* for git pull to avoid such a race whether or not it +currently does. But another way to look at it is the remote has two +writers, one of which is git annex exporttree and the other something +else, and it's up to the user to avoid them conflicting. + Really fixing this race needs locking or an atomic operation. Locking seems unlikely to be a portable enough solution.