update on variant
authorJoey Hess <joeyh@joeyh.name>
Fri, 27 Dec 2019 17:17:53 +0000 (13:17 -0400)
committerJoey Hess <joeyh@joeyh.name>
Fri, 27 Dec 2019 17:17:53 +0000 (13:17 -0400)
doc/bugs/A_case_where_file_tracked_by_git_unexpectedly_becomes_annex_pointer_file/comment_3_37fb6d6361068f57c2068270ceb72ac0._comment [new file with mode: 0644]

diff --git a/doc/bugs/A_case_where_file_tracked_by_git_unexpectedly_becomes_annex_pointer_file/comment_3_37fb6d6361068f57c2068270ceb72ac0._comment b/doc/bugs/A_case_where_file_tracked_by_git_unexpectedly_becomes_annex_pointer_file/comment_3_37fb6d6361068f57c2068270ceb72ac0._comment
new file mode 100644 (file)
index 0000000..f4bebf7
--- /dev/null
@@ -0,0 +1,16 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 3"""
+ date="2019-12-27T17:11:42Z"
+ content="""
+A variant of this where an annexed unlocked file is added first,
+then the file is touched, and then some other file is added
+with -c annex.largefiles=nothing does result in the clean filter sending
+the whole annexed file content back to git, rather than keeping it annexed.
+For whatever reason, git does not store that content in .git/objects or
+update the index for that file though, so it doesn't show up as a change.
+
+So *apparently* that variant is only potentially an expensive cat of a
+large annexed file, and does not need to be dealt with. Unless git
+sometimes behaves otherwise.
+"""]]