summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorJoey Hess <joeyh@joeyh.name>2016-09-27 09:48:30 -0400
committerJoey Hess <joeyh@joeyh.name>2016-09-27 09:48:30 -0400
commit97f7caa4bab95faa194d116f6691ac0fe0106551 (patch)
treed48259e133a151260ed073d5db3c1dd40bf63222
parent99524bf75d3c2b1713a3e82bd29b779e3d535a88 (diff)
comment
-rw-r--r--doc/todo/include_mtime_into_metadata/comment_1_9e75c36722c194fecef316ae99d1d75e._comment25
1 files changed, 25 insertions, 0 deletions
diff --git a/doc/todo/include_mtime_into_metadata/comment_1_9e75c36722c194fecef316ae99d1d75e._comment b/doc/todo/include_mtime_into_metadata/comment_1_9e75c36722c194fecef316ae99d1d75e._comment
new file mode 100644
index 0000000..ebe5e94
--- /dev/null
+++ b/doc/todo/include_mtime_into_metadata/comment_1_9e75c36722c194fecef316ae99d1d75e._comment
@@ -0,0 +1,25 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 1"""
+ date="2016-09-27T13:31:57Z"
+ content="""
+Are there any config files in /etc whose mtime influences the program they
+configure in some way? That would be a very good reason to include the
+mtime.
+
+If the goal is to keep track of multiple manual changes to /etc, the best
+thing to do is to manually run etckeeper commit after making the change.
+Relying on the daily autocommit is at best a fallback.
+
+Need to consider the resource impact of adding mtimes to /etc/.etckeeper
+would not be very bad. The files are already statted, so no extra overhead
+there.
+
+How about the size increase of /etc/.etckeeper?
+For each file in /etc, a line like `maybe touch -d '1970-01-01 00:00Z' foo` would
+be needed. (That seems the most compact available way to specifiy a time to
+touch.) On my system, that adds 136kb to /etc/.etckeeper, a little bit more
+than doubling its size.
+
+That doesn't seem too bad a resource impact.
+"""]]