Merge branch 'master' of ssh://git-annex.branchable.com

doc/bugs/assistant_doesn__39__t_notice_changes_to_gitignore_files.mdwn

@@ -0,0 +1,41 @@
+### Please describe the problem.
+
+Once the assistant daemon is started, it won't notice updates to any gitignore files.  This is probably because it spawns a long-lived `git check-ignore -z --stdin --verbose --non-matching` process which only reads gitignore files on startup.
+
+### What steps will reproduce the problem?
+
+- Start the assistant
+- Add a pattern to a relevant `.gitignore` or `.git/info/exclude`
+- (Just to be safe) wait for the assistant to commit and sync that change
+- Create a file matching the new ignore pattern
+- Observe the assistant committing the file to the repository despite it matching the new ignore pattern
+
+### What version of git-annex are you using? On what operating system?
+
+7.20191230-gef6d1e327 on openSUSE Leap 15.1
+
+### Please provide any additional information below.
+
+I assume that the obvious fix would be to watch for changes to gitignore files, and restart the `check-ignore` process accordingly.
+
+A workaround is to restart the assistant.  I've tested this and it works fine.
+
+Unsurprisingly the daemon log just shows the normal behaviour so it's probably not very useful, but including anyway just in case:
+
+[[!format sh """
+# If the problem is with the git-annex assistant, paste in .git/annex/daemon.log
+Everything up-to-date
+[2020-01-06 15:40:21.863988735] Committer: Committing changes to git
+(recording state in git...)
+[2020-01-06 15:40:22.024703938] Pusher: Syncing with peer1 
+[2020-01-06 15:40:23.025650786] Committer: Committing changes to git
+(recording state in git...)
+To tor-annex::<censored>.onion:5227
+   841db50..d47d59e  master -> synced/master
+
+# End of transcript or log.
+"""]]
+
+### Have you had any luck using git-annex before? (Sometimes we get tired of reading bug reports all day and a lil' positive end note does wonders)
+
+Yes, worked great for many years, but struggling to get the assistant to a point I can use it.  Main blockers are [[assistant_sometimes_removes_and_re-adds_whole_file]] and [[todo/wishlist__58___disable_automatic_commits]].

doc/todo/provide_windows_build_with_MagicMime.mdwn

@@ -0,0 +1,7 @@
+I forgot already why windows build is without having MagicMime support added, and failed to find a reason in the archives.
+But [there is demand](https://github.com/datalad/datalad/pull/3956#pullrequestreview-335604805) from Windows users! ;)
+
+Without such functionality we cannot consistently (cross-platform) use git-annex for many DataLad datasets, where text files (well -- soon "non-binary files") are configured to be committed directly to git.
+
+[[!meta author=yoh]]
+[[!tag projects/datalad]]

doc/todo/wishlist__58___disable_automatic_commits/comment_4_e021de3647a00aa0259a247441de50eb._comment

@@ -0,0 +1,8 @@
+[[!comment format=mdwn
+ username="AdamSpiers"
+ avatar="http://cdn.libravatar.org/avatar/ae41dba34ee6000056f00793c695be75"
+ subject="Would you accept a patch implementing an annex.autocommit gitattribute?"
+ date="2020-01-06T16:12:54Z"
+ content="""
+In a semi-mythical future where I find the time to learn Haskell, in principle would you consider accepting a patch implementing my above suggestion of a new `annex.autocommit` gitattribute which offers finer control of autocommit than the current binary toggle?  It would be very useful to me, and maybe I'm not the only one.
+"""]]