Signed-off-by: Justin Lebar <jlebar@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

* jh/fetch-head-update:
  write first for-merge ref to FETCH_HEAD first

The FETCH_HEAD refname is supposed to refer to the ref that was fetched
and should be merged. However all fetched refs are written to
.git/FETCH_HEAD in an arbitrary order, and resolve_ref_unsafe simply
takes the first ref as the FETCH_HEAD, which is often the wrong one,
when other branches were also fetched.

The solution is to write the for-merge ref(s) to FETCH_HEAD first.
Then, unless --append is used, the FETCH_HEAD refname behaves as intended.
If the user uses --append, they presumably are doing so in order to
preserve the old FETCH_HEAD.

While we are at it, update an old example in the read-tree documentation
that implied that each entry in FETCH_HEAD only has the object name, which
is not true for quite a while.

[jc: adjusted tests]

Signed-off-by: Joey Hess <joey@kitenet.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

Since 34110cd4 (Make 'unpack_trees()' have a separate source and
destination index) it is no longer true that a subdirectory with
the same prefix must not exist.

Signed-off-by: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

* nd/sparse-doc:
  git-read-tree.txt: update sparse checkout examples

* mg/maint-doc-sparse-checkout:
  git-read-tree.txt: correct sparse-checkout and skip-worktree description
  git-read-tree.txt: language and typography fixes
  unpack-trees: print "Aborting" to stderr

* nd/sparse-doc:
  git-read-tree.txt: update sparse checkout examples

* mg/maint-doc-sparse-checkout:
  git-read-tree.txt: correct sparse-checkout and skip-worktree description
  git-read-tree.txt: language and typography fixes
  unpack-trees: print "Aborting" to stderr

The negation example uses '*' to match everything. This used to work
before 9037026 (unpack-trees: fix sparse checkout's "unable to match
directories") because back then, the list of paths is used to match
sparse patterns, so with the patterns

    *
    !subdir/

subdir/ always matches any path that start with subdir/ and "*" has no
chance to get tested. The result is subdir is excluded.

After the said commit, a tree structure is dynamically created and
sparse pattern matching now follows closely how read_directory()
applies .gitignore. This solves one problem, but reveals another one.

With this new strategy, "!subdir/" rule will be only tested once when
"subdir" directory is examined. Entries inside subdir, when examined,
will match "*" and are (correctly) re-added again because any rules
without a slash will match at every directory level. In the end, "*"
can revert every negation rules.

In order to correctly exclude subdir, we must use

    /*
    !subdir

to limit "match all" rule at top level only.

"*" rule has no actual use in sparse checkout and can be confusing to
users. While we can automatically turn "*" to "/*", this violates
.gitignore definition. Instead, discourage "*" in favor of "/*" (in
the second example).

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

The description of .git/info/sparse-checkout and
skip-worktree is exactly the opposite of what is true, which is:

If a file matches a pattern in sparse-checkout, then (it is to be
checked out and therefore) skip-worktree is unset for that file;
otherwise, it is set (so that it is not checked out).

Currently, the opposite is documented, and (consistently) read-tree's
behavior with respect to bit flips is descibed incorrectly.

Fix it.

In hindsight, it would have been much better to have a "sparse-ignore"
or "sparse-skip" file so that an empty file would mean a full checkout,
and the file logic would be analogous to that of .gitignore, excludes
and skip-worktree.

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>