Refactor git_cache to use an rwlock

This significantly reduces contention when many threads are trying to
read from the cache simultaneously.

Fix bugs with negative ignores inside an ignored parent directory

add Julia to the language bindings list

pack: return the correct final offset

The callers of git_packfile_unpack() expect the obj_offset argument to
be set to the beginning of the next object. We were mistakenly returning
the the offset of the object's data, which causes the CRC function to
try to use the wrong offset.

Set obj_offset to curpos instead of elem->offset to point to the next
element and bring back expected behaviour.

merge base: Correctly raise an error if a non-commit object is passed.

Allow mkdir helper to skip parent errors

Check if the refspec matches before transforming

Our mkdir helper was failing is a parent directory was not
accessible even if the child directory could be created.
This changes the helper to keep trying child directories
even when the parent is unwritable.

Export `git_oid_tostr_s` instead of `_allocfmt`

This reverts commit 0dc54e14.

We don't really use this at all, and it breaks packaging in Windows.

The old `allocfmt` is of no use to callers, as they are not able to free
the returned buffer. Export a new API that returns a static string that
doesn't need to be freed.

http: make sure we can consume the data we request

Removing a completed starter project.

The recv buffer (parse_buffer) and the buffer have independent sizes and
offsets. We try to fill in parse_buffer as much as possible before
passing it to the http parser. This is fine most of the time, but fails
us when the buffer is almost full.

In those situations, parse_buffer can have more data than we would be
able to put into the buffer (which may be getting full if we're towards
the end of a data sideband packet).

To work around this, we check if the space we have left on our buffer is
smaller than what could come from the network. If this happens, we make
parse_buffer think that it has as much space left as our buffer, so it
won't try to retrieve more data than we can deal with.

As the start of the data may no longer be at the start of the buffer, we
need to keep track of where it really starts (data_offset) and use that
in our calculations for the real size of the data we received from the
network.

This fixes #2518.

Removing the starter project for adding support for the symref
extension (#2006) from PROJECTS.md, as this seems to have been
completed with the merge of #2376.

git_remote_ls() crashes before connect is called

SPNEGO authentication via GSSAPI

Decode base64-encoded text into a git_buf

The online::push::notes test pushes a note but leaves it hanging
around for other tests to stumble across when they're validating
that they're seeing the refs they expect to see.  Clean it up on
exit.

Custom transport: minor cleanups

 * Move the transport registration mechanisms into a new header under
   'sys/' because this is advanced stuff.
 * Remove the 'priority' argument from the registration as it adds
   unnecessary complexity.  (Since transports cannot decline to operate,
   only the highest priority transport is ever executed.)  Users who
   require per-priority transports can implement that in their custom
   transport themselves.
 * Simplify registration further by taking a scheme (eg "http") instead
   of a prefix (eg "http://").

Don't include the unreadable tests on win32

Allow NULL error message prefix when class=GITERR_OS