Make sure error messages are set for most error returns

By default do not set tagopt when cloning

allow checkout to proceed when a dir to be removed is in use (win32)

branch.h: fix typo in docs

Strict/lax commit API

Peeling to an invalid type is now checked via an assert so this
test is no longer relevant.

fix some leaks

braces

puns are not funny; type punning especially so

Switch to index_version as "git_pack_file is ready" flag

We use p->index_map.data to check whether the struct has been set up
and all the information about the index is stored there. This variable
gets set up halfway through the setup process, however, and a thread
can come along and use fields that haven't been written to yet.

Crucially, pack_entry_find_offset() needs to read the index version
(which is written after index_map) to know the offset and stride
length to pass to sha1_entry_pos(). If these values are wrong,
assertions in it will fail, as it will be reading bogus data.

Make index_version the last field to be written and switch from using
p->index_map.data to p->index_version as "git_pack_file is ready" flag
as we can use it to know if every field has been written.

refdb_fs: do not require peeled packed refs to be tags

We only set our negative flag for PEELING_FULL; we can fall
back to the lesser PEELING_STANDARD if our ref is in the
refs/tags/ hierarchy.

This makes parsing easier! :p

Fixes #1532

This reverts commit 8c535f3f.

Older versions of git would only write peeled entries for
items under refs/tags/. Newer versions will write them for
all refs, and we should be prepared to handle that.

More packfile locking

There is an occasional assertion failure in sha1_entry_pos from
pack_entry_find_index when running threaded.  Holding the mutex
around the code that grabs the index_map data and processes it
makes this assertion failure go away.

We need to hammer the packfile open phase harder in the thread
tests, in addition to the cache API.

We were still seeing a few issues in threaded access to packs.
This adds extra locks around the opening of the mwindow to
avoid a different race.

There are many paths through revparse that may return an error
code without reporting an error, I believe.  This fixes one of
them.  Because of the backtracking in revparse, it is pretty
complicated to fix the others.