Fix broken tests on Windows

Fix negative ignores withing ignored dirs

doc: add documentation to all the public structs and enums

Adjust the local transport for the common refspec parser

Notes API consistency fixes

The iterator is the last argument. There is also no returned notes, just
ids, so the comment about freeing is out of place.

Make it consistent between git_note_create() and git_note_remote() by
putting it after the repository.

This makes them show up in the reference, even if the text itself isn't
the most descriptive.

These have been found with

    grep -Przon '\n\ntypedef struct.*?\{' -- include
    grep -Przon '\n\ntypedef enum.*?\{' -- include

A rule can only negate something which was explicitly mentioned in the
rules before it. Change our parsing to ignore a negative rule which does
not negate something mentioned in the rules above it.

While here, fix a wrong allocator usage. The memory for the match string
comes from pool allocator. We must not free it with the general
allocator. We can instead simply forget the string and it will be
cleaned up.

Given

    top
    !top/foo

in an ignore file, we should not unignore top/foo. This is an
implementation detail of the git code leaking, but that's the behaviour
we should show.

A negation rule can only negate an exact rule it has seen before.

This is not our code and it adds unecessary changes from the upstream
code.

Spelling fixes

init: return the number of initializations

Add missing else directive

Add missing else directive to fix compiler warning: control reaches
end of non-void function

win32: clear connection data on close

fix cl_git_mkfile() to test for -1 rather than 0 for errors from p_creat...

describe: check error codes

Regain non-threadsafe builds and make the valgrind build failable

peel: reject bad queries with EPEEL

We no longer have NULL strings, but empty ones and duplicate the sides
if necessar, so the first check will never do anything.

While in the area, remove unnecessary ifs and early returns.

Valgrind is now clean except for libssl and libgcrypt.

There are some combination of objects and target types which we know
cannot be fulfilled. Return EINVALIDSPEC for those to signify that there
is a mismatch in the user-provided data and what the object model is
capable of satisfying.

If we start at a tag and in the course of peeling find out that we
cannot reach a particular type, we return EPEEL.

When flipping the default to threadsafe, we never made sure that the
tests were building non-threadsafe variants. Specify which thead-safety
we want for each.

We don't need to know the result of the travis run before determining
whether we want to consider the merge as successful.

Fix for misleading "missing delta bases" error.

Do not assume blob contents are NULL terminated