An earlier change to `git_diff_from_iterators` introduced a
memory leak where the allocated spoolandsort iterator was not
returned to the caller and thus not freed.

One proposal changes all iterator APIs to use git_iterator** so
we can reallocate the iterator at will, but that seems unexpected.
This commit makes it so that an iterator can be changed in place.
The callbacks are isolated in a separate structure and a pointer
to that structure can be reassigned by the spoolandsort extension.

This means that spoolandsort doesn't create a new iterator; it
just allocates a new block of callbacks (along with space for its
own extra data) and swaps that into the iterator.

Additionally, since spoolandsort is only needed to switch the
case sensitivity of an iterator, this simplifies the API to only
take the ignore_case boolean and to be a no-op if the iterator
already matches the requested case sensitivity.

Teach UNC paths to git_path_dirname_r()

Fix libgit2/libgit2sharp#256

odb: check if object exists before writing

Update the procondition of git_odb_backend::write.
It may now be assumed that the object has already been hashed.

Mark travis build as failed if any test fails

fetchhead reading/iterating

netops: on SSL teardown only send shutdown alert

Make `goto cleanup` more consistent

RFC: Some cleanups/features in CMakeLists.txt

Fix some leaks in tests

Also adds some asserts.

... so we can clean up correctly.

Instead tell MSVC to group the source files by directory.

- Also document the -DSTDCALL even better.

With the src/compat/*.c glob removed, there is no longer a difference to
the default globs we use for the IF( UNIX ) case.

This directory doesn't exist.

Up to now, on windows we don't even bother to look if the user has a zlib
available somwhere.

In almost all larger commercial projects i've participated in, it was not
at all uncommon to have such a dependency somewhere in the source tree and
use it whereever required.
Searching for it, even if it's unlikely to be present, allows for such a
scenario (i.e. by prefilling the CMake-Cache).

There may be some question about whether this is likely to be needed
at all, but that's above my head at the moment.

don't deref before we've asserted

just sayin'.

Trivial comment fix

I was totally flaunting @ben's 3-space tab advice.

This looked wrong to me.  I *think* this is more appropriate
commentary.

Reset global variable to NULL after free'ing resource

Fix diff constructor names

Based on feedback from the ObjectiveGit folks, these are some
further updates to diff.h areas that are poorly documented.

The diff constructor functions had some confusing names, where the
"old" side of the diff was coming after the "new" side.  This
reverses the order in the function name to make it less confusing.

Specifically...

* git_diff_index_to_tree becomes git_diff_tree_to_index
* git_diff_workdir_to_index becomes git_diff_index_to_workdir
* git_diff_workdir_to_tree becomes git_diff_tree_to_workdir