A second round of 0.23 maint updates

When we have an error renaming the lockfile, we need to make sure
that we remove it upon cleanup. For this, we need to keep track of
whether we opened the file and whether the rename succeeded.

If we did create the lockfile but the rename did not succeed, we
remove the lockfile. This won't protect against all errors, but
the most common ones (target file is open) does get handled.

When we fail to rename, we currently leave the lockfile laying
around. This shows that behaviour.

For such a path '/~/...' the leading '/' is stripped so the server will
get a path starting with '~' and correctly handle it.

Without this change, compiling with gcc and pedantic generates warning:
ISO C does not allow extra ‘;’ outside of a function.

Fixes issue #3338

We don't want the iterator to make us stop whenever we hit an unreadable
dir. We should instead move over to the next item.

An error here will typically mean that the directory was removed between
the time we iterated the parent and the time we wanted to visit it in
which case we should ignore it.

Other kinds of errors such as permissions (or transient errors) also
better dealt with by pretending we didn't see it.

Maintenance updates for v0.23

(fixes issue #3316)

Signed-off-by: Sven Strickroth <email@cs-ware.de>

git_fetch_options was missing from the API docs because it lacked a
documentation comment above the struct declaration.
I used the git_checkout_options docstring as a template.

Also fixes a typo in git_remote_prune_refs (remote, not reamote).

Our path functions expect to work with slashes, so convert a
path with backslashes into one with slashes at the top of
the function.

The header src/cc-compat.h defines portable format specifiers PRIuZ, PRIdZ, and PRIxZ. The original report highlighted the need to use these specifiers in examples/network/fetch.c.  For this commit, I checked all C source and header files not in deps/ and transitioned to the appropriate format specifier where appropriate.

When the stream list init or write fail, we must also make sure to close
the stream, as that's the function contract.

Similarly to how git itself does it, allow the index update operation to
stage a change in a submodule's HEAD.

This also affects `git_index_add_bypath()` by providing a better error
message and a specific error code when a directory is passed.

This is to be returned when the operation which the user asked for is
not possible to do on a directory.

If we get the path from the gitmodules file, look up the submodule we're
interested in by path, rather then by name. Otherwise we might get
duplicate results.

When we rename a submodule, we should be merging two sets of information
based on whether their path is the same. We currently only deduplicate
on equal name, which causes us to double-report.

Upgrade xdiff to version used in core git 2.4.5 (0df0541).

Corrects an issue where an LF is added at EOF while applying
an unrelated change (ba31180), cleans up some unused code (be89977 and
e5b0662), and provides an improved callback to avoid leaking internal
(to xdiff) structures (467d348).

This also adds some additional functionality that we do not yet take
advantage of, namely the ability to ignore changes whose lines are
all blank (36617af).

When invoked with three files that each lack a trailing newline,
the merge result should also lack a trailing newline.

Versions prior to 0.9.8f  did not have this function, rhel/centos5 are still on a
heavily backported version of 0.9.8e and theoretically supported until March 2017

Without this ifdef, I get the following link failure:
```
CMakeFiles/libgit2_clar.dir/src/openssl_stream.c.o: In function `openssl_connect':
openssl_stream.c:(.text+0x45a): undefined reference to `SSL_set_tlsext_host_name'
collect2: error: ld returned 1 exit status
make[6]: *** [libgit2_clar] Error 1
```

filter: add docs for streaming filters

These functions are available on the public API but don't have any
documentation, so they don't appear on the API reference. Fix that.

iterator_walk: cast away constness for free

filter::stream: free the filter sanely

Don't use the filter's free callback to free the actual data structure
holding the filter, as we may not always actually initialize it (the
test may be skipped).

Correctly delimit the keys for submodule lookup

The regex we use to look at the gitmodules file does not correctly
delimit the name of submodule which we want to look up and puts '.*'
straight after the name, maching on any submodule which has the seeked
submodule as a prefix of its name.

Add the missing '\.' in the regex so we want a full stop to exist both
before and after the submodule name.

When two submodules are fairly similar, we may end up loading the wrong
one.