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

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

That's a bad assumption to make, even though right now it holds
(because of the way we've implemented decompression of packfiles),
this may change in the future, given that ODB objects can be
binary data.

Furthermore, the ODB object can return a NULL pointer if the object
is empty. Copying the NULL pointer to the strbuf lets us handle it
like an empty string. Again, the NULL pointer is valid behavior because
you're supposed to check the *size* of the object before working
on it.

Run shell scripts through shellcheck

Fix for memory leak issue in indexer.c, that surfaces on windows

odb: `git_odb_object` contents are never NULL

This is a contract that we made in the library and that we need to uphold. The
contents of a blob can never be NULL because several parts of the library (including
the filter and attributes code) expect `git_blob_rawcontent` to always return a
valid pointer.

Update message for error during push

Clarify the git_remote_ls() documentation

fetch: clear the connection data on close

travis: disable the cert callback on OSX

The sshd on OSX has stopped responding to the setting which this test
assumes is there, so let's disable it until we can figure out what's
happening.

When we fetch twice with the same remote object, we did not properly
clear the connection flags, so we would leak state from the last
connection.

This can cause the second fetch with the same remote object to fail if
using a HTTP URL where the server redirects to HTTPS, as the second
fetch would see `use_ssl` set and think the initial connection wanted to
downgrade the connection.

On disconnect we simply ask the transport to close the connection, we do
not free it.

Make it clear that this is not the ls-remote command but a way to access
the data we have and how long it's kept around.

Fix missing object in tests/resources/crlf by changing the tail commit

Provide a convenience function `git_remote_push()`

Fixed active_refspecs field not initialized on new git_remote objects

push: use the common refspec parser