When ffi.new() is used to build a new pointer object, the returned
pointer object has ownership on the allocated memory. When it is
garbage-collected, then the memory is freed. Thus, we need to make sure
the original object survives its use, otherwise the casted pointer will
point to garbage.
This fixes one test which was failing with the latest CFFI version, see
issue #694. Thus, this commit also reverts 803b1cb (cffi 1.10 not yet
supported, 2017-03-22) where the latest CFFI version was marked as
unsupported.
Signed-off-by: Lukas Fleischer <lfleischer@lfos.de>
Removed a chunk of code from `checkout` that did nothing, but had a bug.
When checking out from a branch-less state (like the state when a repository is first initialized) the code failed.
The failure was due to trying to get some properties of the current branch, which were never used in the code.
As happened in support request https://github.com/libgit2/libgit2/issues/3963 it can be easily overseen,
that push returns True, when the remote has installed a hook that denies the commits.
Libgit2 partially forwards OS error message texts.
On non-english Windows OSes these errors may contain non-ascii characters (i.e. umlauts).
To avoid that a UnicodeDecodeError is raised the error message is interpreted as UTF-8.
The solution should not be necessary on linux/osx as they return always ascii (as far as I know).
Thus this solution will not change the behaviour on linux/osx.
In Python versions older than 3.2 hasattr is not robust as it masks real
errors. And usually is not efficient as a call to hasattr is often
followed by another call to getattr. It is best to avoid using it
completely.
See https://docs.python.org/3/whatsnew/3.2.html#other-language-changes
This was implemented for clone, but not for fetch or push, so it was
deleted during the conversion, which shows why we need to unify these
callback structures.
This represents what's going on much better than the remnants from the
older methods. What we do is pass a list of callbacks to libgit2 for it
to call, and they are valid for a single operation, not for the remote
itself.
This should also make it easier to re-use callbacks which have already
been set up.
