Commit b1a6c316a6 moved auto-refresh into
the pack backend, and added a comment accordingly. Commit
43820f204e moved auto-refresh back *out*
of backends into the ODB layer, but didn't update the comment.
This change introduces git_odb_write_multi_pack_index(), which creates a
`multi-pack-index` file from all the `.pack` files that have been loaded
in the ODB.
Fixes: #5399
Custom backends can allocate memory when reading objects and providing
them to libgit2. However, if an error occurs in the custom backend
after the memory has been allocated for the custom object but before
it's returned to libgit2, the custom backend has no way to free that
memory and it must be leaked.
Provide a free function that corresponds to the alloc function so that
custom backends have an opportunity to free memory before they return an
error.
The `git_odb_backend_malloc` name is a system function that is provided
for custom ODB backends and allows them to allocate memory for an ODB
object in the read callback. This is important so that libgit2 can
later free the memory used by an ODB object that was read from the
custom backend.
However, the name _suggests_ that it actually allocates a
`git_odb_backend`. It does not; rename it to make it clear that it
actually allocates backend _data_.
Since we now always build the library with cdecl calling conventions,
our callbacks should be decorated as such so that users will not be able
to provide callbacks defined with other calling conventions.
The `GIT_CALLBACK` macro will inject the `__cdecl` attribute as
appropriate.
git_object_t is the future; update the public API to use it. This will
also ensure that we can build our tests which make use of the old API
without modification (and without compiler warnings).
As refdb and odb backends can be allocated by client code, libgit2
can’t know whether an alternative memory allocator was used, and thus
should not try to call `git__free` on those objects.
Instead, odb and refdb backend implementations must always provide
their own `free` functions to ensure memory gets freed correctly.
Restricting files to size_t is a silly limitation. The loose backend
writes to a file directly, so there is no issue in using 63 bits for the
size.
We still assume that the header is going to fit in 64 bytes, which does
mean quite a bit smaller files due to the run-length encoding, but it's
still a much larger size than you would want Git to handle.
When given an ODB from which to read objects, the indexer will attempt
to inject the missing bases at the end of the pack and update the
header and trailer to reflect the new contents.
Previously, `git_object_read()`, `git_object_read_prefix()` and
`git_object_exists()` were implementing an auto refresh logic. When the
expected object couldn't be found in any backend, a call to
`git_odb_refresh()` was triggered and the lookup was once again performed
against all backends.
This commit removes this auto-refresh logic from the odb layer and pushes
it down into the pack-backend (as it's the only one currently exposing
a `refresh()` endpoint).
Clarify the role of each function and in particular mention that there
is no need for the backend or stream to worry about the object's id,
as it will be given when `finalize_write` is called.
The frontend is in charge of calculating the id of the objects. Thus
the backends should treat it as a read-only value. The positioning in
the function signature made it seem as though it was an output
parameter.
Make the id const and move it from the front to behind the subject
(backend or stream).
This moves some of the odb_backend stuff that is related to the
internals of an odb_backend implementation into include/git2/sys.
Some of the stuff related to streaming I left in include/git2
because it seemed like it would be reasonably needed by a normal
user who wanted to stream objects into and out of the ODB.
Also, I added APIs for traversing the list of backends so that
some of the tests would not need to access ODB internals.
