When I was writing threading tests for the new cache, the main
error I kept running into was a pack file having it's content
unmapped underneath the running thread. This adds a lock around
the routines that map and unmap the pack data so that threads can
effectively reload the data when they need it.
This also required reworking the error handling paths in a couple
places in the code which I tried to make consistent.
These offsets are needed for REF_DELTA objects, which encode which
object they use as a base, but not where it lies in the packfile, so
we need a list.
These objects are mostly from older packfiles, before OFS_DELTA was
widely spread. The time spent in indexing these packfiles is greatly
reduced, though remains above what git is able to do.
Somewhat surprisingly, this can increase the speed considerably, as we
don't bother trying to decide what to evict, and the most used entries
are quickly back into the cache.
The indexer needs to call the packfile's free function so it takes care of
freeing the caches.
We still need to close the mwf descriptor manually so we can rename the
packfile into its final name on Windows.
Many delta bases are re-used. Cache them to avoid inflating the same
data repeatedly.
This version doesn't limit the amount of entries to store, so it can
end up using a considerable amound of memory.
To paraphrase @peff:
You can get both size and type from a packed object reasonably cheaply.
If you have:
* An object that is not a delta; both type and size are available in the
packfile header.
* An object that is a delta. The packfile type will be OBJ_*_DELTA, and
you have to resolve back to the base to find the real type. That means
potentially a lot of packfile index lookups, but each one is
relatively cheap. For the size, you inflate the first few bytes of the
delta, whose header will tell you the resulting size of applying the
delta to the base.
For simplicity, we just decompress the whole delta for now.
This updates all the `foreach()` type functions across the library
that take callbacks from the user to have a consistent behavior.
The rules are:
* A callback terminates the loop by returning any non-zero value
* Once the callback returns non-zero, it will not be called again
(i.e. the loop stops all iteration regardless of state)
* If the callback returns non-zero, the parent fn returns GIT_EUSER
* Although the parent returns GIT_EUSER, no error will be set in
the library and `giterr_last()` will return NULL if called.
This commit makes those changes across the library and adds tests
for most of the iteration APIs to make sure that they follow the
above rules.
Once a file is registered, there is no way to deregister it, even
after the structure that contains it is no longer needed and has been
freed. This may be the source of #624.
Allow and use the deregister function to remove our file from the
global list.
On RAM: the .idx and .pack files become links to a .lock and the original download respectively.
Assume some feature (such as record locking) supported by SFS but not JXFS or RAM: is required.
There are three changes here:
- correctly propogate error code from failed object lookups
- make zlib inflate use our allocators
- add OID to notfound error in ODB lookups
This migrates odb.c, odb_loose.c, odb_pack.c and pack.c to
the new style of error handling. Also got the unix and win32
versions of map.c. There are some minor changes to other
files but no others were completely converted.
This also contains an update to filebuf so that a zeroed out
filebuf will not think that the fd (== 0) is actually open
(and inadvertently call close() on fd 0 if cleaned up).
Lastly, this was built and tested on win32 and contains a
bunch of fixes for the win32 build which was pretty broken.
This is legacy compat stuff for when `deflateBound` is not defined, but
we're not embedding zlib and that function is always available. Kill
that with fire.
This takes all of the functions that look up simple data about
paths (such as `git_futils_isdir`) and moves them over to path.h
(becoming `git_path_isdir`). This leaves fileops.h just with
functions that actually manipulate the filesystem or look at
the file contents in some way.
As part of this, the dir.h header which is really just for win32
support was moved into win32 (with some minor changes).
off_t is always 32 bits in Windows, which is beyond stupid, but we just
don't care anymore because we're using `git_off_t` which is assured to
be 64 bits on all platforms, regardless of compilation mode. Just
ensure that no casts to `off_t` are performed.
Also, the check for `off_t` overflows has been dropped, once again,
because the size of our offsets is always 64 bits on all platforms.
Fixes#534
See `global.c` for a description of what we're doing.
When libgit2 is built with GIT_THREADS support, the threading system
must be explicitly initialized with `git_threads_init()`.
This makes libgit2 more closely match Git, which only checks for
ambiguous pack entries when given short hashes.
Note that the only time this is ever relevant is when a pack has the
same object more than once (it's happened in the wild, I promise).
There were quite a few places were spaces were being used instead of
tabs. Try to catch them all. This should hopefully not break anything.
Except for `git blame`. Oh well.
1. The license header is technically not valid if it doesn't have a
copyright signature.
2. The COPYING file has been updated with the different licenses used in
the project.
3. The full GPLv2 header in each file annoys me.
When indexing a file with ref deltas, a temporary cache for the
offsets has to be built, as we don't have an index file yet. If the
user takes the responsiblity for filling the cache, the packing code
will look there first when it finds a ref delta.
Signed-off-by: Carlos Martín Nieto <carlos@cmartin.tk>
The callers immediately throw away the offset, so we don't need any
logical changes in any of them. This will be useful for the indexer,
as it does need to know where the compressed data ends.
Signed-off-by: Carlos Martín Nieto <carlos@cmartin.tk>
