Ok, this is the real deal. Hopefully. Here's how it's going to work:
- One main method, called `git__throw`, that sets the error
code and error message when an error happens.
This method must be called in every single place where an error
code was being returned previously, setting an error message
instead.
Example, instead of:
return GIT_EOBJCORRUPTED;
Use:
return git__throw(GIT_EOBJCORRUPTED,
"The object is missing a finalizing line feed");
And instead of:
[...] {
error = GIT_EOBJCORRUPTED;
goto cleanup;
}
Use:
[...] {
error = git__throw(GIT_EOBJCORRUPTED, "What an error!");
goto cleanup;
}
The **only** exception to this are the allocation methods, which
return NULL on failure but already set the message manually.
/* only place where an error code can be returned directly,
because the error message has already been set by the wrapper */
if (foo == NULL)
return GIT_ENOMEM;
- One secondary method, called `git__rethrow`, which can be used to
fine-grain an error message and build an error stack.
Example, instead of:
if ((error = foobar(baz)) < GIT_SUCCESS)
return error;
You can now do:
if ((error = foobar(baz)) < GIT_SUCCESS)
return git__rethrow(error, "Failed to do a major operation");
The return of the `git_lasterror` method will be a string in the
shape of:
"Failed to do a major operation. (Failed to do an internal
operation)"
E.g.
"Failed to open the index. (Not enough permissions to access
'/path/to/index')."
NOTE: do not abuse this method. Try to write all `git__throw`
messages in a descriptive manner, to avoid having to rethrow them to
clarify their meaning.
This method should only be used in the places where the original
error message set by a subroutine is not specific enough.
It is encouraged to continue using this style as much possible to
enforce error propagation:
if ((error = foobar(baz)) < GIT_SUCCESS)
return error; /* `foobar` has set an error message, and
we are just propagating it */
The error handling revamp will take place in two phases:
- Phase 1: Replace all pieces of code that return direct error codes
with calls to `git__throw`. This can be done semi-automatically
using `ack` to locate all the error codes that must be replaced.
- Phase 2: Add some `git__rethrow` calls in those cases where the
original error messages are not specific enough.
Phase 1 is the main goal. A minor libgit2 release will be shipped once
Phase 1 is ready, and the work will start on gradually improving the
error handling mechanism by refining specific error messages.
OTHER NOTES:
- When writing error messages, please refrain from using weasel
words. They add verbosity to the message without giving any real
information. (<3 Emeric)
E.g.
"The reference file appears to be missing a carriage return"
Nope.
"The reference file is missing a carriage return"
Yes.
- When calling `git__throw`, please try to use more generic error
codes so we can eventually reduce the list of error codes to
something more reasonable. Feel free to add new, more generic error
codes if these are going to replace several of the old ones.
E.g.
return GIT_EREFCORRUPTED;
Can be turned into:
return git__throw(GIT_EOBJCORRUPTED,
"The reference is corrupted");
Typical use is git_reference_resolve(&ref, ref). Currently, if there is
an error, ref will point to NULL, causing the user to lose that
reference.
Always update resolved_ref instead of just on finding an OID ref,
storing the last valid reference in it.
This change helps simplify the code for allowing root commits.
Signed-off-by: Carlos Martín Nieto <cmn@elego.de>
Allow any well-formed reference name to live under refs/ removing the
condition that they be under refs/{heads,tags,remotes}/ as was the
design of git.
An exception is made for HEAD which is allowed to contain an OID
reference in detached HEAD state.
Signed-off-by: Carlos Martín Nieto <cmn@elego.de>
Add internal reference create and rename functions which take a force
parameter, telling them to overwrite an existing reference if it
exists.
These functions try to update the reference if it's of the same type
as the one it's going to be replaced by. Otherwise the old reference
becomes invalid.
Signed-off-by: Carlos Martín Nieto <cmn@elego.de>
List all the references in the repository, calling a custom
callback for each one.
The listed references may be filtered by type, or using
a bitwise OR of several types. Use the magic value
`GIT_REF_LISTALL` to obtain all references, including
packed ones.
The `callback` function will be called for each of the references
in the repository, and will receive the name of the reference and
the `payload` value passed to this method.
It's no longer retarded. All object interdependencies are stored as OIDs
instead of actual objects. This should be hundreds of times faster,
specially on big repositories. Heck, who knows, maye it doesn't even
segfault -- wouldn't that be awesome?
What has changed on the API?
`git_commit_parent`, `git_commit_tree`, `git_tag_target` now return
their values through a pointer-to-pointer, and have an error code.
`git_commit_set_tree` and `git_tag_set_target` now return an error
code and may fail.
`git_repository_free__no_gc` has been deprecated because it's
stupid. Since there are no longer any interdependencies between
objects, we don't need internal reference counting, and GC
never fails or double-free's pointers.
`git_object_close` now does a very sane thing: marks an object
as unused. Closed objects will be eventually free'd from the
object cache based on LRU. Please use `git_object_close` from
the garbage collector `destroy` method on your bindings. It's
100% safe.
`git_repository_gc` is a new method that forces a garbage collector
pass through the repo, to free as many LRU objects as possible.
This is useful if we are running out of memory.
We now store only one sorting callback that does entry comparison. This
is used when sorting the entries using a quicksort, and when looking for
a specific entry with the new search methods.
The following search methods now exist:
git_vector_search(vector, entry)
git_vector_search2(vector, custom_search_callback, key)
git_vector_bsearch(vector, entry)
git_vector_bsearch2(vector, custom_search_callback, key)
The sorting state of the vector is now stored internally.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
The methods previously known as
git_repository_lookup
git_repository_newobject
git_repository_lookup_ref
are now part of their respective namespaces:
git_object_lookup
git_object_new
git_reference_lookup
This makes the API more consistent with the new references API.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
The following methods have been implemented:
git_reference_packall
git_reference_rename
git_reference_delete
The library now has full support for packed references, including
partial and total writing. Internal documentation has been updated with
the details.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
These two reference types are now stored separately to eventually allow
the removal/renaming of loose references and rewriting of the refs
packfile.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
The old hash table with chained buckets has been replaced by a new one
using Cuckoo hashing, which offers guaranteed constant lookup times.
This should improve speeds on most use cases, since hash tables in
libgit2 are usually used as caches where the objects are stored once and
queried several times.
The Cuckoo hash implementation is based off the one in the Basekit
library [1] for the IO language, but rewritten to support an arbritrary
number of hashes. We currently use 3 to maximize the usage of the nodes pool.
[1]: https://github.com/stevedekorte/basekit/blob/master/source/CHash.c
Signed-off-by: Vicent Marti <tanoku@gmail.com>
The new `git_filebuf` structure provides atomic high-performance writes
to disk by using a write cache, and optionally a double-buffered scheme
through a worker thread (not enabled yet).
Writes can be done 3-layered, like in git.git (user code -> write cache
-> disk), or 2-layered, by writing directly on the cache. This makes
index writing considerably faster.
The `git_filebuf` structure contains all the old functionality of
`git_filelock` for atomic file writes and reads. The `git_filelock`
structure has been removed.
Additionally, the `git_filebuf` API allows to automatically hash (SHA1)
all the data as it is written to disk (hashing is done smartly on big
chunks to improve performance).
Signed-off-by: Vicent Marti <tanoku@gmail.com>
The `git__joinpath` function has been changed to use a statically
allocated buffer; we assume the buffer to be 4096 bytes, because fuck
you.
The new method also supports an arbritrary number of paths to join,
which may come in handy in the future.
Some methods which were manually joining paths with `strcpy` now use the
new function, namely those in `index.c` and `refs.c`.
Based on Emeric Fermas' original patch, which was using the old
`git__joinpath` because I'm stupid. Thanks!
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Several changes have been committed to allow the user to create
in-memory references and write back to disk. Peeling of symbolic
references has been made explicit. Added getter and setter methods for
all attributes on a reference. Added corresponding documentation.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
