Are Git's pack files deltas rather than snapshots?

One of the key differences between Git and most other version control systems is that the others tend to store commits as a series of deltas - changesets between one commit and the next. This seems logical, since it's the smallest possible amount of information to store about a commit. But the longer the commit history gets, the more calculation it takes to compare ranges of revisions.

By contrast, Git stores a complete snapshot of the whole project in each revision. The reason this doesn't make the repo size grow dramatically with each commit is each file in the project is stored as a file in the Git subdirectory, named for the hash of its contents. So if the contents haven't changed, the hash hasn't changed, and the commit just points to the same file. And there are other optimizations as well.

All this made sense to me until I stumbled on this information about pack files, into which Git puts data periodically to save space:

In order to save that space, Git utilizes the packfile. This is a format where Git will only save the part that has changed in the second file, with a pointer to the file it is similar to.

Isn't this basically going back to storing deltas? If not, how is it different? How does this avoid subjecting Git to the same problems other version controls systems have?

For example, Subversion uses deltas, and rolling back 50 versions means undoing 50 diffs, whereas with Git you can just grab the appropriate snapshot. Unless git also stores 50 diffs in the packfiles... is there some mechanism that says "after some small number of deltas, we'll store a whole new snapshot" so that we don't pile up too large a changeset? How else might Git avoid the disadvantages of deltas?

Solution 1:

Summary:
Git’s pack files are carefully constructed to effectively use disk caches and provide “nice” access patterns for common commands and for reading recently referenced objects.

Git’s pack file format is quite flexible (see Documentation/technical/pack-format.txt, or The Packfile in The Git Community Book). The pack files store objects in two main ways: “undeltified” (take the raw object data and deflate-compress it), or “deltified” (form a delta against some other object then deflate-compress the resulting delta data). The objects stored in a pack can be in any order (they do not (necessarily) have to be sorted by object type, object name, or any other attribute) and deltified objects can be made against any other suitable object of the same type.

Git’s pack-objects command uses several heuristics to provide excellent locality of reference for common commands. These heuristics control both the selection of base objects for deltified objects and the order of the objects. Each mechanism is mostly independent, but they share some goals.

Git does form long chains of delta compressed objects, but the heuristics try to make sure that only “old” objects are at the ends of the long chains. The delta base cache (whose size is controlled by the core.deltaBaseCacheLimit configuration variable) is automatically used and can greatly reduce the number of “rebuilds” required for commands that need to read a large number of objects (e.g. git log -p).

Delta Compression Heuristic

A typical Git repository stores a very large number of objects, so it can not reasonably compare them all to find the pairs (and chains) that will yield the smallest delta representations.

The delta base selection heuristic is based on the idea that the good delta bases will be found among objects with similar filenames and sizes. Each type of object is processed separately (i.e. an object of one type will never be used as the delta base for an object of another type).

For the purposes of delta base selection, the objects are sorted (primarily) by filename and then size. A window into this sorted list is used to limit the number of objects that are considered as potential delta bases. If a “good enough”¹ delta representation is not found for an object among the objects in its window, then the object will not be delta compressed.

The size of the window is controlled by the --window= option of git pack-objects, or the pack.window configuration variable. The maximum depth of a delta chain is controlled by the --depth= option of git pack-objects, or the pack.depth configuration variable. The --aggressive option of git gc greatly enlarges both the window size and the maximum depth to attempt to create a smaller pack file.

The filename sort clumps together the objects for entries with with identical names (or at least similar endings (e.g. .c)). The size sort is from largest to smallest so that deltas that remove data are preferred to deltas that add data (since removal deltas have shorter representations) and so that the earlier, larger objects (usually newer) tend to be represented with plain compression.

¹ What qualifies as “good enough” depends on the size of the object in question and its potential delta base as well as how deep its resulting delta chain would be.

Object Ordering Heuristic

Objects are stored in the pack files in a “most recently referenced” order. The objects needed to reconstruct the most recent history are placed earlier in the pack and they will be close together. This usually works well for OS disk caches.

All the commit objects are sorted by commit date (most recent first) and stored together. This placement and ordering optimizes the disk accesses needed to walk the history graph and extract basic commit information (e.g. git log).

The tree and blob objects are stored starting with the tree from the first stored (most recent) commit. Each tree is processed in a depth first fashion, storing any objects that have not already been stored. This puts all the trees and blobs required to reconstruct the most recent commit together in one place. Any trees and blobs that have not yet been saved but that are required for later commits are stored next, in the sorted commit order.

The final object ordering is slightly affected by the delta base selection in that if an object is selected for delta representation and its base object has not been stored yet, then its base object is stored immediately before the deltified object itself. This prevents likely disk cache misses due to the non-linear access required to read a base object that would have “naturally” been stored later in the pack file.

Solution 2:

The use of delta storage in the pack file is just an implementation detail. At that level, Git doesn't know why or how something changed from one revision to the next, rather it just knows that blob B is pretty similar to blob A except for these changes C. So it will only store blob A and changes C (if it chooses to do so - it could also choose to store blob A and blob B).

When retrieving objects from the pack file, the delta storage is not exposed to the caller. The caller still sees complete blobs. So, Git works the same way it always has without the delta storage optimisation.