The timestamp index (_darcs/index) should not be confused with the patch index. It was introduced in Petr Ročkai’s 2009 Google Summer of Code project and made its way into Darcs around Darcs 2.3.1 or so (TODO: check history). José Neder added information about the inode numbers of the files in the 2013 Google Summer of Code project.
Darcs tries and save space and make copying faster by hard-linking certain files (this is safe because the files are internal ones that darcs knows will not change). Unfortunately, this hardlinking is a potential source of confusion, because Darcs relies on timestamps to know if it should diff a file for whatsnew or not.
The index is a binary file that overlays a hashed tree over the working copy. This means that every working file and directory has an entry in the index, that contains its path, hash and fileid, and validity data. The validity data is a timestamp plus the file size. The file hashes are sha256’s of the file’s content.
It is used by Darcs to keep track of timestamps rather than trusting the filesystem. This means Darcs doesn’t get confused so easily and start trying to diff files left and right. It is also used to keep track of files renames tracking files by his fileid.
You can query the index with the
darcs show index command
There are two entry types, a file entry and a directory entry. Both have a common binary format.
For each file, the index has a copy of the file’s last modification timestamp taken at the instant when the hash has been computed. This means that when file size and timestamp of a file in working copy matches those in the index, we assume that the hash stored in the index for given file is valid. These hashes are then exposed in the resulting ‘Tree’ object, and can be leveraged by eg. ‘diffTrees’ to compare many files quickly.
You may have noticed that we also keep hashes of directories. These are assumed to be valid whenever the complete subtree has been valid. At any point, as soon as a size or timestamp mismatch is found, the working file in question is opened, its hash (and timestamp and size) is recomputed and updated in-place in the index file (everything lives at a fixed offset and is fixed size, so this isn’t an issue). This is also true of directories: when a file in a directory changes hash, this triggers recomputation of all of its parent directory hashes; moreover this is done efficiently – each directory is updated at most once during an update run.
The Index is organised into "lines" where each line describes a single indexed item. Cf. ‘Item’.
The first word on the index "line" is the length of the file path (which is the only variable-length part of the line). Then comes the path itself, then fixed-length hash (sha256) of the file in question, then three words, one for size, one for “aux”, which is used differently for directories and for files, and one for the fileid(inode or fhandle depending on the operating system) of the file.
With directories, this aux holds the offset of the next sibling line in the index, so we can efficiently skip reading the whole subtree starting at a given directory (by just seeking aux bytes forward). The lines are pre-ordered with respect to directory structure – the directory comes first and after it come all its items. Cf. ‘readIndex’’.
For files, the aux field holds a timestamp.