Track backup targets explicitly specified by the user and prevent
excluding them. This for example ensures that `restic backup
--exclude-if-present .git /home/user/data` backs up the `data` folder
even if there is a `.git` folder in `/home/user`.
Note that this does not suffice for commands like `restic backup --exclude data /home/user/data`
as the exclude pattern will still match every single file within `data`.
The TreeNodeIterator decodes nodes while iterating over a tree blob.
This should reduce peak memory usage as now only the serialized tree
blob and a single node have to alive at the same time. Using the
iterator has implications for the error handling however. Now it is
necessary that all loops that iterate through a tree check for errors
before using the node returned by the iterator.
The other change is that it is no longer possible to iterate over a tree
multiple times. Instead it must be loaded a second time. This only
affects the tree rewriting code.
wg.Go() may not be called after wg.Wait(). This prevents connecting two
errgroups such that the errors are propagated between them if the child
errgroup dynamically starts goroutines. Instead use just a single errgroup,
and sequence the shutdown using a sync.WaitGroup. This is far simpler
and does not require any "clever" tricks.
The new method combines both step into a single wrapper function. Thus
it ensures that both are always called in pairs. As an additional
benefit this slightly reduces the boilerplate to upload blobs.
This adds proper support for filenames that include directories. For
example, `/foo/bar` would result in an error when trying to open `/foo`.
The directory tree is now build upfront. This ensures let's the
directory tree construction be handled only once. All accessors then
only have to look up the constructed directory entries.
Only the `Sys()` value from os.FileInfo is kept as field `sys` to
support Windows. The os.FileInfo removal ensures that for values like
`ModTime` that existed in both data structures there's no more confusion
which value is actually used.
The actual implementation still relies on file paths, but with the
abstraction layer in place, an FS implementation can ensure atomic file
accesses in the future.
Depending on parameters the paths in a snapshot do not directly
correspond to real paths on the filesystem. Therefore, reject funcs must
use the FS interface to work correctly.
Files were not included in the backup if the extended metadata for the
file could not be read. This is rather drastic. Instead settle on
returning a warning but still including the file in the backup.
On FreeBSD, limited users may not be able to even list xattrs for the
parent directories above the snapshot source paths. As this can cause
the backup to fail, just ignore those errors.
The deviceID can change e.g. when backing up from filesystem snapshot.
It is only used for hardlink detection. Thus there it is not necessary
to store it for everything else.
Michael Eischer
dmotte
Martin Smith