amorfo77/restic
1
0
Fork 0
mirror of https://github.com/restic/restic.git synced 2026-09-29 07:20:30 +00:00
Code Issues Projects Releases Wiki Activity

diff: switch to efficient DualTreeIterator

Browse Source 
The previous implementation stored the whole tree in a map and used it
for checking overlap between trees. This is now replaced with the
DualTreeIterator, which iterates over two trees in parallel and returns
the merge stream in order. In case of overlap between both trees, it
returns both nodes at the same time. Otherwise, only a single node is
returned.
This commit is contained in:
Michael Eischer
2026-01-31 20:03:38 +01:00
parent 350f29d921
commit 24d56fe2a6
2 changed files with 118 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files
cmd/restic/cmd_diff.go
+16 -43
View File
@@ -5,7 +5,6 @@ import (
"encoding/json"
"path"
"reflect"
"sort"
"github.com/restic/restic/internal/data"
"github.com/restic/restic/internal/debug"
@@ -240,37 +239,6 @@ func (c *Comparer) collectDir(ctx context.Context, blobs restic.AssociatedBlobSe
return ctx.Err()
}
func uniqueNodeNames(tree1, tree2 data.TreeNodeIterator) (tree1Nodes, tree2Nodes map[string]*data.Node, uniqueNames []string, err error) {
names := make(map[string]struct{})
tree1Nodes = make(map[string]*data.Node)
for item := range tree1 {
if item.Error != nil {
return nil, nil, nil, item.Error
}
node := item.Node
tree1Nodes[node.Name] = node
names[node.Name] = struct{}{}
}
tree2Nodes = make(map[string]*data.Node)
for item := range tree2 {
if item.Error != nil {
return nil, nil, nil, item.Error
}
node := item.Node
tree2Nodes[node.Name] = node
names[node.Name] = struct{}{}
}
uniqueNames = make([]string, 0, len(names))
for name := range names {
uniqueNames = append(uniqueNames, name)
}
sort.Strings(uniqueNames)
return tree1Nodes, tree2Nodes, uniqueNames, nil
}
func (c *Comparer) diffTree(ctx context.Context, stats *DiffStatsContainer, prefix string, id1, id2 restic.ID) error {
debug.Log("diffing %v to %v", id1, id2)
tree1, err := data.LoadTree(ctx, c.repo, id1)
@@ -283,24 +251,29 @@ func (c *Comparer) diffTree(ctx context.Context, stats *DiffStatsContainer, pref
return err
}
tree1Nodes, tree2Nodes, names, err := uniqueNodeNames(tree1, tree2)
if err != nil {
return err
}
for _, name := range names {
for dt := range data.DualTreeIterator(tree1, tree2) {
if dt.Error != nil {
return dt.Error
}
if ctx.Err() != nil {
return ctx.Err()
}
node1, t1 := tree1Nodes[name]
node2, t2 := tree2Nodes[name]
node1 := dt.Tree1
node2 := dt.Tree2
var name string
if node1 != nil {
name = node1.Name
} else {
name = node2.Name
}
addBlobs(stats.BlobsBefore, node1)
addBlobs(stats.BlobsAfter, node2)
switch {
case t1 && t2:
case node1 != nil && node2 != nil:
name := path.Join(prefix, name)
mod := ""
@@ -346,7 +319,7 @@ func (c *Comparer) diffTree(ctx context.Context, stats *DiffStatsContainer, pref
c.printError("error: %v", err)
}
}
case t1 && !t2:
case node1 != nil && node2 == nil:
prefix := path.Join(prefix, name)
if node1.Type == data.NodeTypeDir {
prefix += "/"
@@ -360,7 +333,7 @@ func (c *Comparer) diffTree(ctx context.Context, stats *DiffStatsContainer, pref
c.printError("error: %v", err)
}
}
case !t1 && t2:
case node1 == nil && node2 != nil:
prefix := path.Join(prefix, name)
if node2.Type == data.NodeTypeDir {
prefix += "/"