From 5101ead24cdabb8606da3a3b89901d5ec00c5bbd Mon Sep 17 00:00:00 2001
From: Alexander Weiss <alex@weissfam.de>
Date: Wed, 9 Feb 2022 08:49:20 +0100
Subject: [PATCH] add check command

---
 src/commands/check.rs | 111 ++++++++++++++++++++++++++++++++++++++++++
 src/commands/mod.rs   |   5 ++
 src/id.rs             |   4 ++
 3 files changed, 120 insertions(+)
 create mode 100644 src/commands/check.rs

diff --git a/src/commands/check.rs b/src/commands/check.rs
new file mode 100644
index 0000000..9d6f895
--- /dev/null
+++ b/src/commands/check.rs
@@ -0,0 +1,111 @@
+use anyhow::Result;
+use clap::Parser;
+use std::collections::HashMap;
+
+use crate::backend::{FileType, ReadBackend};
+use crate::blob::tree_iterator_once;
+use crate::index::{AllIndexFiles, BoomIndex, ReadIndex};
+use crate::repo::{IndexBlob, SnapshotFile};
+
+#[derive(Parser)]
+pub(super) struct Opts {
+    /// read all data blobs
+    #[clap(long)]
+    read_data: bool,
+}
+
+pub(super) fn execute(be: &impl ReadBackend, opts: Opts) -> Result<()> {
+    println!("checking packs...");
+    check_packs(be)?;
+
+    println!("loading index...");
+    let index = BoomIndex::from_iter(AllIndexFiles::new(be.clone()).into_iter());
+
+    println!("checking snapshots and trees...");
+    check_snapshots(be, &index)?;
+
+    if opts.read_data {
+        unimplemented!()
+    }
+
+    Ok(())
+}
+
+// calculate the pack size from the contained blobs
+fn pack_size(blobs: &[IndexBlob]) -> u32 {
+    let mut size = 4 + 32; // 4 + crypto overhead
+    for blob in blobs {
+        size += blob.length() + 37 // 37 = length of blob description
+    }
+    size
+}
+
+// check if packs correspond to index
+fn check_packs(be: &impl ReadBackend) -> Result<()> {
+    let mut packs = AllIndexFiles::new(be.clone())
+        .into_iter()
+        .map(|p| (*p.id(), pack_size(p.blobs())))
+        .collect::<HashMap<_, _>>();
+
+    for (id, size) in be.list_with_size(FileType::Pack)? {
+        match packs.remove(&id) {
+            None => println!("pack {} not contained in index", id.to_hex()),
+            Some(index_size) if index_size != size => println!(
+                "pack {}: size computed by index: {}, actual size: {}",
+                id.to_hex(),
+                index_size,
+                size
+            ),
+            _ => {} //everything ok
+        }
+    }
+
+    for (id, _) in packs {
+        println!(
+            "pack {} is referenced by the index but not presend!",
+            id.to_hex()
+        );
+    }
+
+    Ok(())
+}
+
+// check if all snapshots and contained trees can be loaded and contents exist in the index
+fn check_snapshots(be: &impl ReadBackend, index: &impl ReadIndex) -> Result<()> {
+    let snap_ids = be
+        .list(FileType::Snapshot)?
+        .into_iter()
+        .map(|id| SnapshotFile::from_backend(be, id).unwrap().tree)
+        .collect();
+
+    for path_node in tree_iterator_once(be, index, snap_ids) {
+        let fullname = path_node.path().join(path_node.node().name());
+        let node = path_node.node();
+
+        match node.tpe() as &str {
+            "file" => {
+                for (i, id) in node.content().iter().enumerate() {
+                    if id.is_null() {
+                        println!("file {:?} blob {} has null ID", fullname, i);
+                    }
+
+                    if index.get_id(id).is_none() {
+                        println!("file {:?} blob {} is missig in index", fullname, id);
+                    }
+                }
+            }
+
+            "dir" => {
+                if node.subtree().is_null() {
+                    println!("dir {:?} subtree has null ID", fullname);
+                }
+            }
+
+            "symlink" | "socket" | "chardev" | "dev" | "fifo" => {} // nothing to check
+
+            tpe => println!("file {:?} unkown type {}", fullname, tpe),
+        }
+    }
+
+    Ok(())
+}
diff --git a/src/commands/mod.rs b/src/commands/mod.rs
index 91a53f5..bf6c7e3 100644
--- a/src/commands/mod.rs
+++ b/src/commands/mod.rs
@@ -7,6 +7,7 @@ use crate::backend::{DecryptBackend, LocalBackend};
 use crate::repo;
 
 mod cat;
+mod check;
 mod list;
 mod ls;
 mod snapshots;
@@ -39,6 +40,9 @@ enum Command {
 
     /// ls snapshots
     Ls(ls::Opts),
+
+    /// check repository
+    Check(check::Opts),
 }
 
 pub fn execute() -> Result<()> {
@@ -54,5 +58,6 @@ pub fn execute() -> Result<()> {
         Command::Cat(opts) => cat::execute(&be, &dbe, opts),
         Command::Snapshots(opts) => snapshots::execute(&dbe, opts),
         Command::Ls(opts) => ls::execute(&dbe, opts),
+        Command::Check(opts) => check::execute(&dbe, opts),
     }
 }
diff --git a/src/id.rs b/src/id.rs
index 6f04657..6343260 100644
--- a/src/id.rs
+++ b/src/id.rs
@@ -33,4 +33,8 @@ impl Id {
     pub fn to_hex(self) -> String {
         hex::encode(self.0)
     }
+    
+    pub fn is_null(self) -> bool {
+        self == Id::default()
+    }
 }