gc.rs 91.5%

130 of 142 lines run, 6 of 12 functions · run #292 on main · run missed

//! Garbage collection (spec §21): delete objects
//! nothing reachable refers to.
//!
//! Roots: the current view, the views of every operation newer than
//! `core.op-retention` (their working copies only within the shorter
//! `core.snapshot-retention`), reference attestations, and the policy; receipts
//! live as long as the changes of their state. Operation
//! objects are always kept so the log stays walkable, but an operation older
//! than the retention window no longer keeps its states alive, so undoing
//! that far back may fail once its objects are gone.

use super::*;
use crate::store::{PackReport, PruneReport};
use std::collections::HashSet;

#[derive(Debug, Clone, Default, PartialEq, Eq)]
pub struct GcReport {
    pub packed: PackReport,
    pub pruned: PruneReport,
    pub reachable: usize,
    /// Packs folded into the one a full repack writes.
    pub merged_packs: usize,
}

impl Repo {
    /// Prune unreachable objects older than the grace period (`now`
    /// ignores it), then repack everything into one pack (spec §13.4).
    pub fn gc(&self, now: bool) -> Result<GcReport> {
        self.gc_with(now, &mut |_| {})
    }

    /// [`Repo::gc`], reporting each phase and the packing's progress
    /// (oxygen#66), for a spinner or a log.
    pub fn gc_with(&self, now: bool, progress: crate::store::Progress) -> Result<GcReport> {
        self.gc_keeping(now, &[], progress)
    }

    /// [`Repo::gc_with`], keeping `extra` blobs and their chunks as well:
    /// objects a host stored for itself and names from outside the
    /// repository's history, such as a forge's job artifacts
    /// (oxygenhub#126).
    pub fn gc_keeping(
        &self,
        now: bool,
        extra: &[ObjectRef],
        progress: crate::store::Progress,
    ) -> Result<GcReport> {
        progress("finding what is reachable");
        let mut keep = self.reachable()?;
        for r in extra {
            if !self.store.has(r) || !keep.insert(*r) {
                continue;
            }
            if let Ok(body) = self.store.get_typed(r, ObjectType::Blob) {
                if let Ok(blob) = Blob::decode(&body) {
                    keep.extend(blob.chunks);
                }
            }
        }
        let grace = if now {
            std::time::Duration::ZERO
        } else {
            std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
        };
        let pruned = self.store.prune_with(&keep, grace, progress)?;
        // Then everything left in one pack, compressed harder than a
        // commit's (spec §13.4), as `git gc` repacks.
        let (packed, merged_packs) = self
            .store
            .repack_all_with(crate::store::REPACK_ZSTD_LEVEL, progress)?;
        Ok(GcReport {
            packed,
            pruned,
            reachable: keep.len(),
            merged_packs,
        })
    }

    /// Every object that must survive garbage collection.
    pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
        let mut keep: HashSet<ObjectRef> = HashSet::new();
        let mut states: Vec<ObjectRef> = Vec::new();
        let mut changes: Vec<ChangeHash> = Vec::new();
        let mut attests: Vec<ObjectRef> = Vec::new();
        let mut policies: Vec<ObjectRef> = Vec::new();

        // Operations: always kept; views within retention are roots. Past
        // working copies only within the shorter `core.snapshot-retention`
        // (oxygen#98): a file only ever snapshotted, such as a build output
        // a status took in, goes once they expire.
        let now_ms = Hlc::now().wall_ms;
        let cutoff = now_ms.saturating_sub(self.config.op_retention_days * 86_400_000);
        let wc_cutoff = now_ms
            .saturating_sub(self.config.snapshot_retention_days * 86_400_000)
            .max(cutoff);
        for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
            keep.insert(r);
            if i == 0 || op.hlc.wall_ms >= cutoff {
                states.extend(op.view.bookmarks.values().copied());
                states.extend(op.view.remotes.values().flat_map(|b| b.values().copied()));
                policies.extend(op.view.policy);
            }
            if i == 0 || (self.config.snapshot_retention_days > 0 && op.hlc.wall_ms >= wc_cutoff) {
                changes.extend(op.view.workspaces.values().copied());
            }
        }
        // Checkpoints: the object, its payload blob, and the blob's chunks.
        for r in self.checkpoint_refs()? {
            if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
                keep.insert(r);
                let cp = crate::objects::Checkpoint::decode(&bytes)?;
                keep.insert(cp.payload);
                if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
                    keep.extend(Blob::decode(&b)?.chunks);
                }
                // A version 2 payload's graph blobs and their chunks.
                for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
                    keep.insert(g);
                    if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
                        keep.extend(Blob::decode(&b)?.chunks);
                    }
                }
            }
        }
        // Tags: their states and attestations.
        let view_now = self.view()?;
        states.extend(view_now.tags.values().copied());
        for name in view_now.tags.keys() {
            if let Some((r, _)) = self.tag_attest(name)? {
                attests.push(r);
            }
        }
        // Reference attestations.
        for name in self.view()?.bookmarks.keys() {
            if let Some((r, _)) = self.latest_ref_attest(name)? {
                attests.push(r);
            }
        }
        // A historical view may name objects pruned under an earlier, shorter
        // retention; there is nothing to keep for those.
        for st in states {
            if keep.insert(st) {
                match self.store.get_state(&st) {
                    Ok(s) => changes.extend(s.heads),
                    Err(StoreError::NotFound(_)) => continue,
                    Err(e) => return Err(e.into()),
                }
            }
        }
        for p in policies {
            if keep.insert(p) {
                let bytes = match self.store.get_typed(&p, ObjectType::Policy) {
                    Ok(b) => b,
                    Err(StoreError::NotFound(_)) => continue,
                    Err(e) => return Err(e.into()),
                };
                let policy = crate::policy::Policy::decode(&bytes)?;
                for id in &policy.identities {
                    keep.insert(id.hash());
                }
                for sig in self.store.signatures(&p)? {
                    keep.insert(sig.key);
                }
            }
        }
        // Changes: their closure, identities, blobs, chunks, seals.
        let mut stack = changes;
        while let Some(h) = stack.pop() {
            if !keep.insert(h) {
                continue;
            }
            // Beyond the shallow boundary, or pruned under an earlier
            // retention, there is nothing to keep.
            let c = match self.store.get_change_parts(&h) {
                Ok((c, texts)) => {
                    keep.extend(texts);
                    c
                }
                Err(StoreError::NotFound(_)) => continue,
                Err(e) => return Err(e.into()),
            };
            stack.extend(c.deps.iter().copied());
            if !c.author.is_zero() {
                keep.insert(c.author);
            }
            for sig in self.store.signatures(&h)? {
                keep.insert(sig.key);
            }
            for e in &c.edits {
                if let Edit::Replace { old, new, .. } = e {
                    for b in [old, new].into_iter().flatten() {
                        if keep.insert(*b) {
                            if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
                                keep.extend(Blob::decode(&bytes)?.chunks);
                            }
                        }
                    }
                }
            }
            attests.extend(self.store.seals_for(&h)?);
        }
        self.reachable_receipts(&mut keep)?;
        for a in attests {
            if keep.insert(a) {
                for sig in self.store.signatures(&a)? {
                    keep.insert(sig.key);
                }
            }
        }
        Ok(keep)
    }
}

`/`	Search
`g` `c`	Code
`g` `h`	History
`g` `b`	Bookmarks
`g` `t`	Trust
`y`	Permalink to this state (files and trees)
`?`	This sheet

1		//! Garbage collection (spec §21): delete objects
2		//! nothing reachable refers to.
3		//!
4		//! Roots: the current view, the views of every operation newer than
5		//! `core.op-retention` (their working copies only within the shorter
6		//! `core.snapshot-retention`), reference attestations, and the policy; receipts
7		//! live as long as the changes of their state. Operation
8		//! objects are always kept so the log stays walkable, but an operation older
9		//! than the retention window no longer keeps its states alive, so undoing
10		//! that far back may fail once its objects are gone.
11
12		use super::*;
13		use crate::store::{PackReport, PruneReport};
14		use std::collections::HashSet;
15
16		#[derive(Debug, Clone, Default, PartialEq, Eq)]
17		pub struct GcReport {
18		pub packed: PackReport,
19		pub pruned: PruneReport,
20		pub reachable: usize,
21		/// Packs folded into the one a full repack writes.
22		pub merged_packs: usize,
23		}
24
25		impl Repo {
26		/// Prune unreachable objects older than the grace period (`now`
27		/// ignores it), then repack everything into one pack (spec §13.4).
28	14×	pub fn gc(&self, now: bool) -> Result<GcReport> {
29	78×	self.gc_with(now, &mut \|_\| {})
30	14×	}
31
32		/// [`Repo::gc`], reporting each phase and the packing's progress
33		/// (oxygen#66), for a spinner or a log.
34	15×	pub fn gc_with(&self, now: bool, progress: crate::store::Progress) -> Result<GcReport> {
35	15×	self.gc_keeping(now, &[], progress)
36	15×	}
37
38		/// [`Repo::gc_with`], keeping `extra` blobs and their chunks as well:
39		/// objects a host stored for itself and names from outside the
40		/// repository's history, such as a forge's job artifacts
41		/// (oxygenhub#126).
42	16×	pub fn gc_keeping(
43	16×	&self,
44	16×	now: bool,
45	16×	extra: &[ObjectRef],
46	16×	progress: crate::store::Progress,
47	16×	) -> Result<GcReport> {
48	16×	progress("finding what is reachable");
49	16×	let mut keep = self.reachable()?;
50	16×	for r in extra {
51	1×	if !self.store.has(r) \|\| !keep.insert(*r) {
52	0×	continue;
53	1×	}
54	1×	if let Ok(body) = self.store.get_typed(r, ObjectType::Blob) {
55	1×	if let Ok(blob) = Blob::decode(&body) {
56	1×	keep.extend(blob.chunks);
57	1×	}
58	0×	}
59		}
60	16×	let grace = if now {
61	14×	std::time::Duration::ZERO
62		} else {
63	2×	std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
64		};
65	16×	let pruned = self.store.prune_with(&keep, grace, progress)?;
66		// Then everything left in one pack, compressed harder than a
67		// commit's (spec §13.4), as `git gc` repacks.
68	16×	let (packed, merged_packs) = self
69	16×	.store
70	16×	.repack_all_with(crate::store::REPACK_ZSTD_LEVEL, progress)?;
71	16×	Ok(GcReport {
72	16×	packed,
73	16×	pruned,
74	16×	reachable: keep.len(),
75	16×	merged_packs,
76	16×	})
77	16×	}
78
79		/// Every object that must survive garbage collection.
80	16×	pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
81	16×	let mut keep: HashSet<ObjectRef> = HashSet::new();
82	16×	let mut states: Vec<ObjectRef> = Vec::new();
83	16×	let mut changes: Vec<ChangeHash> = Vec::new();
84	16×	let mut attests: Vec<ObjectRef> = Vec::new();
85	16×	let mut policies: Vec<ObjectRef> = Vec::new();
86
87		// Operations: always kept; views within retention are roots. Past
88		// working copies only within the shorter `core.snapshot-retention`
89		// (oxygen#98): a file only ever snapshotted, such as a build output
90		// a status took in, goes once they expire.
91	16×	let now_ms = Hlc::now().wall_ms;
92	16×	let cutoff = now_ms.saturating_sub(self.config.op_retention_days * 86_400_000);
93	16×	let wc_cutoff = now_ms
94	16×	.saturating_sub(self.config.snapshot_retention_days * 86_400_000)
95	16×	.max(cutoff);
96	116×	for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
97	116×	keep.insert(r);
98	116×	if i == 0 \|\| op.hlc.wall_ms >= cutoff {
99	78×	states.extend(op.view.bookmarks.values().copied());
100	78×	states.extend(op.view.remotes.values().flat_map(\|b\| b.values().copied()));
101	78×	policies.extend(op.view.policy);
102	38×	}
103	116×	if i == 0 \|\| (self.config.snapshot_retention_days > 0 && op.hlc.wall_ms >= wc_cutoff) {
104	75×	changes.extend(op.view.workspaces.values().copied());
105	75×	}
106		}
107		// Checkpoints: the object, its payload blob, and the blob's chunks.
108	16×	for r in self.checkpoint_refs()? {
109	2×	if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
110	2×	keep.insert(r);
111	2×	let cp = crate::objects::Checkpoint::decode(&bytes)?;
112	2×	keep.insert(cp.payload);
113	2×	if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
114	2×	keep.extend(Blob::decode(&b)?.chunks);
115	0×	}
116		// A version 2 payload's graph blobs and their chunks.
117	2×	for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
118	1×	keep.insert(g);
119	1×	if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
120	1×	keep.extend(Blob::decode(&b)?.chunks);
121	0×	}
122		}
123	0×	}
124		}
125		// Tags: their states and attestations.
126	16×	let view_now = self.view()?;
127	16×	states.extend(view_now.tags.values().copied());
128	16×	for name in view_now.tags.keys() {
129	3×	if let Some((r, _)) = self.tag_attest(name)? {
130	3×	attests.push(r);
131	3×	}
132		}
133		// Reference attestations.
134	16×	for name in self.view()?.bookmarks.keys() {
135	16×	if let Some((r, _)) = self.latest_ref_attest(name)? {
136	0×	attests.push(r);
137	16×	}
138		}
139		// A historical view may name objects pruned under an earlier, shorter
140		// retention; there is nothing to keep for those.
141	97×	for st in states {
142	97×	if keep.insert(st) {
143	47×	match self.store.get_state(&st) {
144	44×	Ok(s) => changes.extend(s.heads),
145	3×	Err(StoreError::NotFound(_)) => continue,
146	0×	Err(e) => return Err(e.into()),
147		}
148	50×	}
149		}
150	16×	for p in policies {
151	4×	if keep.insert(p) {
152	1×	let bytes = match self.store.get_typed(&p, ObjectType::Policy) {
153	1×	Ok(b) => b,
154	0×	Err(StoreError::NotFound(_)) => continue,
155	0×	Err(e) => return Err(e.into()),
156		};
157	1×	let policy = crate::policy::Policy::decode(&bytes)?;
158	1×	for id in &policy.identities {
159	1×	keep.insert(id.hash());
160	1×	}
161	1×	for sig in self.store.signatures(&p)? {
162	1×	keep.insert(sig.key);
163	1×	}
164	3×	}
165		}
166		// Changes: their closure, identities, blobs, chunks, seals.
167	16×	let mut stack = changes;
168	240×	while let Some(h) = stack.pop() {
169	224×	if !keep.insert(h) {
170	113×	continue;
171	111×	}
172		// Beyond the shallow boundary, or pruned under an earlier
173		// retention, there is nothing to keep.
174	111×	let c = match self.store.get_change_parts(&h) {
175	101×	Ok((c, texts)) => {
176	101×	keep.extend(texts);
177	101×	c
178		}
179	10×	Err(StoreError::NotFound(_)) => continue,
180	0×	Err(e) => return Err(e.into()),
181		};
182	101×	stack.extend(c.deps.iter().copied());
183	101×	if !c.author.is_zero() {
184	30×	keep.insert(c.author);
185	71×	}
186	101×	for sig in self.store.signatures(&h)? {
187	13×	keep.insert(sig.key);
188	13×	}
189	111×	for e in &c.edits {
190	111×	if let Edit::Replace { old, new, .. } = e {
191	18×	for b in [old, new].into_iter().flatten() {
192	18×	if keep.insert(*b) {
193	18×	if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
194	18×	keep.extend(Blob::decode(&bytes)?.chunks);
195	0×	}
196	0×	}
197		}
198	93×	}
199		}
200	101×	attests.extend(self.store.seals_for(&h)?);
201		}
202	16×	self.reachable_receipts(&mut keep)?;
203	16×	for a in attests {
204	5×	if keep.insert(a) {
205	4×	for sig in self.store.signatures(&a)? {
206	4×	keep.insert(sig.key);
207	4×	}
208	1×	}
209		}
210	16×	Ok(keep)
211	16×	}
212		}

Coverage / crates/oxy-core/src/repo/gc.rs 91.5%

Keyboard