CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.28.0 (2022-04-05)

New Features

• Handle::packed_object_count() Provide packed objects numbers and cache the value for fast access later on.

Changed (BREAKING)

• Remove deprecated compound and linked object databases The dynamic/general store is the only maintained can-do-it-all DB now.

Commit Statistics

• 2 commits contributed to the release over the course of 1 calendar day.
• 2 days passed between releases.
• 2 commits where understood as conventional.
• 1 unique issue was worked on: #298

Commit Details

view details

• #298
  • Handle::packed_object_count() (84ec54e)
• Uncategorized
  • Remove deprecated compound and linked object databases (8c5ae77)

0.27.0 (2022-04-03)

• fixes a race condition around the first initialization of an ODB, which could leave the loosing thread without any packs are loose databases.
• Adds support for lookups by prefix.

New Features

• loose::Store::lookup_prefix(…)

Bug Fixes

• support Rust 1.52

Commit Statistics

• 45 commits contributed to the release over the course of 68 calendar days.
• 69 days passed between releases.
• 2 commits where understood as conventional.
• 6 unique issues were worked on: #298, #301, #329, #331, #333, #364

Thanks Clippy

Clippy helped 1 time to make code idiomatic.

Commit Details

view details

• #298
  • docs (a45f378)
  • Implement ODB::disambiguate_prefix(…) (7d4d281)
  • Add missing docs (c7cd4ab)
  • refactor (4455af3)
  • refactor (fe87704)
  • lookup_prefix for ODB (a4ccd18)
  • base for testing git_odb::Handle::lookup_prefix() (6244c06)
  • basics for ODB lookup prefix (5c228e1)
  • add docs to loose::Store::lookup_prefix(…) (9fa4817)
  • upgrade parking_lot and cargo_toml (f95c1a0)
  • loose::Store::lookup_prefix(…) (996bfb3)
  • simplify error of loose::Iter (622abd7)
  • support Rust 1.52 (9c14de3)
• #301
  • adapt to changes in git-quote (ba48629)
  • use git-quote crate in git-odb alternate parsing (8e49aa6)
  • all path-related tests are green (81d2bf2)
  • Update git-odb changelog to include information about bugfix (055b117)
  • fix the first race-condition around initialization in ODB (a891315)
  • conversions from Rc to arc for Handle (c19331e)
  • less restrictive ref-delta-base resolution (917480b)
  • more safety around recursion and invariants when resolving ref-deltas (dddb4a5)
  • allow delta base objects to be out-of-pack in general odb (d4f1590)
  • elaborate odb info and simple entries printing (0f65282)
  • a first sketch of access odb information using a sub-command (89b628a)
• #329
  • Document all features related to serde1 (72b97f2)
• #331
  • Adapt to changes in git_features::path to deal with Result (bba4c68)
• #333
  • Use git_features::path everywhere where there is a path conversion (2e1437c)
• #364
  • A statement about replacement objects (2d32f4d)
  • Implement object replacement (b16d5e9)
  • add some precaution to avoid strange interactions with packs (b052a9a)
  • An API and a test for replacement configuration (f2d6db1)
  • Initial test to assure we don't replace objects by default (6cb9ecc)
• Uncategorized
  • Release git-diff v0.14.0, git-bitmap v0.1.0, git-index v0.2.0, git-tempfile v2.0.1, git-lock v2.0.0, git-mailmap v0.1.0, git-traverse v0.13.0, git-pack v0.17.0, git-quote v0.2.0, git-odb v0.27.0, git-packetline v0.12.4, git-url v0.4.0, git-transport v0.16.0, git-protocol v0.15.0, git-ref v0.12.0, git-worktree v0.1.0, git-repository v0.15.0, cargo-smart-release v0.9.0, safety bump 5 crates (e58dc30)
  • Merge branch 'for-onefetch' (8e5cb65)
  • Release git-hash v0.9.3, git-features v0.20.0, git-config v0.2.0, safety bump 12 crates (f0cbb24)
  • make fmt (7cf3545)
  • Merge branch 'describe-rev' (77b7cd9)
  • adapt to breaking changes in git-actor (40c48c3)
  • thanks clippy (4618f8a)
  • Release git-tempfile v2.0.0, safety bump 6 crates (90b1c42)
  • Merge branch 'AP2008-implement-worktree' (f32c669)
  • Release git-bitmap v0.0.1, git-hash v0.9.0, git-features v0.19.0, git-index v0.1.0, safety bump 9 crates (4624725)
  • Release git-hash v0.9.2, git-object v0.17.1, git-pack v0.16.1 (0db19b8)
  • Implement git_odb::Find for git_odb::Handle (3522aef)
  • Merge branch 'index-information' (025f157)

0.26.0 (2022-01-23)

Refactor

• replace bare u32 data::Id typedef

Other

• try LRU-like contains implementation Which unfortunately isn't really faster at all even though it totally should be.
• Try to make Handle usable for pack creation It's nearly there, but for some reason the boxed dyn traits don't get to be Send even though it's specified.
• :Find for Arc and Rc
• :Find implementation for linked::Store

Chore

• remove unused dependencies

New Features

• git_pack::Find::try_find_cached(…, pack_cache) With this method it's easier to bypass local caches and control the cache oneself entirely.
• add linked::Store::rc_iter() For completeness in case of single-threaded operations
• linked::Store sorts bundles by modification date, newest first
• A simplified version of the Find trait It's meant for the next generation of object db handles which keep a local cache of all the details of the actual object database.

Bug Fixes

• support Rust 1.52
• linked::Store now assures unique IDs across compound stores

Changed (BREAKING)

• cleanup and unify verify_integrity() method signature Previously they used many different ways of handling their parameters despite all boiling down to calling the same 'index::File::traverse()` method.

  This allows for more reuse of Options structs and generally makes clearer how these optinos are used.

• Change accessors named hash_kind() to object_hash() for consistency

• consistently use object_hash instead of hash_kind

• remove Write::*(…, hash_kind) The hash_kind is now intrinsic to the implementation of the write trait and thus isn't passed along anymore in parameters.

  The sink() function now takes the kind of hash as parameter.

• various changes to the loose::Store.

  Change path field to read-only path() method add hash_kind parameter to loose::Store::at(…, hash_kind)

• move sink::Sink to the top-level exclusively

• move loose::iter::Iter to loose::Iter

• remove make_object_cache parameter from git_pack::data::output::count::objects() It now is an implementation detail of the Find trait.

• Rename Handle to Cache Because this is exactly what it is effectively. Also add some basic instantiation for the new object store.

• remove pack-cache from Find::try_find(…) With the new architecture this can be an implementation detail without forcing it to be Sync.

• move git_pack::data::Object to git_object::Data, massively alter git_odb::Find trait This will break a lot, but has to happen to prepare these traits for the next generation of object databases.

New Features (BREAKING)

• Cache::inner removed in favor of Deref/Mut and into_inner() Making the inner field available allows changing it, which would make it potentially incompatible with existing caches. The new implementation makes it essentially read-only while allowing more convenient access to methods on inner.

Commit Statistics

• 162 commits contributed to the release over the course of 51 calendar days.
• 55 days passed between releases.
• 23 commits where understood as conventional.
• 5 unique issues were worked on: #260, #266, #279, #287, #293

Thanks Clippy

Clippy helped 4 times to make code idiomatic.

Commit Details

view details

• #260
  • linked::Store now assures unique IDs across compound stores (b605c1f)
• #266
  • More explicit information about how much garbaged is in the slotmap (cfd36ee)
  • assure stable handles can actually access the indices hey need (9474a43)
  • a failing test to show the handle-stability doesn't quite work yet (5562e88)
  • refactor (c499843)
  • docs for dynamic object store (2c2a2e9)
  • Default handle refresh mode is the least surprising, with option to configure (1b74c14)
  • remove unused dependencies (c800fdd)
  • refactor (b88f253)
  • refactor (52a4dcd)
  • refactor (3da91ce)
  • move sink::Sink to the top-level exclusively (ab4e726)
  • dynamic store module cleanu (494772c)
  • adapt to changes in git-odb (a44dd4b)
  • move loose::iter::Iter to loose::Iter (8bb5c9a)
  • minor improvements to module layout, docs (0364f48)
  • fix docs (360bf9d)
  • Make single-threaded programs possible to use with git-repository (dde5c6b)
  • A more suitable iterator implementation for general store (af0cc5f)
  • A quick and dirty version index iteration (0384007)
  • Use new store in git-repository (2f9e342)
  • Use new odb in place of the old one and it works (8ad25c5)
  • remaining methods of git-pack::Find (92b9764)
  • Make find::Entry self-contained (ad36fb9)
  • Remove iterator access in favor of fully owned data (62d3f10)
  • It shows that we can't return anything referenced from the interior-mutable handle (b9f308b)
  • refactor (6cb474e)
  • prepare implementation of location-dependent methods (5de29f4)
  • refactor (2c23f42)
  • Impl git_odb::Write for general::Handle (b7a6ab7)
  • cleanup (a4f3670)
  • A way to predict the amount of slots needed for smooth operation (a3a16d6)
  • finding objects and dynamically loading packs seems to work (8f58c30)
  • add docs for handle-related functions (cf1b1e6)
  • first sketch of looking up objects with pack cache (a0aae84)
  • add contains checks for libgit2 (c64a45a)
  • Add MRU to contains() even faster contains checks (6525847)
  • Load an index right after refreshing items from disk (0c40eb3)
  • some more assertions for contains() regarding refresh mode (3f8c540)
  • Try reusing mappings, but no LRU (bb602e8)
  • try LRU-like contains implementation (2d6960f)
  • Don't abort run into assertion unnecessarily… (4e87a56)
  • Adjust object-acess to test new contains method (8488b41)
  • Looks like 'contains()' is implemented well enough (c24015a)
  • First successful loading of indices… (8cbef57)
  • first stab at loading indices while dealing with inherent raciness (94be3a0)
  • assure loops can't happen anymore (f04ff80)
  • The first green test for loose object lookup (0c6b7b1)
  • impl Write for Arc, Rc and shared borrows (5cdc27d)
  • more clarity around generations; actually trash slots or unset them (7bce101)
  • Adapt to changes in git-repository (3ab9b03)
  • Create a new index snapshot and store it (41d91f9)
  • Sort out race condition around slots that change identity (6e678e7)
  • Uncover slightly disturbing races which make it hard to ever release/unload maps (cbf2d13)
  • remaining comments about multi-pack-index handling (569c40b)
  • handling of multi-pack index slot-map moves as they change (456f1e7)
  • Also use object cache if there is no pack-cache (3e1ae25)
  • prepare correct handling of multi-pack indices when the time comes (6388ba2)
  • More thoughts about how to continue a search… (eac8c45)
  • remove make_object_cache parameter from git_pack::data::output::count::objects() (3f05fea)
  • get closer to actually setting up slots (55645ae)
  • Cache-creators are indeed shared across threads, must be sync (c326cb3)
  • Add load-pack method frame to not forget (e1ec535)
  • Try to make Handle usable for pack creation (424c9b3)
  • Some steps towards recording the disk state in the slot-map (5074f4c)
  • make all handle caches and cache creators optional (3c30769)
  • better errors for disk consolidation; setup loose object dbs (4c13e14)
  • Make odb handle more general (2be6725)
  • Detail the load-indices flow more (11d3325)
  • Make slot-count configurable (c910af5)
  • :Find for Arc and Rc (b1c82a7)
  • Provide handle with a snapshot of the store's state (6e0cd6d)
  • MultiPackIndex compatible pack::Find trait definition (5fa1a9d)
  • refactor (d5565da)
  • git_pack::Find::try_find_cached(…, pack_cache) (58c2edb)
  • construct a handle from a linked store (9702ed4)
  • refactor (ab14f70)
  • A first tiny stab at Handle for linked store (ef08f7f)
  • Support for metrics in general store handle (11b98b8)
  • refactor (1361c31)
  • first test to trigger all major code-paths (25b56c5)
  • refactor (3310d8f)
  • more trustworthy state-id hashing (4eb43d0)
  • :Find implementation for linked::Store (9235106)
  • Use handle registration to avoid unloading packs; fix state-id hash (a1070de)
  • add linked::Store::rc_iter() (36fde72)
  • Handle registration (df4e4eb)
  • Use Deref instead of Borrow in linked ODB iterator (a96b1e5)
  • Bring in the slotmap (3a5cb5f)
  • put down more types for loading of indices and refresh logic (9909eaf)
  • remove pack-cache from Find::try_find(…) (ebc7f47)
  • Add all types the handle would have to store (e2f0cb0)
  • linked::Store sorts bundles by modification date, newest first (a81b333)
  • Rename Handle to Cache (580e96c)
  • First sketch of general store (fc1b640)
  • move git_pack::data::Object to git_object::Data, massively alter git_odb::Find trait (2290d00)
  • More affirmative notes about multi-pack indices (dceaea2)
  • A simplified version of the Find trait (e25f4ea)
  • Remove CRC32 check entirely as it doesn't seem to be important in the big picture (22d35bd)
  • Notes about multi-pack indices in the current data::entry::location (7eff6bf)
  • Adjust to new name/place of bundle::Location (1f8954d)
  • Add 'contains()' method to Find (dfdd6fb)
• #279
  • add missing docs (4137327)
  • cargo fmt (8b9da35)
  • very first experimental support for multi-pack index verification (bb35c69)
  • multi-index verify checksum (853d468)
  • Another test to validate index stability (1fb08df)
  • Change accessors named hash_kind() to object_hash() for consistency (2ef9a84)
  • refactor (a4ad5bf)
  • handle 'move' of multi-pack indices (aec0895)
  • trigger last portion of multi-index logic (b46979b)
  • load slots early to avoid races with the 'generation' field (10bc3ab)
  • Only open multi-pack indices if the mtime changed (efe2579)
  • Add remaining Store docs (e440bcd)
  • first odb lookup with multi-index works (a8773df)
  • Generalize intra-pack offset lookup (dff05ef)
  • consistently use object_hash instead of hash_kind (49998cc)
  • replace bare u32 data::Id typedef (e0b8636)
  • A rough implementation of everything multi-index support would need… (56f174f)
  • Make opened multi-pack indices representable (28e648d)
  • first sketch towards reading in multi-indices (25eb157)
  • Cache::inner removed in favor of Deref/Mut and into_inner() (bf73a94)
  • refactor (7331e99)
  • adapt to changes in git-pack (28dba20)
  • adjust to changes in git-pack (b8f109e)
  • refactor (9b5451a)
  • loose object iteration respects hash kind (72eb9da)
  • remove Write::*(…, hash_kind) (67c42fb)
  • various changes to the loose::Store (ad1b9ea)
  • loose odb doesn't hard-code sha1 anymore for find() and contains() (68f1031)
  • First pieces of header parsing; allow to respect multi-index desired hash kind in git-odb (1a2a049)
  • Respect core.multiPackIndex option (1495efc)
  • only load multi-pack indices if allowed (b22e146)
• #287
  • basic output for 'repo verify' json only (9f8d61f)
  • way nicer progress messages for repo verification (4b4f9f8)
  • upgrade to prodash 17 (47860b7)
  • better verify progress printing (4a464f2)
  • refactor (831397c)
  • refactor (38426a1)
  • Very rough version of repository verification (80a4a7a)
  • support for loose object statistics in odb store (53d835a)
  • bare-bones loose object integrity check (3dfec81)
  • frame for loose-db validation (a24307d)
  • First basic validation of all packs within an odb store (d63176f)
  • Don't reset generations, instead make them match the current one (1d995ef)
  • frame for interity check on object store (b5dd059)
  • cleanup and unify verify_integrity() method signature (91d0476)
• #293
  • refactor (9b28b18)
• Uncategorized
  • Release git-odb v0.26.0, git-packetline v0.12.3, git-url v0.3.5, git-transport v0.15.0, git-protocol v0.14.0, git-ref v0.11.0, git-repository v0.14.0, cargo-smart-release v0.8.0 (42ebb53)
  • Release git-diff v0.13.0, git-tempfile v1.0.4, git-chunk v0.3.0, git-traverse v0.12.0, git-pack v0.16.0, git-odb v0.26.0, git-packetline v0.12.3, git-url v0.3.5, git-transport v0.15.0, git-protocol v0.14.0, git-ref v0.11.0, git-repository v0.14.0, cargo-smart-release v0.8.0 (1b76119)
  • Release git-actor v0.8.0, git-config v0.1.10, git-object v0.17.0, git-diff v0.13.0, git-tempfile v1.0.4, git-chunk v0.3.0, git-traverse v0.12.0, git-pack v0.16.0, git-odb v0.26.0, git-packetline v0.12.3, git-url v0.3.5, git-transport v0.15.0, git-protocol v0.14.0, git-ref v0.11.0, git-repository v0.14.0, cargo-smart-release v0.8.0 (8f57c29)
  • Release git-features v0.19.1, git-actor v0.8.0, git-config v0.1.10, git-object v0.17.0, git-diff v0.13.0, git-tempfile v1.0.4, git-chunk v0.3.0, git-traverse v0.12.0, git-pack v0.16.0, git-odb v0.26.0, git-packetline v0.12.3, git-url v0.3.5, git-transport v0.15.0, git-protocol v0.14.0, git-ref v0.11.0, git-repository v0.14.0, cargo-smart-release v0.8.0 (d78aab7)
  • Release git-hash v0.9.1, git-features v0.19.1, git-actor v0.8.0, git-config v0.1.10, git-object v0.17.0, git-diff v0.13.0, git-tempfile v1.0.4, git-chunk v0.3.0, git-traverse v0.12.0, git-pack v0.16.0, git-odb v0.26.0, git-packetline v0.12.3, git-url v0.3.5, git-transport v0.15.0, git-protocol v0.14.0, git-ref v0.11.0, git-repository v0.14.0, cargo-smart-release v0.8.0, safety bump 4 crates (373cbc8)
  • prepare changelogs for release (674ec73)
  • prepar changelogs for cargo-smart-release release (8900d69)
  • Release git-chunk v0.2.0, safety bump 4 crates (b792fab)
  • refactor (c09a44d)
  • thanks clippy (bf4694c)
  • thanks clippy (17af184)
  • make fmt (066f3ff)
  • thanks clippy (123a95e)
  • thanks clippy (4ca9e07)

0.25.0 (2021-11-29)

Maintenance release due, which isn't really required but one now has to be careful what's committed at once.

Refactor (BREAKING)

• move loose header manipulation from git-pack to git-object

Commit Statistics

• 7 commits contributed to the release over the course of 25 calendar days.
• 12 days passed between releases.
• 1 commit where understood as conventional.
• 2 unique issues were worked on: #250, #259

Commit Details

view details

• #250
  • move loose header manipulation from git-pack to git-object (598698b)
• #259
  • btree/hashmap free lookup of packs in store, keeping things more bundled (a88981b)
• Uncategorized
  • Release git-actor v0.7.0, git-config v0.1.9, git-object v0.16.0, git-diff v0.12.0, git-traverse v0.11.0, git-pack v0.15.0, git-odb v0.25.0, git-packetline v0.12.2, git-transport v0.14.0, git-protocol v0.13.0, git-ref v0.10.0, git-repository v0.13.0, cargo-smart-release v0.7.0 (d3f9227)
  • Release git-features v0.18.0, git-actor v0.7.0, git-config v0.1.9, git-object v0.16.0, git-diff v0.12.0, git-traverse v0.11.0, git-pack v0.15.0, git-odb v0.25.0, git-packetline v0.12.2, git-transport v0.14.0, git-protocol v0.13.0, git-ref v0.10.0, git-repository v0.13.0, cargo-smart-release v0.7.0, safety bump 12 crates (acd3737)
  • Adjust changelogs prior to release (ec38950)
  • Merge branch 'git-loose-objects' of https://github.com/xmo-odoo/gitoxide into xmo-odoo-git-loose-objects (ee737cd)
  • Move "loose object header" ser/de to git-object (3d1565a)

0.24.0 (2021-11-16)

A maintenance release triggered by changes to git-pack and changelog rewrites.

Commit Statistics

• 4 commits contributed to the release.
• 27 days passed between releases.
• 0 commits where understood as conventional.
• 1 unique issue was worked on: #254

Commit Details

view details

• #254
  • Adjust changelogs prior to git-pack release (6776a3f)
• Uncategorized
  • Release git-config v0.1.8, git-object v0.15.1, git-diff v0.11.1, git-traverse v0.10.1, git-pack v0.14.0, git-odb v0.24.0, git-packetline v0.12.1, git-transport v0.13.1, git-protocol v0.12.1, git-ref v0.9.1, git-repository v0.12.0, cargo-smart-release v0.6.0 (f606fa9)
  • better changelog descriptions. (f69b2d6)
  • Adjusting changelogs prior to release of git-config v0.1.8, git-object v0.15.1, git-diff v0.11.1, git-traverse v0.10.1, git-pack v0.14.0, git-odb v0.24.0, git-packetline v0.12.1, git-transport v0.13.1, git-protocol v0.12.1, git-ref v0.9.1, git-repository v0.12.0, cargo-smart-release v0.6.0, safety bump 5 crates (39b40c8)

v0.23.0 (2021-10-19)

A maintenance release to properly dealing with previously breaking changes in git-hash.

Commit Statistics

• 4 commits contributed to the release.
• 3 days passed between releases.
• 0 commits where understood as conventional.
• 1 unique issue was worked on: #222

Commit Details

view details

• #222
  • update changelogs prior to release (9a493d0)
  • stabilize changelogs (920e832)
  • Update changelogs prior to release (b3e2252)
• Uncategorized
  • Release git-hash v0.8.0, git-features v0.17.0, git-actor v0.6.0, git-object v0.15.0, git-diff v0.11.0, git-traverse v0.10.0, git-pack v0.13.0, git-odb v0.23.0, git-packetline v0.12.0, git-transport v0.13.0, git-protocol v0.12.0, git-ref v0.9.0, git-repository v0.11.0, git-commitgraph v0.6.0, gitoxide-core v0.12.0, gitoxide v0.10.0, cargo-smart-release v0.5.0, safety bump 16 crates (0e02953)

v0.22.0 (2021-10-15)

Dependency Upgrade (BREAKING)

• git-traverse saw a breaking change moving to v0.9, which triggered this crate to signal a breaking change, too.

Improvements

• pack-ids as generated when instantiating a compound database are now sequential. That way, they never clash which was possible before when it was using a CRC32 over the filename.

  The latter was done to allow deduplicating packs, but it turns out not to be necessary just yet.

Commit Statistics

• 24 commits contributed to the release over the course of 31 calendar days.
• 35 days passed between releases.
• 0 commits where understood as conventional.
• 2 unique issues were worked on: #198, #67

Commit Details

view details

• #198
  • Adjust all changelogs to fulfil requirements for publishing (04b9ca0)
  • deduplicate conventional message ids (e695eda)
  • regenerate all changelogs to get links (0c81769)
  • Mention actual issues that where worked on (a517e39)
  • Allow 'refactor' and 'other' in conventional messages if they have breaking changes (4eebaac)
  • Rebuild all changelogs to assure properly ordered headlines (4a9a05f)
  • Sort all commits by time, descending… (f536bad)
  • greatly reduce changelog size now that the traversal fix is applied (a0bc98c)
  • Fixup remaining changelogs… (2f75db2)
  • Generate changelogs with details (e1861ca)
  • Update all changelogs with details (58ab2ae)
  • Update changelogs (c857d61)
  • Avoid adding newlines which make writing unstable (6b5c394)
  • Fix section headline level (9d6f263)
  • Write first version of changlogs thus far… (719b6bd)
  • Parse more user generated section content, adapt existing changelogs to work correctly (2f43a54)
• #67
  • Assure pack-ids are actually unique, the simple way… (0509b4f)
• Uncategorized
  • Release git-hash v0.7.0, git-features v0.16.5, git-actor v0.5.3, git-config v0.1.7, git-validate v0.5.3, git-object v0.14.1, git-diff v0.10.0, git-tempfile v1.0.3, git-lock v1.0.1, git-traverse v0.9.0, git-pack v0.12.0, git-odb v0.22.0, git-packetline v0.11.0, git-url v0.3.4, git-transport v0.12.0, git-protocol v0.11.0, git-ref v0.8.0, git-repository v0.10.0, cargo-smart-release v0.4.0 (59ffbd9)
  • Adjusting changelogs prior to release of git-hash v0.7.0, git-features v0.16.5, git-actor v0.5.3, git-validate v0.5.3, git-object v0.14.1, git-diff v0.10.0, git-tempfile v1.0.3, git-lock v1.0.1, git-traverse v0.9.0, git-pack v0.12.0, git-odb v0.22.0, git-packetline v0.11.0, git-url v0.3.4, git-transport v0.12.0, git-protocol v0.11.0, git-ref v0.8.0, git-repository v0.10.0, cargo-smart-release v0.4.0, safety bump 3 crates (a474395)
  • Don't let 'a' leak out of the tempdir to fix #212 (682dff8)
  • More proper fix for #212 (dfc1493)
  • naive fix for #212 (b31863c)
  • Update changelogs just for fun (21541b3)
  • Bump git-traverse v0.9.0, safety bump 8 crates (d39fabb)

v0.21.3 (2021-09-10)

Commit Statistics

• 2 commits contributed to the release.
• 1 day passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • Release git-odb v0.21.3 (223f930)
  • [smart-release #195] fix docs (8d7e132)

v0.21.2 (2021-09-08)

Commit Statistics

• 7 commits contributed to the release.
• 1 day passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • Release git-odb v0.21.2 (d443644)
  • Bump git-pack v0.11.0 (5ae6ff5)
  • Bump git-object v0.14.0 (d4fc81f)
  • [repository #164] Prepare commit() for a possible less-allocating future (0fd01f7)
  • [repository #164] generic write_object() (c569f83)
  • [repository #164] Support for refreshing the object database (46e10f8)
  • [odb #164] Add refresh() functionality (ee16d04)

v0.21.1 (2021-09-07)

Commit Statistics

• 4 commits contributed to the release over the course of 1 calendar day.
• 10 days passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • Release git-odb v0.21.1 (cb33c2f)
  • [odb #190] Read all eligble packed refs, no "pack-" prefix needed (ab250f7)
  • Bump git-pack v0.10.0 (e5e3c80)
  • Bump git-hash v0.6.0 (6efd90d)

v0.21.0 (2021-08-27)

Commit Statistics

• 16 commits contributed to the release over the course of 4 calendar days.
• 10 days passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • [repository #174] adjust various changelogs (081faf5)
  • [odb #180] fix docs (bd50752)
  • [odb #180] refactor (eff21da)
  • Bump git-odb v0.21.0 (7b9854f)
  • [odb #180] add changelog (acf1193)
  • [pack #179] refactor bundle (420dca2)
  • [pack #179] refactor (ab6554b)
  • [object #177] move mutable objects to crate::* (c551c02)
  • [object #177] migrate immutable::tree to crate::tree (fa5cd06)
  • [object #177] move immutable::* to crate::*Ref, start iter adjustments (461dc53)
  • [object #177] rename immutable::* to immutable::*Ref (6deb012)
  • Release git-object v0.13.0 (708fc5a)
  • Merge pull request #172 from mellowagain/main (61aebbf)
  • [actor #173] rename immutable::Signature to SignatureRef! (96461ac)
  • Release git-pack v0.9.0 (7fbc961)
  • [pack #67] refactor (14717f6)

v0.20.2 (2021-08-17)

Commit Statistics

• 2 commits contributed to the release over the course of 1 calendar day.
• 3 days passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • Release git-odb v0.20.2 (6fb8bbb)
  • Apply nightly rustfmt rules. (5e0edba)

v0.20.1 (2021-08-13)

Commit Statistics

• 2 commits contributed to the release.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • Release git-odb v0.20.1 (ca3f736)
  • remove dev-dependency cycles by removing their version (c40faca)

v0.20.0 (2021-08-12)

Commit Statistics

• 4 commits contributed to the release.
• 1 day passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • Release git-diff v0.8.0, git-odb v0.20.0, git-pack v0.8.0, git-traverse v0.7.0 (f123f69)
  • Release git-diff v0.7.0, git-odb v0.19.0, git-pack v0.7.0, git-traverse v0.6.0 (c67291f)
  • Release git-object v0.12.0 (7006150)
  • Release git-actor-0.3.1 (727087d)

v0.18.0 (2021-08-11)

Commit Statistics

• 2 commits contributed to the release.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • (cargo-release) version 0.18.0 (b327590)
  • (cargo-release) version 0.6.0 (d704bca)

v0.17.0 (2021-08-11)

Commit Statistics

• 7 commits contributed to the release.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • (cargo-release) version 0.17.0 (c52a491)
  • (cargo-release) version 0.5.0 (c2f94a5)
  • (cargo-release) version 0.4.0 (d69d0ac)
  • (cargo-release) version 0.11.0 (a5be31c)
  • (cargo-release) version 0.3.0 (64efc05)
  • (cargo-release) version 0.4.0 (70ef344)
  • [utils #154] refactor: bool.then(||this) - neat (1dec1c4)

v0.16.1 (2021-08-10)

Commit Statistics

• 1 commit contributed to the release.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • (cargo-release) version 0.16.1 (8cd173b)

v0.16.0 (2021-08-10)

Commit Statistics

• 113 commits contributed to the release over the course of 83 calendar days.
• 93 days passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • (cargo-release) version 0.3.0 (0e9c73a)
  • (cargo-release) version 0.5.0 (ae02dab)
  • (cargo-release) version 0.16.0 (1231dbd)
  • (cargo-release) version 0.2.0 (8ff5115)
  • clippy on tests and thanks clippy (a77a71c)
  • thanks clippy (e1964e4)
  • [pack] a way to obtain whole bundles for offset-to-index lookup (15fcbe2)
  • [pack] refactor (64b1dcd)
  • [pack] bundle::Location with pack offset; order counts by that… (f92f285)
  • Don't use ASM on windows for Sha1 as it fails to build there. (ba1fb7a)
  • Merge branch 'remove-unnecessary-unsafe' (7a3c5c1)
  • Remove unnecessary pub(crate) exports (3d2456e)
  • Bump thiserror from 1.0.25 to 1.0.26 (9682590)
  • [actor] fix gix hours (b4e95fd)
  • [actor] git-object uses git-actor (d01dd2f)
  • thanks clippy (3f7e27b)
  • (cargo-release) version 0.3.0 (6b33678)
  • Merge branch 'dependabot/cargo/crc-2.0.0' (683c44d)
  • (cargo-release) version 0.2.0 (3286e42)
  • thanks clippy (c5b4de8)
  • [git-odb] linked::Store can now check if an object exists (bb95c79)
  • refactor (a25a774)
  • (cargo-release) version 0.4.0 (866f86f)
  • [git-ref] the first failing test (7e802a0)
  • [git-odb] fix test compiilation (639bc10)
  • [git-odb] much better docs; cleanup exposed API (3d5b229)
  • (cargo-release) version 0.2.0 (b213628)
  • [git-odb] refactor (2958145)
  • [git-odb] refactor (1eab15d)
  • [git-odb] refactor (4967c22)
  • [git-odb] refactor (2e68e0c)
  • [git-odb] fix docs (936cfd3)
  • [git-pack] compilation (b392a55)
  • [git-pack] refactor (ea2b3de)
  • [git-pack] refactor (5ca2547)
  • [git-pack] refactor (157b6ff)
  • (cargo-release) version 0.16.0 (769c649)
  • [git-pack] refactor (be6ddaa)
  • [git-pack] refactor (1fab6af)
  • [git-pack] refactor (e5b00ee)
  • [git-pack] the world compiles again (f0c0e36)
  • [git-pack] used by git-odb (5d6ee07)
  • [git-features] refactor to help understand a zlib-related logic bug (ae826e8)
  • [git-features] a first step towards supporting a pure rust zlib backend (040cab7)
  • [git-odb] refactor (e07478c)
  • [git-odb] fix docs (05347d4)
  • [git-odb] refactor (721303d)
  • [git-odb] refactor (ea224e9)
  • [git-odb] refactor (6a1b16a)
  • [git-odb] refactor (bae3980)
  • [git-odb] refactor (6b7400b)
  • [git-odb] refactor (19ab0cb)
  • [git-odb] refactor (47c4042)
  • [pack-gen] refactor (b5618ca)
  • (cargo-release) version 0.15.0 (d69d9fb)
  • Put prodash behind a feature toggle, too (966058d)
  • Put 'walkdir' behind a feature flag/make it optional. (1a3cc5b)
  • Put 'sha1' behind a feature toggle (4f326bc)
  • Put crc functionality behind a feature toggle (458fa6e)
  • Revert "[pack-gen] quick hack for obtaining the entry size more quickly" (4c36f92)
  • [pack-gen] quick hack for obtaining the entry size more quickly (ad6d007)
  • Revert "[pack-gen] remove tree-diff as traversal option." (2907a5f)
  • [pack-gen] remove tree-diff as traversal option. (8373671)
  • [pack-gen] fix docs (2548b48)
  • [pack-gen] basic progress for entry generation (953190d)
  • [pack-gen] the first barely working progress (5b89a0e)
  • [pack-gen] tests are green (34b6a2e)
  • [pack-gen] thanks clippy (3f948a4)
  • [pack-gen] the basics to get the program going (03b67b0)
  • [pack-gen] Use more light-weight lookups for all blobs (80ce34d)
  • [pack-gen] refactor (e0caf8d)
  • [pack-gen] a way to get the pack location by ID right away (5619efb)
  • [pack-gen] refactor (fcb9c5f)
  • [pack-gen] refactor (11ce2d8)
  • [pack-gen] And the fix - all green (202c704)
  • [pack-gen] with the addition of write-oack checks it actually fails (a9e46a6)
  • [pack-gen] it compiles and all tests are green, with tests for all parts (b3a0344)
  • [pack-gen] minor but relevant differences between 'existing' and 'existing_object' (5f18124)
  • [pack-gen] very close to a basic impl of count + entries-gen… (c927429)
  • [pack-gen] Fill the relevant information for later (932b439)
  • [pack-gen] the first test for object counts (67b1512)
  • [pack-gen] first sketch of how counting could look like (6ef0072)
  • [pack-gen] prep for counting stage (93fd425)
  • [pack-gen] tag handling for diff based traversal (e55906c)
  • [pack-gen] tag support for tree traversal (28ed260)
  • (cargo-release) version 0.10.0 (5d7ee6a)
  • [pack-gen] the first green test for Tag iterators (df5ef8a)
  • [pack-gen] A test to see we can handle tag objects (1898319)
  • refactor (9f0a8cc)
  • [pack-gen] Finally traversal based pack gen works too (086422b)
  • [pack-gen] diff-based traversal now finds all reachable objects (e819c92)
  • thanks clippy (760febf)
  • [pack-gen] add more objects during diff traversal (bc2ef19)
  • [pack-gen] pickup more trees (2da57bd)
  • [pack-gen] Specific tests show that something is off in the changes case… (b131c9e)
  • [pack-gen] central object synchronization for diff based algo as well (6de3558)
  • [pack-gen] have to keep track of all seen objects (dc645c6)
  • [pack-gen] updating tests to verify something shows that objects are duplicated (cef1a58)
  • [pack-gen] tree diff based pack generation passes the trivial test (ad0e345)
  • [pack-gen] refactor (cac002a)
  • [git-traverse] accept tree iterators instead of object ids (f343dad)
  • [pack-gen] Most of changes based pack gen (9ade04c)
  • [pack-gen] prepare diff based pack-gen (fa2ae2c)
  • [git-diff] refactor (087e853)
  • [git-traverse] refactor (85de287)
  • (cargo-release) version 0.3.0 (684de4b)
  • [pack-gen] Speed up tree-traversal :D (90b4093)
  • thanks clippy (009a342)
  • [pack-gen] Probably a valid impl of tree traversal (4c72a17)
  • [pack-gen] quite a bit closer to tree-traversal for pack gen (ecd37ee)
  • [pack-gen] refactor (325d63e)
  • [pack-gen] a test for upcoming traversal modes (8d1e9ac)
  • [pack-gen] refactor (08b136f)

Thanks Clippy

Clippy helped 5 times to make code idiomatic.

v0.15.0 (2021-05-09)

Commit Statistics

• 18 commits contributed to the release over the course of 6 calendar days.
• 6 days passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Thanks Clippy

Clippy helped 2 times to make code idiomatic.

Commit Details

view details

• Uncategorized
  • (cargo-release) version 0.15.0 (d91b241)
  • (cargo-release) version 0.9.0 (84897fd)
  • Merge branch 'patch-2' (f01dc54)
  • refactor (a9e4feb)
  • Fix formatting (a341995)
  • Use Seek to skip large objects during indexing. (95e2af7)
  • Remove almost all unsafe code from Tree. (42b6033)
  • thanks clippy (17258cc)
  • thanks clippy (09decde)
  • Convenience methods for iterators (aa6c9e6)
  • refactor (d9783b9)
  • A sketch of convenient finding of commits (db21062)
  • refactor (3af7b7b)
  • sketch of convenience function for Find trait. (2bf4958)
  • refactor (bd4d21e)
  • refactor (8b10434)
  • Fix order of operations in git-odb::hash::Write (a31d8c7)
  • (cargo-release) version 0.14.0 (a760f8c)

v0.14.0 (2021-05-02)

Commit Statistics

• 4 commits contributed to the release over the course of 2 calendar days.
• 2 days passed between releases.
• 0 commits where understood as conventional.
• 0 issues like '(#ID)' where seen in commit messages

Commit Details

view details

• Uncategorized
  • (cargo-release) version 0.14.0 (d9514ee)
  • rename 'Locate' to 'Find' - shorter and just as good (60f72f5)
  • (cargo-release) version 0.13.0 (5c791af)
  • [traversal] remove git-odb::traversal (now git-traverse::iter) (35b74d2)

v0.12.0 (2021-04-30)

Other

• :borrowed::Object => git-odb::data::Object
• :Db::init() with a few tests

Commit Statistics

• 153 commits contributed to the release over the course of 18 calendar days.
• 21 days passed between releases.
• 2 commits where understood as conventional.
• 1 unique issue was worked on: #67

Commit Details

view details

• #67
  • The very first version of complete pack writing (4d76d53)
  • A sketch of the pack::generation function signature (21b0aab)
• Uncategorized
  • prepare test utilities for release… (d35e654)
  • (cargo-release) version 0.8.0 (a1ce210)
  • (cargo-release) version 0.3.0 (e9665c7)
  • [traversal] all the caching (0890403)
  • [tree-diff] first prototype of traversal experiment (ece43b4)
  • thanks clippy (2d5e205)
  • [tree-diff] more tests for the tree iterator (91b5a02)
  • [tree-diff] Now the commit graph traversal works with zero-allocations (8078910)
  • make it easy to get a commit iterator (33213f3)
  • [tree-diff] refactor into iterator based model (29b527a)
  • [tree-diff] The least intrusive way to allow dealing with tree iterators (d41dd3c)
  • refactor (a4d5f99)
  • refactor (03ee510)
  • Better ergonomics for accessing decoded objects (ae3eab6)
  • refactor (c1013dd)
  • refactor (f37c31f)
  • refactor (ac70651)
  • refactor (77764f3)
  • refactor (3185cc9)
  • Thanks, cargo audit (4f293f5)
  • refactor (edf7d38)
  • refactor (ca98221)
  • refactor (b4027e3)
  • refacto (6e328da)
  • fix docs (a54bab4)
  • refactor (3f2ee4c)
  • refactor (d6ab581)
  • refactor (d490b65)
  • Pack V2 writing (base objects only) seems to work now #(67) (e68dd84)
  • The first more thorough and indirect validation of the newly written pack… (d43687e)
  • refactor (08fafaa)
  • test newly written pack data alone (01fdd70)
  • Write pack data entries #(67) (722202e)
  • refactor (eed1e3c)
  • Write pack data header #(67) (717726b)
  • refactor (28cbeb3)
  • refactor (4261c7d)
  • All logic needed to write a valid pack within an iterator (775ab29)
  • sketch of pack data write API (dfeda87)
  • refactor (f33fa10)
  • [experiment/object-access] allow bare repositories (401690d)
  • thanks clippy (c86823a)
  • refactor zlib (4587b82)
  • refactor zlib (496e6bb)
  • refactor (3a4469c)
  • First basic pack entry generation (base only) works… (75cb32b)
  • refactor (d4bf8ae)
  • refactor (2d89222)
  • refactor (eb3a8da)
  • Allow calling 'finalize()' on the entries iterator (3c617bc)
  • refactor (b7d0323)
  • Reduce size of data::Object (7aa783a)
  • First pack-to-pack copy and crc32 verification (37619f0)
  • It's possible to get entry data within pack generation (a2a5927)
  • git-odb without cargo warnings due to using the same test twice (8945f95)
  • A way to obtain entry information using previous lookup information (a55d113)
  • refactor (95ab11b)
  • A probably more secure way of accessing pack data (7a01bd8)
  • sketch of pack-entry retrieval api (d1e9248)
  • refactor (08cf90a)
  • Let's be a bit more conservative with this information for now (efef417)
  • Objects know their pack location (73f1c66)
  • Chunks based iteration for pack generation (23c2694)
  • Some notes about how to treat defaults of file versions (cfa5cf6)
  • run git-odb tests in parallel, too; improved threaded error handling (40802fd)
  • the first test for pack generation (2a2fdde)
  • refactor (385f52d)
  • refactor Please enter the commit message for your changes. Lines starting (f65c68c)
  • fix doc links (ec35743)
  • thanks clippy (563e445)
  • The first seemingly working iteration over all objects in an odb #(67) (6b34f62)
  • refactor (01d9d91)
  • impl size_hint for linked db iter (ada259b)
  • refactor (82c2f42)
  • refactor (7a6b514)
  • First sketch of object iterator for linked::Db (a316eed)
  • Set environment in testtools to freeze repositories generation scripts (eaad3ab)
  • faster repeated tests if fixtures don't change (792277f)
  • refactor (e1a92ad)
  • Allow the use of shared test utilities across crates (b117626)
  • refactor (40b86a7)
  • refactor (8b00094)
  • fix doc links (7479071)
  • thanks clippy (6f901f5)
  • ancestor iterator is now generic over Locate trait (bbfd616)
  • [fail] try to abstract ancestor::Iter over Locate trait (f8c0375)
  • refactor (5ef1f22)
  • Improve interface for building packs based on Locate trait #(67) (5b66b6e)
  • A version of the Locate trait we can do today #(67) (d752be2)
  • [git-odb] Associated types with lifetimes also don't seem to work (0e68a9d)
  • [git-odb] Trying to use offical traits won't work with our kind of object (29a5054)
  • :borrowed::Object => git-odb::data::Object (747a13e)
  • An even better name for decode errors (f270850)
  • Make clear it's a decode error we are using there (f45cb4b)
  • rename git-object::(owned->mutable)|(borrowed|immutable) #(67) (91ee558)
  • bump git-odb minor version (5c833ce)
  • thanks clippy (547af6e)
  • Fix test breakage for loose object reading (222c7a2)
  • fix docs #(67) (01db10a)
  • thanks clippy (60a7689)
  • refactor (ef674ff)
  • Remove loose::Object entirely #(67) (5cf4840)
  • Start using loose::Db::locate2 - definitely still bugs in there (d6f07b7)
  • An alternative version of loose::Db::locate() for use with borrowed::Object (5b40a32)
  • refactor (bad3ce4)
  • replace loose::Object::stream() with *::data() #(67) (040b347)
  • sketch loose::Object::data() to start refactoring #(67) (ee1701f)
  • Sketch of trait for locating objects #(67) (31445d7)
  • refactor (2754dd6)
  • refactor (3e908bd)
  • refactor (409d763)
  • refactor (896ab94)
  • Remove unsafe interface for stepped computation #(67) (c856613)
  • Show that with custom iterators, Arc's are natively supported #(67) (0c49007)
  • thanks clippy (405dd9d)
  • multi-tip support #(67) (2254ecc)
  • cache support for traversal #(67) (1e9300a)
  • cycle and duplicate check #(67) (334a72d)
  • a new failing test (86b6c24)
  • refactor (5408b62)
  • The first basic traversal utility #(67) (ea6610b)
  • Fix iteration signature due to shadowed naming (fe8b459)
  • thanks clippy (a463a43)
  • Sketch of machinery for producing pack entries #(67) (ac8e7fb)
  • A step towards using SteppedReduce #(67) (0d5595a)
  • (cargo-release) version 0.13.0 (ac2eddb)
  • Allow parallel reducers to produce something during 'feed()' #(67) (6c04fcd)
  • Allow more fine-grained stepping over the pack generator #(67) (22eb892)
  • Allow to obtain object information without fetching the data #(67) (6553850)
  • sketch a version to abstract object data retrieval #(67) (ad90446)
  • Implement Write trait for linked::Db (21362c3)
  • Docs for linked::Db (9d936de)
  • Support for caches within linked::Db (3635a3e)
  • locate() for linked::Db without cache for now (014bc3c)
  • refactor (7b443d1)
  • refactor (d077ead)
  • :Db::init() with a few tests (4c77e4c)
  • Frame for linked::Db (e64d148)
  • Make cycles in alternate object chains fatal (67e679a)
  • Resolve alternates as paths, not as repositories (73352c3)
  • Remove support for Polonius in preparation for a new repo type (871c803)
  • (cargo-release) version 0.11.0 (fd698e3)
  • Introduce pack_id for use in pack cache, preventing (most collisions) (ad04ad3)
  • Fix benchmark to get valid test results (20abb3a)
  • First use of memory-cap based LRU cache for object access (b057494)
  • Add hash-map based LRU to allow bigger/more effective object caches (5affdd5)
  • Feature toggle for uluru based Lru cache (98eec48)
  • refactor (d624d09)
  • Improve docs to prevent people to 'misuse' the Lru cache. (fff62ed)
  • LruCache with const-generics (93618d1)
  • [experiment] cached version of compound::locate() (ec988dc)

Thanks Clippy

Clippy helped 8 times to make code idiomatic.

v0.10.0 (2021-04-08)

Other

• add link to simplified/polonius version in the docs
• Only check alternates for objects not found in packs or loose This matches the behavior of git.
• Avoid double-lookup in packs without polonius Split object lookup into two steps: looking up the object index, and looking up the object itself given the index. This avoids passing in the buffer (and thus looking like an unconditional borrow to non-polonius) until we're committed to returning from the loop.

Commit Statistics

• 23 commits contributed to the release over the course of 4 calendar days.
• 4 days passed between releases.
• 3 commits where understood as conventional.
• 1 unique issue was worked on: #63

Thanks Clippy

Clippy helped 1 time to make code idiomatic.

Commit Details

view details

• #63
  • Impl == and != for common combinations of ObjectId/oid (2455178)
  • git-protocol uses oid type (3930a6f)
  • Use new oid where possible in git-odb (68a709e)
  • refactor; better errors for invalid hash sizes (be84b36)
  • Make ObjectId/oid happen! (ca78d15)
  • Remove all public exports of git-hash types in git-object (accf89d)
  • Remove re-export of git_object::borrowed::Id (a3f2816)
• Uncategorized
  • (cargo-release) version 0.10.0 (3161777)
  • (cargo-release) version 0.7.0 (b900914)
  • (cargo-release) version 0.12.0 (3b71e7e)
  • (cargo-release) version 0.2.0 (4ec09f4)
  • Greatly reduce compound::Object size (afa8156)
  • The git-odb compound Object clearly is too large (8f0e813)
  • add link to simplified/polonius version in the docs (d53c4b0)
  • Only check alternates for objects not found in packs or loose (b317200)
  • Avoid double-lookup in packs without polonius (eaae9c1)
  • thanks clippy (0c5f404)
  • Remove locate(…) -> Option<Result<…>> in favor of Result<Option<…>> (40ee743)
  • A locate returning Result<Option> for compound DB (a1dafa6)
  • Use Result<Option> in Bundle::locate() (2dfef8f)
  • A trial for Result<Option> for loose object databases (3842859)
  • Assure loose objects are actually not found when opening (7a4f2bf)
  • Add feature toggle for polonius and make it part of the test suite (c825c11)

  v0.9.1 (2021-04-03)

  Commit Statistics

  • 4 commits contributed to the release.
  • 5 days passed between releases.
  • 0 commits where understood as conventional.
  • 1 unique issue was worked on: #59

  Commit Details

  view details
  • #59
    • Fix initializing pack bundles in compound db (5a48e08)
    • Add failing test (d629339)
    • Move pack fixtures into place which resembles an actual object db (fb5cea4)
  • Uncategorized
    • (cargo-release) version 0.9.1 (e0feb28)

  v0.9.0 (2021-03-29)

  Commit Statistics

  • 13 commits contributed to the release over the course of 14 calendar days.
  • 63 days passed between releases.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.9.0 (efc8983)
    • thanks clippy (0fc239c)
    • refactor (f2e9add)
    • upgrade depdendencies (e4a7711)
    • Fix compile warnings produced by +nightly build (e387d2c)
    • Conform imports (fd73731)
    • Fix error type argument order and spell fields out (819568e)
    • [git-odb] Return error on invalid packs (88de64d)
    • [git-odb] Fix Inflate::once (36f6bbd)
    • [git-odb] Remove unnecessary tests (ebe41ca)
    • [gix] Use flate2 by default (f1158a1)
    • [gix] Add optional zlib feature (f1f9665)
    • [git-odb] Add feature flag for zlib-ng (96b3810)

  Thanks Clippy

  Clippy helped 1 time to make code idiomatic.

  v0.8.0 (2021-01-24)

  Commit Statistics

  • 1 commit contributed to the release.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.8.0 (1ccfdcd)

  v0.7.1 (2021-01-24)

  Commit Statistics

  • 8 commits contributed to the release over the course of 38 calendar days.
  • 38 days passed between releases.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.7.1 (2c38ff9)
    • (cargo-release) version 0.11.0 (1aa1f5e)
    • Require latest version of git-features in git-odb (e664e93)
    • Remove usage of gitfeatures::fs in organize subcommand (b567d37)
    • refactor; planning (5df492c)
    • thanks clippy (343ab9a)
    • refactor (5b1328f)
    • Add missing '.' at end of doc comments (7136854)

  Thanks Clippy

  Clippy helped 1 time to make code idiomatic.

  v0.7.0 (2020-12-16)

  Commit Statistics

  • 3 commits contributed to the release.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • fix git-odb tests (35c1209)
    • All crates use git-hash::Kind and its types, sometimes through git-object (124c171)
    • use git-hash in git-features (5b307e0)

  v0.6.0 (2020-12-15)

  Commit Statistics

  • 3 commits contributed to the release.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.6.0 (27f5955)
    • (cargo-release) version 0.9.0 (a89fdb9)
    • (cargo-release) version 0.5.0 (fc7d600)

  v0.5.0 (2020-12-15)

  Commit Statistics

  • 8 commits contributed to the release over the course of 27 calendar days.
  • 27 days passed between releases.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.5.0 (c767e07)
    • more docs for owned git-object (b79101d)
    • thanks clippy (ba9b3c2)
    • refactor (d5d7cf9)
    • more docs of git-object::owned (0620dce)
    • (cargo-release) version 0.8.0 (47c00c2)
    • cargo clippy Rust 1.48 (475a68c)
    • finish refactoring git-odb (ec282ae)

  Thanks Clippy

  Clippy helped 1 time to make code idiomatic.

  v0.4.2 (2020-11-18)

  Other

  • try to get rid of tree-traversal Boxed error… …which really complicates things downstream as these now have to deal with another type argument, or of to try to turn it into a Box anyway.

    The latter seems to be…troubling so I can't make it compile.

  Commit Statistics

  • 79 commits contributed to the release over the course of 55 calendar days.
  • 67 days passed between releases.
  • 1 commit where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Thanks Clippy

  Clippy helped 3 times to make code idiomatic.

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.7.0 (7fa7bae)
    • (cargo-release) version 0.4.2 (173c957)
    • Minor fixes to git-odb docs (3788512)
    • complete docs for git-odb (0cf8496)
    • prefer empty doc strings for modules over [allow missing docs] (9b3f04f)
    • add remaining doc strings for git-odb (428f0ad)
    • Some more docs (2d87124)
    • try to document all the bits an pieces of git-odb (1b353fa)
    • Finish docs on top-level traversal method (2ef1c99)
    • start describing how pack traversal works (5e990f2)
    • refactor (a681335)
    • document pack::index::write (f5edc60)
    • dependency update (bc336d9)
    • refactor (6b909a2)
    • refactor (b511a2b)
    • document index integrity checking (9336ab9)
    • docs for index access (996acbf)
    • docs for top level pack index module (d2dd72f)
    • document pack data verification (27962ca)
    • document pack entry iteration (c869ee9)
    • docs for pack header (9505b40)
    • some more pack data file documentation (05e05f4)
    • docs for Bundle::write_* (ac41253)
    • remove special Error with just one variant… (d05a416)
    • Docs for Bundle::locate (066787c)
    • some more docs for 'pack' module (c32850d)
    • some more documentation (201f67c)
    • Merge branch 'commit-graph' into main (9cb09b2)
    • the daily commit (single handedly) (b528c2e)
    • specify the hash to create with 'hash::bytes_of_file' (c000294)
    • move 'git_odb::hash::bytes_of_file' into git_features::hash (c5f6b45)
    • document loose::Object entirely (d5eef9c)
    • thanks clippy (b9e0a87)
    • docs for Sink (e7a09f0)
    • docs for compound object databases (813df71)
    • Document borrowed odb objects (7626f7f)
    • Document alternates implementation (60666e8)
    • docs for git-odb crate (top-level) (71af366)
    • remove dash in all repository links (98c1360)
    • thanks clippy (e355b4a)
    • refactor (5a1cbf2)
    • And octal values unquoting works too (5effc7b)
    • All explicit escapes (1841544)
    • First bunch of simple unescapes (a45c594)
    • prepare for actual unescaping (284da44)
    • basic infrastructure for unquoting c-style strings (f81bb03)
    • fix incorrect cycle detection, which worked on MacOS by accident (a6e7765)
    • Also use alternates for looking up objects… (ade929d)
    • prepare for unquoting c-strings (47e2fa0)
    • refactor (8877b77)
    • Read multiple alternates from single file; ignore comments (1f8d367)
    • support for relateive alternates (b20e9ee)
    • refactor (4a0d034)
    • Ignore all cycles and be happy if we have found at least one actual odb (1effdfd)
    • refactor (485aa91)
    • prepare for multi-line parsing and all the bells and whistles (08f9ec4)
    • refactor (c1d2f41)
    • Make compound DB initialization less lazy… (6dc57b3)
    • Revert "FAIL: try to get rid of tree-traversal Boxed error…" (1b42b31)
    • refactor (07aff14)
    • Use parallel walkdir (via jwalk) when parallel feature is enabled (f444c85)
    • try to get rid of tree-traversal Boxed error… (13159eb)
    • refactor (57d463f)
    • alternate now handles cycles (71167e4)
    • Parameterize traversal error with Processor error (1513a13)
    • refactor (c6be43d)
    • first simple alternate tests (7372118)
    • Switch to prodash 10 and safe a lot of trait bounds in the process (e2fb1d9)
    • refactor (524d0fe)
    • test for circular alternates (fc92709)
    • refactor (a8f4cd7)
    • Checksum verification for compound object (3be08b0)
    • thanks clippy (4ddc64f)
    • refactor (59d989a)
    • Actually resolve alternates when creating a compound DB (9be7aed)
    • More methods for compound object (84d2b0e)
    • refactor (c1eff58)
    • refactor (e5a9343)
    • first sketch of alternate resolution (6cc8a94)

  v0.4.1 (2020-09-18)

  Other

  • See if tree compaction saves considerable amounts of memory No, it's not worth it.

  Commit Statistics

  • 17 commits contributed to the release over the course of 36 calendar days.
  • 36 days passed between releases.
  • 1 commit where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.4.1 (60ac8b0)
    • refactor (ad17bfd)
    • (cargo-release) version 0.6.0 (9ef184e)
    • refactor (91d9f78)
    • refactor (6ebb5d1)
    • [clone] All it took was a an intermediary to call 'read' as expected (7c8ecb7)
    • [clone] minor refactor; it's definitely the read() that doesn't work… (406829b)
    • [clone] This actually works: first MVP of retrieving packs via clone (c06d819)
    • [clone] support for progress that can handle writing pack files (46e0055)
    • Use fast-reset for miniz oxide to gain about 4s when resolving the kernel pack (e5b6ce4)
    • fix build (6178133)
    • refactor (174baa7)
    • bump git-features to 0.4 to allow publishes after breaking changes (9d6b879)
    • Allow dual-licensing with Apache 2.0 (ea353eb)
    • refactor (b4a6e16)
    • remove tree compaction code (dfc6c7d)
    • See if tree compaction saves considerable amounts of memory (0092c25)

  v0.4.0 (2020-09-12)

  Commit Statistics

  • 32 commits contributed to the release.
  • 0 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • (cargo-release) version 0.4.0 (2272fa4)
    • (cargo-release) version 0.4.0 (0d7b60e)
    • remove quickerror dependency from git-odb (7e27495)
    • refactor (7874c35)
    • refactor (6a84f13)
    • refactor (3ec99dc)
    • refactor (4e89c3b)
    • refactor (519dd12)
    • Document why we won't use nightly for fixing NLL issue (ca29368)
    • refacator (7ac2153)
    • (cargo-release) version 0.5.0 (82b7313)
    • Revert "Fix NLL issue by using nightly" (6864a55)
    • refactor (d4f288c)
    • Fix NLL issue by using nightly (8c5bd09)
    • refactor (3a8fb61)
    • Update tasks, prepare for NLL fix (52af8d1)
    • Thanks clippy (6c4d1ec)
    • This works, but locates twice… (4e709f6)
    • refactor (98b3f4a)
    • Also the imperative version doesn't borrowcheck… (c5720f1)
    • Looks like the functional approach to locate(…) doesn't borrowcheck (5df6867)
    • refactor (127b8b2)
    • refactor (9e68c6b)
    • refactor (669b726)
    • refactor (f219d5a)
    • refactor (7bc321e)
    • sketch compound::Db::locate; sort packs by size (6609a53)
    • refactor (4a09754)
    • refactor (0752b45)
    • Implement Write in terms of writing to the loose object DB (02b88c2)
    • First sketch of compound Db (9bf2279)
    • refactor (203ba99)

  v0.3.0 (2020-08-12)

  Refactor

  • Use borrowed::Id in trees for full type safety

  Other

  • try to use a customized version of just pieces of Miniz-oxide
  • fanout table, but slowly I get it :D
  • discard idea of making traversal even more generic

  Commit Statistics

  • 413 commits contributed to the release over the course of 31 calendar days.
  • 31 days passed between releases.
  • 4 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • bump minor version to 0.3 (4351e28)
    • thanks clippy (6725104)
    • Also run file hashing in indexed mode in parallel (like with lookup) (8f8d14f)
    • first step towards parallelizing file hashes and traversal! (9573836)
    • allow hashing to be interrupted (df4dfd7)
    • unify file based file verification of data and index (e1b4105)
    • update to quick-error 2.0 (4b1b784)
    • Haha, funny, silly me… (a4a1244)
    • better usability for units (b226253)
    • better progress for Sha1 of pack and index (310a59e)
    • Make obvious that interrupt request was received (34b2373)
    • Conditionally use an eager iterator… (e9b5511)
    • Reduce progress information for threads (e9a1b31)
    • Revert "A test to see how much time can be saved by not zeroing zlib buffers" (3d51d59)
    • A test to see how much time can be saved by not zeroing zlib buffers (fd41a51)
    • Implement optionally keeping the compressed bytes (fc26914)
    • first step towards more control over allocation in iterator (cacf76c)
    • never keep decompressed bytes while streaming… (65c3856)
    • Only keep base objects, not the deltas (like originally intended) (fc8334b)
    • reduce footprint of sha1 when writing the index (12aa454)
    • first successful test of moving the streaming iterator into its own thread (c9fcb68)
    • first sketch of order-destroying eager iterator (20fca45)
    • add object size progress when resolving with index (b2f8c9e)
    • add decompression progress (0e5c534)
    • Print read throughput automatically (0a71b48)
    • Allow 'read' progress to go out of scope while keeping it accessible! (d7a7828)
    • Fix throughput display of otherwise stepped progress indicators (399f81d)
    • unify used ranges for line renderer amond pretty and lean interface (f59f66e)
    • Add percentage and throughput to tasks that matter (763d7ca)
    • Upgrade to latest iteration of prodash (3a4faec)
    • First part of migration to prodash 8.0, but… (6901a09)
    • Fix various issues related to 64bit offset support when writing indices… (da31694)
    • fix unit tests: actually sort the directory entries :D (b69717a)
    • Add convenience method to get a new bundle for the index/data just written (a6d74ad)
    • bundle write with a given directory (7f29c73)
    • first unit test for bundle writing (74bda39)
    • journey tests for restore functionality (1aa63e4)
    • refactor (fc42567)
    • refactor (cf3ebe0)
    • refactor (72ca435)
    • more flexible error types for processors - anything goes (be3a947)
    • refactor (c7dd581)
    • refactor (aae8e79)
    • refactor (0e27763)
    • Make lookup based algorithm gracefully interruptible (8d2e649)
    • Write about user interfaces and the use/non-use of async (91ba045)
    • Use pack hash for index file as well :D (2106c64)
    • support for interruptible operations (a025593)
    • thanks clippy (62d2ff3)
    • organize object type comparisons by probability… (19a5d94)
    • count object types as well (e04a8d1)
    • Revert "Less memory for look up mode, faster start" - too slow (584350a)
    • Less memory for look up mode, faster start (395c7e7)
    • remove petgraph entirely (70ba33a)
    • compute statistics for indexed pack verify (3d31c23)
    • prepare for computing indexed statistics (082c246)
    • refactor (bfbae90)
    • keep all metadata per object needed to compute the usual statistics (961b85e)
    • make 'level' available to support statistics (f7ba51c)
    • refactor (6277318)
    • support for error handling in traversal callbacks (c1d5bf6)
    • indexed traversal now works, in theory, but needs error handling (86f8400)
    • support for progress (62108fd)
    • support for thread local storage in callbacks (1dad088)
    • support for learning about the objects slice in the pack (faec782)
    • And even more caapbilities are required to make tree traversal work natively (90523bb)
    • refactor (2bbfd82)
    • refactor (efa7cd8)
    • first steps towards actually using the new tree traversal during verification (785b0ff)
    • thanks clippy (44b20de)
    • refactor (afe5e44)
    • refactor (fcc660d)
    • reduce memory usage for index considerably (aa802be)
    • and now it works! (f14e10e)
    • use new traversal in index writing, but it doesn't work yet (0dd5570)
    • refactor (4ff69c6)
    • refactor (6cbb7cc)
    • generalized tree traversal can theoretically work (64158e0)
    • make traversal part of the tree for greater ease of use (6629e30)
    • prepare flexible traversal on decompressed objects (7707ea6)
    • refactor (deea36c)
    • refactor (83a0102)
    • refactor (b77d148)
    • generalize tree iteration (fdc06de)
    • Also index using the new tree impl during verify (prepare replacement) (92039b0)
    • refactor (e3ff6af)
    • support for building a tree from offsets (95858bc)
    • refactor (8cfe025)
    • refactor (bb9e518)
    • count sorting into the progress, 7.5 mio entries takes a moment (2fc4cd8)
    • Use bigger buffers when reading from disk. (e76e4eb)
    • Only keep decompressed bytes of base objects… (b39ad89)
    • remove memory mode entirely (and some complexity with it) (8812e91)
    • turns out you never want to keep deltas in memory (657aa2c)
    • Remove support for keeping compressed memory to reduce the index size (1e2ec7e)
    • don't cause re-allocs of the compression buffer (2bb6fd2)
    • Revert "FAIL: try to use a customized version of just pieces of Miniz-oxide" (ea0fdb3)
    • try to use a customized version of just pieces of Miniz-oxide (9945eba)
    • dependency upgrade + update (c6692c6)
    • refactor (133e3ba)
    • Let go of another handbreak - decompression is much faster now (ae9dc16)
    • thanks clippy (393067d)
    • Use call to produce the resolver, allowing to delay opening a file mapping… (dd30e8d)
    • Fix actual memory violation (thanks to unsafe code) (c44c5e1)
    • thanks clippy (1083a0b)
    • Reduce memory consumption (6d1a7a1)
    • Unfortunately working with an optional for data is unwieldy, let's use default (12bbca0)
    • Tree can now be used as sole data structure, collecting all results (3e52d6f)
    • preparation for allowing reuse of the tree data structure (f565512)
    • refactor (9c4bc0a)
    • And it works! The new algorithm is sleeker, and really wants to be backported… (8e025b1)
    • thanks, clippy… (079ce9c)
    • Basis for re-implementing core algorithm using new Tree data structure (be6caf4)
    • refactor (290c29a)
    • incorporate proper filtering of bases (0880998)
    • overhauled iterator logic, still missing 'is_root' filter (2bfbae1)
    • First impl of the Iterator shows it's 'unknown' what a root node is (3f32938)
    • sketch on how children access could look like (16a35df)
    • How a referenced version would look like… (e36021d)
    • refactor (62a01fe)
    • more experimentation towards a safe tree data structure… (d907ce8)
    • first stab at new Tree datastructure… (85d7579)
    • Safety for handling base pack offsets (17d8375)
    • …but there seem to be issues with the kernel pack… (cc147bc)
    • quick and dirty impl of gitoxide layer for bundle writing, aka index-pack (e78386b)
    • cargo clippy (586ba7a)
    • implement in-memory mode; refactor (0c195b9)
    • refactor (c9d9298)
    • Use monomorphic calls only at the expense of code siz (40b28d1)
    • refactor (150d0bc)
    • Also implement the 'no output directory' branch (5a3240f)
    • refactor (68e52f8)
    • For the first time, writing an index could work with persistence (16e045c)
    • Don't write pack to file if everything is kept in memory (f3ddda6)
    • Allow data file to be optional in preparation for in-memory operation (95af105)
    • refactor (413968d)
    • refactor (5d27cdb)
    • optional pack lookup depending on the settings (2b509de)
    • Write-through the pack file as we receive it and move it into place (6180e39)
    • receive progress information when reading packs in bundle (759091d)
    • start supporting writing packs to disk right away (f2203e0)
    • refactor (75c333c)
    • prepare for implementing the bundle with various write modes (de420e4)
    • bundle thread progress underneath reducer progress (76b1b2b)
    • prevent deadlock, interestingly (ca02901)
    • refactor (ea254c0)
    • rough progress for writing the index (f1a7f9b)
    • initial batch of progress usage for index creation… (b10e5c6)
    • refactor (77b3c21)
    • refactor (fb23d15)
    • refactor (7da7e08)
    • refactor (5a3ad3a)
    • refactor (785a23d)
    • header encoding works now! As well as index writing :)! (024b854)
    • initial version of a complete header encoding impl, but… (ce6b46b)
    • looks like CRCs are not correct (3c4e4a0)
    • cargo clippy (a5596fb)
    • Fanout writing works now… (93a7ba9)
    • It's a good idea to remove old code from time to time… (9e47f1b)
    • fanout table, but slowly I get it :D (cfd8a25)
    • Fix decompression; fanout table is still wrong though (77fac1a)
    • Despite writing the CRC32 now, it doesn't work yet (ecd12b9)
    • first stab at streaming pack header encoding (3c6e78b)
    • refactor (5925d46)
    • Simplify offset handling in favor of allocating less (ce4ec62)
    • Only allocate memory for offsets if needed (72e0642)
    • First complete implementation of index writing… (826f996)
    • Reduce contention by using the shared cache only once (c370e13)
    • Optimize CRC handling - no need to assign it after the fact (ffcc03d)
    • Assure we can deltas store theyr resolved buffer (d2a81d9)
    • And it does seem to work! Awesome! (71cd982)
    • Delta-application could work if we handle our buffer better (ac6100b)
    • refactor (400a2a9)
    • One step before applying deltas (a074193)
    • prepare for delta application (9a9fb7a)
    • cargo clippy (d69c973)
    • parse pack header before trying to decompress :D (9d1b44a)
    • refactor (772e9ce)
    • consumer can resolve entries (13adce6)
    • refactor (c87f770)
    • refactor (d9d406d)
    • first version of resolver to copy from a memory map (506b8fd)
    • rethink resolver into something even simpler (4388c6c)
    • use parking_lot where possible (367874e)
    • Consumers can fail gracefully (9082080)
    • refactor (1b4cad0)
    • refactor (4ce13bb)
    • support for decompression in case compressed bytes are stored (c1fcf28)
    • computing hashes for bases from decompressed in-memory store works (7c19fe6)
    • show that all data can be passed for processing in threads (a95ce9c)
    • A cache usable from threads (1d4879a)
    • re-associate CRC32 with the correctly sorted ID output (037e1e5)
    • refactor (b3a365d)
    • refactor (97eb524)
    • Use chunked input and calculate 'optimal' chunk and thread sizes (0cc74d7)
    • generalize chunk iterator (905e85e)
    • first rough cut of in_parallel invocation (8f16081)
    • prepare for parallelization (cb36596)
    • Simplify indexing step (070899c)
    • resolver look ups may now turn out empty… (a991923)
    • Allow us to stop searching for bases early when resolving (e7874da)
    • This should be the interface for building indices from packs directly (f5295d0)
    • Got a good idea on how this will work! (7bb229f)
    • keep track of the pack trailer (cdba61e)
    • Now I understand why there is a separate resolution phase… (1c2bcbd)
    • Fix tests (b9866b6)
    • prepare a way to gradually implement V2 index writing (92a4986)
    • refactor (feba75b)
    • We can now restore (possibly half-written) packs (b1daa46)
    • prepare ability to restore pack files (76583e5)
    • Support for pack trailer verification when iterating (f37f131)
    • Also read the pack trailer during iteration (98a8e17)
    • Only take as many objects as we are allowed (without 'take(…)') (86f5853)
    • refactor (e15bde4)
    • Shift thin pack resolution to another work bucket; test for index writing (2592361)
    • refactor; better tests (12d14bf)
    • refactor (bd66a85)
    • Now keeping track of read bytes works (d32d921)
    • An attempt to intercept bytes read from bufread - FAIL (8db04f6)
    • refactor (2d817d7)
    • refactor (893f65b)
    • refactor (12816bc)
    • refactor (56f763a)
    • Associate HashKind with the kind of pack (d66d139)
    • Move all pack-related file handling to bundle; big refactor (f8b6e75)
    • first step towards putting the index file into position (d994c74)
    • initial interface trial for writing pack index files (936bdcc)
    • refactor; more thorough tests (82d87ce)
    • cargo clippy (b768b56)
    • At least make it configurable if to keep decompressed bytes or not (28ebcae)
    • And streaming iteration works, even though we are forced to allocate… (27d624d)
    • Yes, this really cannot work: StreamingIterator (b4df430)
    • In the moment we tried to actually return Entry<'a>, it didn't let me :D (8367955)
    • First steps towards making the InflateReader reusable (83a97d4)
    • Better error handling in iterator, fuse yourself (5ebacc4)
    • The next() impl shows that we should be less lenient (4521cab)
    • Provide entries which borrow from iterator (86eea13)
    • Provide a lifetime for iterator (and possibly its entries) (7852bd1)
    • first version of expected iterated data types (d5e7d31)
    • improved iterator constructors (fb71f04)
    • better handling of pack headers (0030bdb)
    • frame for a pack iterator (07d1096)
    • some more tests (9095728)
    • verification for pack objects (17bd95e)
    • refactor (3ee947e)
    • 'stream()' now assures all data is decompressed (32e994c)
    • it looks like something is wrong with the object stream implementation (d187b5a)
    • Loose object verifycation - but it doesn't seem to work as expected (9dd5676)
    • refactor (37cfd9b)
    • refactor (8e3b9fc)
    • prepare full 'verify' implementation (ee45c7f)
    • refactor (0a33b24)
    • Always compress values when using a sink when exploding packs (70562fa)
    • support for compression even when using sink (105c845)
    • Another stab at fixing stress tests :) (7db6a33)
    • fix stress test; improve progress messages (37ccd92)
    • Ignore decode errors (if configured) at the right spot (e53141d)
    • tests for relaxed error handling (93c0e26)
    • Nice error message on failure (adbc82c)
    • inform about deleted files using progress (a3ee516)
    • Fix error display - certainly something to watch out for (38eff2c)
    • The first 'explode' implementation… (0d31ad1)
    • Support for skipping various safety checks during traversal (0416666)
    • prepare for configuration of safety checks (06638d0)
    • cargo clippy (95e02c9)
    • Restore original verification functionality (0e3c1b9)
    • nearly there! Interesting that anyhow errors must be sync! (eaee77e)
    • finally it compiles with returning Boxed errors! Ouch… (1fc8252)
    • First sketch of new verify expressed in terms of traversal (4cb570f)
    • refactor (f2832a8)
    • Finally a progress can be passed to the delegate… (a9f4de0)
    • refactor (bbb3e1e)
    • Pass all arguments (but progress) to processor (1e87922)
    • Call a bare version of the traversal processor (95a5cea)
    • preparation for abstracting the 'process object (stateful)' function (fe400f5)
    • discard idea of making traversal even more generic (1525f36)
    • Initial step towards separating verification from traversal (d14b4fc)
    • refactor (bae7781)
    • rename verify-pack to pack-verify (keeping it more formal) (ec8c48a)
    • refactor (f580441)
    • Fast implementation for buffered input (c50b150)
    • Respect object size to be 64 bit where applicable… (61c8aba)
    • better errors for writing disk objects (f7bc137)
    • Try to use HashKind where possible (b32e01d)
    • refactor (a3777ed)
    • clippy happy (a938c70)
    • And writing of loose objects works (bbfe7bf)
    • This seems to be a working deflate write implementation (0acce38)
    • The first succesful inflate run for small input (94e1c5a)
    • what seems to be a reasonable write implementation for deflate (45a28d2)
    • Another test to understand the deflate streamer better (4256038)
    • refactor (dd463df)
    • refactor (0b42237)
    • refactor (5b0bb84)
    • Put down a few tests to understand how deflate wants to be fed (178a018)
    • refactor (0d8d7fe)
    • Improve looks of documentation (11a32eb)
    • Fix tests for now… (79ab945)
    • refactor (0cd7bb7)
    • Complete and unoptimized disk writer for objects, but… (9d0c3f1)
    • refactor (62e75bc)
    • Make use of HashKind in Write trait (0304dd0)
    • Make our Sink API similar to std::io::sink() (a03ae0f)
    • Finish Sink implementation (84f7908)
    • first steps towards serialization tests for sink (e8d52c6)
    • Introduce hash kind, as this should be specified when writing an object (f5d0acf)
    • A simple trait for writing owned objects and streams (68b7d7d)
    • (cargo-release) version 0.2.0 (76fe0ab)
    • (cargo-release) version 0.2.0 (0bb8314)
    • Use 'optimized' chunk size for 'less-time' algorithm (c8c23c0)
    • incorporate dynamic chunking into 'less-time' algorithm (295aa2f)
    • integrate new chunk size code into lookup code (a8422cf)
    • Simplify progress code using inc() (9e8df59)
    • Add 'inc()' convenience methods to progress (2e46c9b)
    • Run clippy first; pacify clippy (0a5b883)
    • use faster algorithm by default (bb45c3d)
    • Properly compute delta chain length by default (a93b894)
    • remove hits_to_live (3a3fae9)
    • attempt to auto-remove unusable deltas… (5dd8243)
    • Now with cache (and due to that, incorrect statistics for now) (efd28d2)
    • Make chunk statistics independent of traversal method (6225f36)
    • First working version of alternate object traversal, without cache (51b5eb6)
    • initial state for indexed lookup (acbcd79)
    • refactor; tests now fail with more than just not-implemented (310a2f7)
    • speedup entry sorting a little; use less memory (b4df372)
    • better index entries sorting progress (b4d7038)
    • prepare sharing even more code (61c76cf)
    • Make use of shared reducer in upcoming indexed verify implementation (290eae1)
    • Use shared reduce implementation in lookup based algorithm (10fc88d)
    • prepare for integration of general reducer (c37832e)
    • refactor; enable testing of reverse-delta lookup (512daf9)
    • Revert "Move deallocation off into own thread" - not worth it! (051da15)
    • Move deallocation off into own thread (90230f1)
    • Implement more cache-friendly pack offset v2 retrieval (00cf84b)
    • refactor (3c25c67)
    • initial refactor of DeltaTree, but… (6384649)
    • measuring performance of sorting index offsets is quite revealing (4b16336)
    • Properly handle the BufReader to make indexing work; FAST (57e95cf)
    • Avoid seek in favor of skimming a file read in bursts (01ae405)
    • Some performance information in progress (20aef2c)
    • Nodes now provide access to the pack offset (61c1497)
    • Basic tree access for the entry graph (c5e5c77)
    • Fix clippy (ec40e09)
    • hookup new indexing step (313064f)
    • frame for running the new streaming code on bigger packs (e0b34eb)
    • refactor (fdfab40)
    • refactor (1fbeb35)
    • refactor (385e935)
    • Now it works :D (008b4de)
    • Initial (failing) implementation of building an index tree (25dc83d)
    • Easy access to sorted offsets in pack index files (d93540f)
    • refactor (cb8d561)
    • refactor (c7ae705)
    • refactor (2fc449c)
    • Change course and do pack streaming first (bcb275e)
    • roundtrip Rust repo in stress test; accept more diverse trees when parsing (0347cdb)
    • Allow some very special trees not to be round-trippable (8fe1358)
    • Consume PGP signature in tags fully (ffd6c31)
    • make tagger signature optional (3358f9a)
    • remove now unused pgp_signature field - it's in extra-headers (c8c937c)
    • proper support for extra-headers (d0feb2b)
    • Switch to latest quick-error (9760856)
    • Fully implement --encode and --re-encode flags (a7cfac8)
    • refactor (56b66ac)
    • prepare for re-encoding each pack object (afae684)
    • fix build with rustc 1.45 (8c2a1ee)
    • refactor (ec5e50f)
    • prepare for writing out owned trees (2b6eced)
    • Use borrowed::Id in trees for full type safety (5d57c1f)
    • refactor (f7b8826)
    • fix odb test (a792f44)
    • Prepare for allowing an owned, processed version of multi-line headers (f966e7f)
    • Use borrowed::Id everywhere (9f876f0)
    • move git_object::Id into git_object::owned::Id - much better already! (50c7136)
    • basic integration of borrowed Id; translate between owned and borrowed (84ff638)
    • prepare to allow Id be owned and borrwed; abstract over hash type (d883c31)
    • introduce the notion of IdRef (7007361)
    • Use statically known borrowed arrays for perfect type safety! (3ead048)
    • refactor (766f3e4)
    • refactor (bca1f16)
    • 'data -> 'a as it's shorter and also more idiomatic (71821e9)
    • refactor (dedd4dc)
    • refactor (de0bc3c)
    • refactor (e5391d3)
    • refactor (163909b)
    • refactor (49f64db)
    • refactor (9f825b8)
    • refactor (2fbc2e1)
    • fix naming change, which was introduced accidentally (fbb9f98)
    • make it easier to validate bundles, for completeness (8ea05de)
    • refactor (34e85f2)
    • refactor (b3bde87)
    • refactor (0b540c2)
    • refactor (2888f1b)
    • refactor (0817b24)
    • refactor (dcacd3b)
    • refactor (b113da9)
    • refactor (6659174)
    • refactor (bed5dc8)
    • refactor (4867740)
    • refactor (f6cc80e)
    • refactor (8b416d4)
    • refactor (23e05d7)
    • refactor (d3b36f4)
    • More tests for various object types (f4703e0)
    • refactor (86fa00f)
    • Basic decode implementation (7ff02cb)
    • Support for in-pack object lookup in Bundle::locate (7e3d6be)
    • First dummy implementation of borrowing a buffer provided by the user (9c31fcb)
    • Make it easy to learn that objects couldn't be located by using options (a916f36)
    • mild refactor - need combined pack + index (6bf8ed4)
    • Respect thread limit in 'in_parallel' (babfd84)
    • apply cargo diet (79b9b73)

  Thanks Clippy

  Clippy helped 5 times to make code idiomatic.

  v0.1.0 (2020-07-12)

  Refactor

  • a simpler implementation to skip the header

  Other

  • first silly attempt to randomly remove an allocation
  • get rid of failure crate in favor of quick-error

  Commit Statistics

  • 167 commits contributed to the release over the course of 90 calendar days.
  • 3 commits where understood as conventional.
  • 0 issues like '(#ID)' where seen in commit messages

  Commit Details

  view details
  • Uncategorized
    • add missing license description (2b80181)
    • Make crates publishable (5688a34)
    • cargo clippy (from CI) (0a28857)
    • Proper implementation of line renderer into 'lean' CLI (e98e7c2)
    • Handle windows newlines in test suite for packs as well. (ebd5176)
    • Add metadata to allow docs.rs build all featueres (10f9386)
    • update tasks (269280a)
    • Allow to limit the logging depth for less cluttered output (fce7035)
    • Looks like this performs much better already, but… ideally subprogress isn't shown (3b96d18)
    • finally speed up logging progress properly - needs input throttling (1a550c6)
    • provide average throughput per second (5b23d17)
    • git-odb with serde support (0da930c)
    • Remove dependency to git-object from git-features - it better remains free (67c3a6a)
    • commit to using bstr whenever something is not data bytes; remove miniserde (3183d1b)
    • Prepare centralization of bstr as optional component (aa857d9)
    • #[forbid(unsafe)] for all crates (afda803)
    • Allow for more screen space when formatting (6794300)
    • prepare next task (74bcbb5)
    • display object throughput per second, even though it won't be visible in TUI… (53b4513)
    • disable LRU cache if we have to get statistics (befba3b)
    • wonderful statistics on compression efficiency! (1bb09c5)
    • count objects per chain level (209d53f)
    • pass average stats through to the top level (5b4979c)
    • refactor (4dd9fd4)
    • closer to actually producing statistics (5f087ec)
    • refactor (7add82c)
    • Also average statistics on chunk level (3b927e5)
    • Provide more detailed information when decoding an entry (80c5da9)
    • no need to say 'begin' before doing something, it's primarily for logging (13eba3a)
    • throughput for pack (81f5c33)
    • print performance stats at the end of hashing the index (9c94417)
    • assure hashing progress is dropped when done (db6e067)
    • First implementation of logging per thread (477dd90)
    • Support for providing progress to threads (2815858)
    • properly count objects (d398e7e)
    • first very basic progress implementation (b820717)
    • Pass progress everywhere, for now just to discard it (da3ae1c)
    • Control which hashing crates to use from the top-level as well. (dfe9b20)
    • Use git-features to toggle 'parallel' mode from the 'gitoxide' level (d944fbf)
    • sketch out Progress trait; don't forget to write docs at some point (534b3c7)
    • refactor (baeb4ef)
    • refactor (e12bfd6)
    • Make in_parallel trait bound more loose: Clone instead of copy (3e91b05)
    • Using all cores actually does speed things up (ed944b9)
    • Also run index+pack validation in parallel; only parallelize bigger packs (dc15b26)
    • avoid running anything in parallel for small packs (c2df183)
    • Don't send every single entry, instead send reasonably sized chunks (56298a6)
    • refactor (down to 6 minutes for big pack verification) (4157b51)
    • first working version of actually parallel in_parallel (145ee39)
    • first implementation of 'parallel' without threads. How will scoped fare? (735744e)
    • A sketch of a minimal helper for parallel work (377252a)
    • refactor (be4795f)
    • refactor (3e2efff)
    • bigger LRU caches are better, but with this one we can't go too large (5e1f7ae)
    • First implementation of an LRU cache - it gets hit, let's see how it fares! (5a21031)
    • also set the cache with bases and deltas (915a3fb)
    • first sketch of cache implementation - get() is there, next is put() (ce54756)
    • Allow delta base resolution to fail (similar to how lookups can fail) (b721424)
    • Allow in-pack lookups for V1 packs (2e51bbb)
    • Add CRC32 reading at index (268f855)
    • Pack offset by index (69e35b1)
    • V2 pack lookup (9e56902)
    • test V1 lookup (e9c7127)
    • Add CRC32 check during pack verification (04ff1a0)
    • prepare for CRC32 check - needs understanding of size of bytes in packed object (3ab2df1)
    • refactor (dd2d623)
    • Finally delta-objects can be read as expected. (81f2f54)
    • definitely an improvement to the way add-deltas are applied… (c6cdb12)
    • Fix one issue with Trees being declared as tags (ada66cd)
    • validate sha1 of pack objects, some work, some don't for some reason… (aa8799a)
    • Capability to write loose object headers, fast (de0aeff)
    • refactor (5364bbe)
    • Fix another implicit assumption that doesn't hold: deltas are NOT… (093637d)
    • Finish delta-application to take into account the biggest possible result… (0ee2b69)
    • first stab at dealing with bigger-than-expected intermediate result sizes… (8027ff4)
    • First simple implementation of fetching all objects in a pack (without validation) (053045b)
    • Support for verifying pack files and index files (b09b4e1)
    • simple index file verification (internal) (1d27050)
    • refactor (4023b02)
    • refactor (855a769)
    • refact[r (c84410b)
    • refactor (c24c79d)
    • test --no-default-features for git-odb (2394bca)
    • refactor; prevent trailing bytes to become part of the digets (043813c)
    • try a version that doesn't rely on memory mapped files for throughput… (d59ddfc)
    • try to speed it up with prefetching - not really :D (8485185)
    • simplify folder names (36fde1f)
    • Fix LSB parsing code with python based code written 6 years ago :D (c12fdad)
    • improved packed header parsing… it works a little better now it seems, but… (ca779ed)
    • refactor; and figured out what the header parsing issue is (d364049)
    • some more tests (85e541f)
    • refactor; better error handling (031df11)
    • first very rough version of full-object decompression without allocation (7c704a7)
    • refactor (dcb1997)
    • refactor (baaf06e)
    • refactor (3edaaec)
    • Finish Object Reader implementation, now for in-memory objects, too (35e69b8)
    • a simpler implementation to skip the header (47ca6ab)
    • Allow skipping the header when decompressing files (streaming) (ff35032)
    • first step towards supporting skipping the header in the stream (8e45f53)
    • Fix stream decoding - it seems to work, but we need to deal with the header (f10ed75)
    • tests for streamed reading of bigger objects (FAIL) (b4a6b72)
    • refactor (80aad4b)
    • Add missing parts to implement Read, need refactoring to make it work though (13d4cdb)
    • First step towards streaming of ZLIB deflated content (a870f7a)
    • cleanup (a2f0a5d)
    • fix clippy (a9c5da7)
    • Make decompression of bigger objects work (on the fly) (7e4f5a9)
    • It becomes obvious that this way of decompressing things won't work (1818bda)
    • Don't do so much logic if we already decompressed everything (26cb36c)
    • refactor (423b885)
    • more convenient access to our four object types (ecda6d2)
    • It's proably OK to make parsed pack entries avaialble, why not (8a64e10)
    • refactor (13f0e77)
    • Memory size checks for objects (ab51616)
    • Reduce loose Object memory footprint (38a81b0)
    • first Blob test for blobs that are already in memory (f503324)
    • Make single-field objects blob and tree more explicit (1aef68f)
    • add Blob type to parsed objects (d3e8e4b)
    • See 'parsed' blobs as in-memory representations… (6a6e105)
    • Make clear that not all objects can be parsed at the expense of convenience (ce3031d)
    • don't conflate errors with 'there is no suitable object' to parse (b9b796f)
    • fix imports (10f2967)
    • try pub use with rename. Not bad in the docs, but maybe a bit confusing (526f3f8)
    • refactor (b9a1647)
    • Integrate Commit object into Loose DB (7e9fe50)
    • test for parsing trees from loose dbs (4f48249)
    • refactor (9f9ccad)
    • refactor (427c480)
    • refactor loose db (6ea4f53)
    • handle commits without newlines; make tag newlines optional (c0b54be)
    • Make Commit available in borrowed object (b2d1b5d)
    • avoid unnecessary allocation when creating SHA1 paths in loose ODB (09d8d3a)
    • first silly attempt to randomly remove an allocation (4ff2168)
    • document existing use of unsafe, deny everywhere else (41f4bce)
    • cleanup integer parsing in loose object database (ecdce1a)
    • the defining property is actually that the object is borrowing data (e0125fd)
    • fix cargo fmt (642dd13)
    • cleanup; all tests work! (7c96603)
    • first version of tag message parsing - it's actually changed now (74b2328)
    • remove itertools in favor of vendoring the little code we need (8340508)
    • optimize macro usage (0c9960b)
    • optimize dependencies (3ea2195)
    • Use git-object in git-odb (07f7c31)
    • Add the latest nom, hoping it will be come out of alpha… (85958f1)
    • refactor; use pretty-assertions for massively more readable test-failures (ea8d311)
    • Switch everything parsed to BStr (62ae90a)
    • refactor (9a86f63)
    • Use btoi to parse all integers, directly from ascii-bytes (4f6ef42)
    • refactor (2990902)
    • move parsing tests close to actual parsing (3ca2c59)
    • move examples into demos, having their very own dependencies; optimize tests (b757712)
    • fix (untested) extraction of delta object information (55a56b7)
    • parallelize git-conut, optimize for speed (debd044)
    • refactor (9fc9fc0)
    • Fix big-pack 64 bit offset handling in index v2 (3b485b5)
    • make refactor (cd6a18a)
    • cargo clippy first pass (8b0a2a8)
    • Finally remove failure and equip example with anyhow (f5e4ec5)
    • remove failure from Index (55034a7)
    • And one more module without failure (d0575bf)
    • A big step towards removing failure (d862bd8)
    • refactor (87c8a2e)
    • get rid of failure crate in favor of quick-error (91c8fc1)
    • Get rid of nightly requirement, just parse tags differently soon (f037c4d)
    • cargo fmt (2aa0857)
    • reorganize repository a bit; use different contact email address (cb9fa28)