50 ancestors of cd5b6e94edc57964
|
2026-07-17
| ||
| 15:58 | Create new branch named "next-verbs" check-in: cd5b6e94ed user: dmestas tags: next-verbs | |
| 15:20 |
Merge recover-inflight: root-URL sync serving + the log one-verb surface
Two independent slices recovered from an interrupted session, each gated green. - Sync now dispatches on method + Content-Type at any URL path (§4 L759-764), so a stock 'fossil clone http://host:port/' at the bare root interoperates; the /xfer-only requirement and its 404 are gone. - 'zit log' is reduced to the signed one-verb surface: one check-in ID, help only, no -n windowing; usage errors exit 2, a bad ID exits 1. Unit + e2e (209/209) green; fmt clean; dst 6x1000 ok. check-in: dbb2735bff user: dmestas tags: trunk | |
| 15:17 |
Reduce 'zit log' to the signed one-verb surface (drop -n windowing)
'zit log' carried a -n K windowed-log flag: an extra grammar mode (a state machine with a positional-only '--' separator and a -n value token in either order) and a bad_count offense, for a bound no other query verb exposes. Its sibling 'update' already documents the target shape — one full check-in ID, help only, no windowing. Collapse log onto that same one-positional grammar. parseLogArgs is now the plain scan parseUpdateArgs uses: -h/--help wins; -n, --all, and a bare '--' are all unknown options (log has no positional-only separator); a second positional is an extra argument. Drop LogArgs.count, the bad_count variant, parseLogCount, and walkLog's max_count / window early-exit — the ancestry walk is unconditionally unbounded again. Usage errors (missing check-in, unknown option, extra argument) now exit 2 via cli.usageError; a well-formed-but-non-canonical ID stays a validation failure at exit 1. Help and 'zit help log' print the trimmed single-verb text. Net -110 lines. Unit + e2e (209/209) green; fmt clean; dst 6x1000 ok. closed check-in: 0973cf7e9e user: dmestas tags: recover-inflight | |
| 15:15 |
Dispatch sync on method + Content-Type at any URL path, not just /xfer
A conformant Fossil client POSTs sync to the repository URL path unchanged and MUST NOT append /xfer (sync §4 L759-764): a bare 'fossil clone http://host:port/' POSTs to "/". The server previously required the /xfer path and answered 404 for anything else, so a root-URL clone from a stock client failed. Drop the path from the dispatch decision entirely. classifyXfer now keys only on method (POST) and a request-valid, outer-compressed Content-Type (application/x-fossil); the URL path is accepted but unused. Remove xfer_path, requestPath, and the not_found/404 disposition; Disposition collapses to ok / method_not_allowed / unsupported_media_type. A non-POST stays 405 rather than a home redirect this sync-only server can't honor (and a 302 to "/" would loop now that "/" is a valid endpoint). The over-real-HTTP e2e clone now targets the bare root "/" to prove a root-URL client interoperates. Unit + e2e green; fmt clean. check-in: 9636157a85 user: dmestas tags: recover-inflight | |
| 13:46 | Merge clone-w4: skip the receive-side inflate for full-form clone content, cutting clone peak RSS ~13% (651 -> 568 MB). Reconciles held()/content() across the pull .raw basis and push upload paths so a corrupt wire-backed container degrades gracefully instead of tripping an unreachable. 1538 tests pass; fossil interop smoke green. check-in: 19af0f2de7 user: dmestas tags: trunk | |
| 13:43 |
Skip the receive-side inflate for full-form clone content
The clone/sync receive path inflated every stored cfile container to prove the RFC1950 stream was intact (§3.9), then threw the result away: the store keeps the container verbatim and reconstructs lazily, and the persist commit hook runs the §7.7 integrity check on the reconstructed bytes. For the full form that inflate was pure waste. Move the decode into the delta branch (which genuinely needs the plaintext to reconstruct) so a full-form cfile is interned and retained without inflating. Retaining an unvalidated container means a held artifact is no longer guaranteed reconstructable: content() can be null for a corrupt wire-backed container, so held() no longer implies content(). Reconcile the consumers that assumed it did. The pull .raw delta basis and the push upload path both degrade gracefully (round abort / silent skip) instead of tripping an unreachable or assert on hostile input; the read/derivation path keeps the decompression-bomb cap, and the eager clone-end image treats a corrupt container as an error. Cuts clone peak RSS ~13% (651 -> 568 MB) on the fossil corpus by dropping the discarded plaintext allocation. Instruction count and wall are flat, since full-form cfiles are the minority of clone content. 1538 tests pass; fossil interop smoke green. closed check-in: d91a3cfe42 user: dmestas tags: clone-w4 | |
| 12:00 | Merge clone-w3: halve clone compute (-53% instructions, 2.3x faster wall) via a bounded persist-time plaintext memo, byte-identical store + trunk hash preserved. check-in: fbfef2b649 user: dmestas tags: trunk | |
| 11:58 | Halve the clone compute by memoising the persist-time plaintext: a bounded PlaintextCache reuses each frame's section 7.7-verify plaintext as the basis for the reverse-deltas built on it, killing the delta-chain reconstruction read-amplification (persist no longer re-inflates + re-applies a basis chain per dependent). libfossil clone: 138.2B -> 64.7B instructions (-53%), wall 13.4s -> 5.7s, RSS 550 -> 371 MB, trunk hash identical, store byte-identical (proven by test). The target's section 7.7 hash is still verified unconditionally after every reconstruction, so a stale basis is caught, never returned. closed check-in: 1bcbee0f86 user: dmestas tags: clone-w3 | |
| 11:30 | Create new branch named "clone-w3" check-in: 57438f967c user: dmestas tags: clone-w3 | |
| 03:44 | Merge windowed-log: 'zit log -n K' bounds the ancestry walk to K check-ins (early-exit before descending), fixing zit's slowest read for the common case. check-in: b2d89cb634 user: dmestas tags: trunk | |
| 03:42 | Add a windowed 'zit log -n K' that early-exits the ancestry walk after K check-ins: walkLog returns before descending into the K-th node's parents, so only K manifests are reconstructed instead of the whole DAG (full log → -n 20 scales linearly, the signature of a true bound). Opt-in — unbounded default unchanged; -n K accepted in either order; 0/non-numeric/overflow rejected. Documented as the K nearest ancestors in DFS order (not fossil's mtime timeline). closed check-in: ed93685e89 user: dmestas tags: windowed-log | |
| 03:24 | Create new branch named "windowed-log" check-in: f0e2cd8720 user: dmestas tags: windowed-log | |
| 02:55 | Merge eager-image: warm the derivation cache at clone-end so the first read after a clone is the zero-parse image path (cold first-read leaves 26s → 0.03s), best-effort and clone-only. check-in: bf8b63f32c user: dmestas tags: trunk | |
| 02:52 | Warm the derivation cache at clone-end so the first read isn't a cold derive: runClone now calls Repository.warmCache after the final log append, which runs open's own miss path (derive → serialize the zero-parse image → store .zit-cache) so the first leaves/branch/name-resolution read takes the image fast path. Measured on a real clone: cold first-read leaves 26.0s → 0.03s, byte-identical. Best-effort — a warm failure leaves the clone correct with the old lazy first-read, never fails the clone (repostate §7.3 rebuild fallback). closed check-in: 85d8080ec8 user: dmestas tags: eager-image | |
| 02:27 | Create new branch named "eager-image" check-in: 5b2ba7331c user: dmestas tags: eager-image | |
| 01:57 | Merge zerocache: materialized-image warm reads (84x faster leaves load) with load-time-validated fallback, closing the read-path perf gap. Phase B (full SoA image) not warranted — the image load is under 1% of the derivation it replaced. check-in: 13cfa27494 user: dmestas tags: trunk | |
| 01:54 | Serve warm reads from a materialized column image instead of re-deriving RepoState, cutting a warm leaves load 178.6us to 2.12us (84x). At cache write, dense integer-addressed columns for the mtime index, leaf image, and symbolic-name table are appended to the versioned .zit-cache; on a fast_read open the freed queries (mtimeOf/isLeaf/isOpenLeaf/leaves/branchOf/resolveName per repostate 7.1/7.4/7.5/8.2) answer from those columns with zero derivation. Load-time span validation range-checks every column, CSR span, and stored id and returns a miss (fall back to derive) on any violation or CRC/layout/log-tail mismatch, so a corrupt or stale image is never a wrong answer. RepoState's public API is unchanged; only leaves opts in, cold verbs keep the derive path. Technique: on-disk zero-parse caching after Richard Feldman's Roc and Andrew Kelley's Zig compiler (common in game engines per Casey Muratori); typed enum(u32) index families after the compact_arena crate. closed check-in: ae242f5563 user: dmestas tags: zerocache | |
| 00:54 | Create new branch named "zerocache" check-in: 6542332151 user: dmestas tags: zerocache | |
| 00:38 | Merge warmprobe: land the warm-read timing bench that gates read-path perf work on measured re-derivation share (98% of warm load on althttpd). check-in: 8ccef5b6bd user: dmestas tags: trunk | |
| 00:36 | Add a warm-read timing bench (zig build bench) that splits a cache-warm read into cache-body I/O, RepoState re-derivation, and formatting, so read-path perf work is gated on measured phase shares. On the althttpd store re-derivation is 98% of warm load. closed check-in: b93d0f9dd6 user: dmestas tags: warmprobe | |
| 00:23 | Create new branch named "warmprobe" check-in: a9a7b2cf4d user: dmestas tags: warmprobe | |
| 00:22 | Merge cli onto trunk: land the zit.Repository facade and the ls, cat, commit --branch/--close, revert, and sync-with-persisted-remote verbs (CLI milestones M0-M4), each built plan-review-commit against the artifact, repostate, and sync RFCs. check-in: 96784cbe6f user: dmestas tags: trunk | |
| 00:19 | Merge trunk into cli: integrate the T-card sign-ordering fix, the F-card read-lenient import shim, and the streaming-clone memory fix ahead of landing the CLI verbs. Unify the duplicated sign->byte helper onto TSign.byte() (dropping the parallel signByte free function) and keep both branches' §4.7.16 ordering tests. closed check-in: 41aeb64521 user: dmestas tags: cli | |
|
2026-07-16
| ||
| 23:59 | Add a bidirectional sync verb and a persisted default remote so pull, push, and sync can omit the URL and resolve it from a new local config field per sync §4/§6 check-in: b5a9919379 user: dmestas tags: cli | |
| 23:06 | Add commit --branch/--close to start or close a branch by emitting the repostate §6.2 branch and sym tags while cancelling inherited parent syms, plus a revert verb restoring managed files to their artifact §8.4 effective check-in blobs check-in: 9f08c51525 user: dmestas tags: cli | |
| 22:18 | Merge fcard-import: read-lenient import path lets zit parse fossil's legacy trailing-space F cards (interop with pre-2009 history) while the strict parser, writer, and RFC stay pure and artifact identity is unchanged. check-in: f98cb9177c user: dmestas tags: trunk | |
| 22:16 | Add a read-lenient import path so zit can read fossil's own pre-2009 history: parseForImport tolerates the single legacy trailing space on fossil F cards (a cvs2fossil quirk, non-conforming per artifact-format 4.2) by stripping one dangling separator per F card ONLY after the strict parse rejects and the original Z self-checksum validates, re-sealing Z on an ephemeral copy that never touches storage, serve, or identity. The strict parser, the writer, and the RFC stay unchanged; artifacts keep their original hash. closed check-in: cca7dc06a2 user: dmestas tags: fcard-import | |
| 21:39 | Add a zit cat verb emitting one file's uncompressed content at a resolved check-in through the Repository facade, per artifact-format §8.4 single-name lookup and §7.2 hash-verified blob reconstruction check-in: d0ef5067bb user: dmestas tags: cli | |
| 21:38 | Create new branch named "fcard-import" check-in: d3be8fe4ac user: dmestas tags: fcard-import | |
| 21:10 | Merge stream-clone: cut clone peak memory 14.2 GB -> 291 MB (fossil parity) by storing received frames wire-backed instead of buffering the whole repo's reconstructed plaintext, with section 7.7 verification moved onto the log-basis reconstruction path (a stronger round-trip parity guard). check-in: 77e2541742 user: dmestas tags: trunk | |
| 21:08 | Stop buffering the whole repository's reconstructed plaintext in RAM during clone: store received frames wire-backed and verify each durable frame's section 7.7 hash against its log-basis reconstruction (sync 7.2/7.4), cutting a sqlite clone's peak memory from 14.2 GB to 291 MB with a byte-identical store and a matching trunk hash. closed check-in: 9a18bdd63d user: dmestas tags: stream-clone | |
| 21:06 | Add a zit ls verb listing a check-in's effective file set through the Repository facade, per artifact-format §8.2 file enumeration check-in: 6ee0c0e855 user: dmestas tags: cli | |
| 20:39 | Merge tcard-order: fix T-card run ordering to key on the full sign-included name token (artifact-format 4.7.16), clearing 159 wrongly-rejected branch-switch check-ins that broke zit log on deep histories. check-in: b4f2ae0ceb user: dmestas tags: trunk | |
| 20:37 | Order T-card runs by the full sign-included name token per artifact-format 4.7.16, not the sign-stripped name, so branch-switch check-ins with interleaved propagate/cancel tags (T *sym-trunk / T -sym-clear-title) parse instead of being wrongly rejected as out-of-order. closed check-in: a2fa7629c6 user: dmestas tags: tcard-order | |
| 20:27 | Fold the store layers and RepoView behind one zit.Repository facade so read verbs open a repo, resolve names, and read content through a single deep interface check-in: 513397e07a user: dmestas tags: cli | |
| 20:15 | Create new branch named "tcard-order" check-in: 83a1201013 user: dmestas tags: tcard-order | |
| 20:06 | Create new branch named "stream-clone" check-in: 14f733376d user: dmestas tags: stream-clone | |
| 19:14 | Merge trunk onto cli: bring the branch tip up to the verbatim-serve, delta-cfile, repostate cache, container codec, and storage-benchmark work before starting the CLI facade milestones. check-in: a81d55773e user: dmestas tags: cli | |
| 16:49 | Make zit pack monotonic so it never inflates a store: preserve each source frame verbatim (re-linking delta containers to their base's new rid) and re-deltify a full-form frame only when the result is strictly smaller, fixing a regression that discarded wire deltas and grew a fresh clone 8.65x. check-in: eb9a1de034 user: dmestas tags: trunk | |
| 16:40 | Make pack monotonic: preserve stored wire deltas verbatim instead of reconstructing every frame to full form, which inflated a delta-bearing store many times over closed check-in: 50a4b5d049 user: dmestas tags: packfix | |
| 14:54 | Add a storage measurement mode to the bench harness (zig build bench -- storage): report bytes at rest by frame kind and artifact type, per-frame overhead, full-form compression headroom, and serve CPU, so storage-efficiency work is chosen on measured data rather than guesses. check-in: a48a0010ef user: dmestas tags: trunk | |
| 14:51 | Add a storage bench mode reporting bytes at rest by frame-kind and artifact-type plus serve CPU, so the storage-efficiency ideas can be weighed against measured numbers before any claim. closed check-in: e782228b2d user: dmestas tags: sbench | |
| 14:25 | Add a zit pack command that repacks the store into a compact new-generation log: whole-frame-verify the source read-only, re-ingest every artifact in offset order re-deltaing full-form frames against lineage bases, then atomically swap via a fsynced temp — a crash before the rename leaves the original byte-identical, and it doubles as an integrity scrub. check-in: de49e67c9f user: dmestas tags: trunk | |
| 14:16 | Add a zit pack verb that repacks the store into a compact new generation: re-ingest every artifact in offset order through a parent-manifest and same-path-blob lineage hint so full-form frames re-delta against their basis, recompute each reconstruction's hash against its name before writing it (repo-state §7.7), and swap the rewritten log in with an atomic rename so a crash before the swap leaves the original store byte-identical. closed check-in: 43dccab01c user: dmestas tags: pack | |
| 13:38 | Store cloned artifacts as compact delta frames: retain fossil's reverse-delta containers through the dangling queue and persist each basis before the delta that points at it, so a clone from a fossil origin keeps the server's deltas instead of re-inflating every artifact to full form — cutting on-disk size to roughly 0.6x of the full-form store, matching fossil's own delta count. check-in: 9853f8b93e user: dmestas tags: trunk | |
| 13:28 | Keep fossil's reverse-delta containers through the dangling queue and persist each basis frame before the delta that points at it, so a clone from a fossil origin stores compact delta frames instead of re-inflating every artifact to full form. closed check-in: 0ebc7991f0 user: dmestas tags: storage | |
| 13:26 | Merge the sync-protocol re-sync onto trunk: ratify the container-capacity and clone-igot clarifications that match the reference implementation and zit's own behavior. check-in: 61fde8dec1 user: dmestas tags: trunk | |
| 13:23 | Re-sync the sync-protocol draft: the container length field is a receiver capacity, so a decoded stream shorter than the advertised value must not be rejected (4.1), and igot is never applicable during clone rounds (8.2 / Table 2) — both aligning the text with the reference implementation. closed check-in: d708b4a2f2 user: dmestas tags: rfc-resync | |
| 11:26 | Stop the clone client re-announcing every held artifact with igot each round: a clone enables no push direction (sync 8.2/8.5) so the server ignores those cards, and dropping them cuts a 50k-artifact clone request from ~88 MB to under 2 KB while convergence still rides the server's RID cursor. check-in: 289e073d9c user: dmestas tags: trunk | |
| 11:16 | Stop the clone client re-announcing every held artifact with igot each round: a clone enables no push direction (RFC 8.2/8.5) so the server ignores those cards, and dropping them cuts a 50k-artifact clone request from ~88 MB to under 2 KB. closed check-in: ebc8401895 user: dmestas tags: igot | |
| 10:33 | Wire zit_storage and zit_logstore into the fuzz executable so the seeded fuzz suite compiles again: its sim helpers and logstore targets import those modules, which every other build artifact already provides but the fuzz exe was missing. check-in: 490977ea08 user: dmestas tags: trunk | |