sift/vendor/modernc.org/sqlite/CHANGELOG.md

Changelog

Entries for v1.38.1 through v1.44.1 and for v1.49.1 were added on 2026-09-05, reconstructed from the git history and the merge requests they cite; they were missing at release time.

• 2026-09-15 v1.59.0:

  • Bump the pinned modernc.org/libc to v1.75.7 and re-vendor lib/ and vec/ from modernc.org/libsqlite3 v1.14.5 and modernc.org/libsqlite_vec v0.5.0, which were transpiled against it. As always, downstream go.mod files must pin the same modernc.org/libc version as this repository's go.mod; see the package documentation and GitLab issue #177. The transpiled SQLite is unchanged: still 3.53.4, byte-identical to v1.58.0 on all 20 targets. The change is beneath it. On the Linux targets libc v1.75.7 replaces the transpiled musl memcpy, memmove, memset, memcmp, strcspn and fabs, loops moving at most four bytes per step, with native Go routines backed by the runtime's vectorized memmove and memclr and by the bytes package; on every target strlen is now a word-at-a-time scan, or bytes.IndexByte on the 64-bit architectures. On the three CPU-bound workloads in the new Performance section of the package documentation, measured on linux/amd64 against the same SQLite 3.53.4 compiled from C with the same options, this driver went from 3.0x, 2.2x and 1.6x the CPU time of the C build to 2.0x, 1.9x and 1.3x. The other operating systems already used a native Go memcpy and memmove and see no such change. vec/ stays at sqlite-vec v0.1.9; its re-vendor, from a transpile made with newer modernc.org/cc and modernc.org/ccgo, differs only in unreferenced macro constants and in one bounds check spelled with INT8_MAX instead of its value, so nothing changes in behavior there either.
  • Hand user-defined function and aggregate callbacks a pooled *FunctionContext instead of allocating a fresh one per call. After the []driver.Value pooling of #226 this was the last driver-side heap allocation per invocation: one 16-byte object for every Scalar, Step, WindowInverse, WindowValue and Final call. The context now also carries the invocation's sqlite3_context, so accessor methods can be added to it later without touching the trampolines. Like the argument slice, it is valid only for the duration of the callback and must not be retained past its return; the documentation on FunctionContext and on the callbacks now says so. On the 1000-row, 3-argument noop scalar UDF benchmark this removes a further 1000 allocs/op (5756 to 4756, and 3756 to 2756 with VolatileArgs) and 16 KB/op; on the reporter's reproducer from #226 it removes about 12% of the remaining allocations (25.3M to 22.4M allocs/op, 553 MB to 505 MB per iteration).
  • Updates GitLab issue #226. See GitLab merge request #137.
  • Add regression tests for the pooled context: two functions evaluated in one statement must receive two different sqlite3_context values and every callback's context must belong to the invoking connection, checked on two connections held at once and without dereferencing the context, so that a stale one is reported rather than faulted on; eight connections held at once calling a function from their own goroutines, and a function that runs a statement invoking another function, both meant for the race detector. See GitLab merge request #138, thanks Ian Chechin!
  • Add a Performance section to the package documentation. It records the measured CPU-time ratios of this driver against the same SQLite compiled from C, that they hold under concurrency, where the gap comes from, and the two consequences for applications: index the columns that ORDER BY, GROUP BY and WHERE use, because a query that is merely sluggish in C can cross a deadline here, and bound the database/sql pool with SetMaxOpenConns, because a periodic query that outlasts its period piles up connections without bound. Prompted by a report that attributed a 103 s query to the driver spilling a temporary b-tree in Go; the workload showed no spill and about 2x the CPU cost of C, the wall time being unbounded overlap of periodic polls.

• 2026-09-01 v1.58.0:

  • Upgrade to SQLite 3.53.4. Upstream's own fix for the journal-rollback data-corruption bug is part of this release, so the local super-journal patch v1.56.0 introduced — and promised to drop once upstream shipped theirs — is dropped; recovery behavior is unchanged. This also bumps the pinned modernc.org/libc to v1.75.6; as always, downstream modules must pin the exact modernc.org/libc version this module's go.mod pins (see GitLab issue #177).
  • Add opt-in support for Linux Open File Description (OFD) locks on database files. A POSIX record lock is owned by the (process, inode) pair, so the kernel drops every lock the process holds on a database file whenever any descriptor of that file is closed: an os.Open/Close for a hash, a backup check or a metadata probe anywhere in the process — third-party libraries included — silently strips SQLite's transaction locks and leaves the file unprotected against other processes. With OFD locking enabled, the locks belong to the open file description that placed them and survive such a close. Off by default, and staying off until the mode has real-world mileage: without opting in, locking behavior is byte-for-byte that of previous releases. Enable it by setting MODERNC_SQLITE_OFD_LOCK=1 in the environment the process starts with (any value but the empty string or one starting with 0; read once, at library initialization), or from Go with the new OFDLocking(true), which overrides the variable and must run before the first connection is opened; OFDLockingEnabled reports the mode in effect. The switch is deliberately process-wide rather than a DSN parameter: POSIX and OFD locks taken by one process are different owners to the kernel and genuinely conflict, so every connection to a database file inside one process must use the same kind, and a per-DSN knob would advertise a granularity the kernel does not offer (see the discussion in #255). Because the two kinds do not release one another, the mode is frozen at the process's first lock attempt: later attempts to change it return the new ErrOFDLockingTooLate, while querying, and setting the value already in effect, keep working. On kernels older than 3.15, and on filesystems that reject OFD locks, the first lock attempt falls back to POSIX locks for good and OFDLocking returns the new ErrOFDLockingUnavailable from then on, which is also how OFDLockingEnabled turning false reports the fallback; the same error is returned on every platform but Linux, where the API exists but OFD locks do not. Two boundaries to note: the immunity covers the locks on the database file itself, while WAL's -shm coordination stays on POSIX locks; and code in the same process that takes fcntl record locks of its own on a database file — which used to never conflict with SQLite's, while quietly destroying them — now conflicts with them loudly instead. The C side — F_OFD_SETLK routing through a designated per-inode locking descriptor that preserves upstream's unixInodeInfo semantics (last-unlocker release, PENDING piggybacking, unix-excl), guarded to __linux__ — ships in the transpiled sources via libsqlite3!3 and its follow-up hardening, with the OFD lock constants from libc!33; the review rounds, the /proc/locks measurements behind the design, and the Tcl lock/WAL gate that runs both modes are recorded in GitLab issue #255 and those merge requests.
  • Add the accompanying regression tests. The four OFD scenarios contributed in merge request #136 — lock survival across a rogue os.Close, the interleaved-readers lock lifecycle, the read-only-first designated-descriptor migration, and the failed-first-lock cleanup — run in a re-executed child process with MODERNC_SQLITE_OFD_LOCK=1, since the mode is process-wide and frozen at the first lock so the suite's own process cannot switch; a positive control asserts the variable really switched the mode on before each scenario runs, and the two scenarios whose invariants hold under both modes also run in the parent process's inherited mode. TestOFDLockingSetter exercises the Go call path end to end in a clean child — off by default, enabled before the first connection, the kernel-visible OFDLCK record in /proc/locks as the positive control, frozen after the first lock — and on every other platform the switch is asserted to report itself unavailable.
  • Resolves GitLab issue #255. See GitLab merge request #136, thanks Nathan Herring (@technosloth), and thanks Gani Georgiev (@ganigeorgiev) for pressing the opt-in default!

• 2026-08-19 v1.57.0:

  • Add an opt-in _defensive DSN query parameter that turns on SQLite's defensive mode for the connection, disabling the SQL-level features that let ordinary statements deliberately corrupt the database file. When _defensive=1 (or any strconv.ParseBool true value) is supplied, the driver calls sqlite3_db_config with SQLITE_DBCONFIG_DEFENSIVE immediately after sqlite3_open_v2 and before every other parameter is applied, so the PRAGMAs the driver itself runs, the _pragma list, and every statement the caller prepares are all subject to it. On such a connection PRAGMA writable_schema=ON, PRAGMA journal_mode=OFF and PRAGMA schema_version=N become silent no-ops, and writes to a virtual table's shadow tables (fts5's _data, _idx and so on) and to sqlite_dbpage fail with "table ... may not be modified"; reading those tables, ordinary use of the virtual tables that own them, and VACUUM are unaffected. The flag has no PRAGMA equivalent, so sqlite3_db_config — and therefore a DSN parameter — is the only way to reach it short of dropping to modernc.org/sqlite/lib. The value is parsed before sqlite3_open_v2, so an invalid one fails the connection without creating the database file, and the parameter must appear at most once: a repeated _defensive is an error rather than letting the first value silently win. Absence of the parameter, or _defensive=0, leaves SQLite's default behavior unchanged; existing DSNs continue to work byte-for-byte. Two limits are worth stating plainly, since the name invites more confidence than the flag earns. Defensive mode is a hardening measure, not a sandbox for hostile database files: it is one of several steps SQLite recommends for that purpose, and this build compiles with neither SQLITE_TRUSTED_SCHEMA=0 nor SQLITE_DQS=0 and exposes no authorizer. And it is a property of the connection, not of the database file — a second handle opened on the same file without the parameter is unrestricted.
  • Reject the one DSN combination defensive mode would otherwise swallow in silence. _defensive=1 together with _journal_mode=OFF (or _journal=OFF) now fails the connection instead of opening one in which neither parameter was honoured: SQLite turns PRAGMA journal_mode=OFF into a no-op that still reports success, so the driver would have accepted the mode, executed it, and left the journal untouched without telling anyone. The check runs in the validation phase introduced in v1.55.0, before any statement executes, so a rejected DSN cannot leave the database half-configured. _pragma remains the exception it has always been: _pragma=journal_mode(OFF) alongside _defensive=1 still runs and is still silently ignored by SQLite. Only DSNs using _defensive can be affected, and that parameter is new, so no DSN that opened before changes behavior.
  • See GitHub pull request #6, thanks wsman!
  • Ship the sqlite-vec license notice this module has been missing. modernc.org/sqlite/vec has bundled the transpiled sqlite-vec sources since v1.47.0, but the module carried only its own BSD-3-Clause LICENSE and the public-domain SQLite notice. sqlite-vec is Copyright (c) 2024 Alex Garcia, dual-licensed Apache-2.0 OR MIT and used here under MIT, whose terms require the copyright and permission notice to accompany substantial portions of the software — which 2.8 MB of transpiled vec/ plainly is. The notice now ships as LICENSE-SQLITE_VEC in the module root, byte-identical to the LICENSE-MIT in the upstream v0.1.9 archive and named after the file modernc.org/libsqlite_vec extracts it into. Attribution was never absent — vec's package documentation has named the extension, pinned the version and linked upstream — but the license text itself was, and the omission was ours: vendor_libs/main.go copied the per-target transpiles and nothing else. It now copies the notice alongside them and fails the vendoring run if it cannot, so a make vendor can no longer quietly drop it. The vec package documentation gained a License section recording that the package is under a different license from the rest of this module.
  • The SQLite notice is renamed from SQLITE-LICENSE to LICENSE-SQLITE; update any direct links to it. Its contents are unchanged and SQLite remains public domain. The name now matches both the new LICENSE-SQLITE_VEC beside it and the LICENSE-<upstream> convention every other modernc.org repository follows, but it is more than cosmetic: go mod vendor selects the files it copies into a downstream vendor/ tree by matching each name against a fixed list of prefixes — LICENSE among them — so a name merely ending in LICENSE was never propagated. Both bundled notices now travel with the code into vendored builds, which is where the MIT terms on vec/ keep applying. No code changes; no behavior changes.
  • Let a caller-constructed Driver register its own functions, collations and virtual table modules. Driver has always held four categories of registration state, but only RegisterConnectionHook could put anything on a constructed one: functions and collations were reachable through the package-level API alone, and modules through the package-level driver only, which left the modules field written and read through that instance and so process-global state wearing a per-instance field. Driver now has RegisterFunction, RegisterScalarFunction, RegisterDeterministicScalarFunction, RegisterCollationUtf8 and RegisterModule, plus Must* variants of the first four, each registering on that Driver alone; the methods are safe to call concurrently, and the zero Driver is ready to use as-is. vtab.RegisterModule also honours its db argument now: a non-nil db registers on the driver backing it when that driver implements the new vtab.ModuleRegisterer, while a nil db keeps targeting the driver this package registers as sqlite. One existing pattern changes behavior, narrowly and loudly: vtab.RegisterModule(db, ...) where db was opened on a caller-constructed Driver used to discard the db argument and land on the sqlite driver, reaching every connection in the process; it now lands on the constructed driver alone, so a sql.Open("sqlite") connection that used to resolve such a module gets no such module instead. The same pattern is also the one way an existing program could hold one module name on both a constructed Driver and the package-level one: there the first of the two registrations used to win and the second was refused as already registered, whereas now the package-level implementation wins on the constructed Driver's connections regardless of the order they ran in. Reaching that case at all means the program ignored an error the older version returned. Two smaller deviations round out the list: Driver.RegisterModule reports no error for such a collision, and vtab.RegisterModule now validates its name and module arguments before the not-implemented check, so a call with an empty name that returned vtab: RegisterModule not wired into engine outside this driver returns vtab: module name must be non-empty instead. Everything else is additive against v1.56.0: the package-level registration functions target the same driver they always did, connections still receive every module registered through the package-level path whichever Driver opened them, and a db opened on the sqlite driver resolves to that same driver. The isolating change discussed in GitLab issue #254 is deliberately not made here.
  • See GitLab merge request #135, thanks Ian Chechin!
  • Promote freebsd/386, freebsd/arm and netbsd/amd64 to fully supported platforms. All three are now listed in the package documentation's platform table, which had carried seventeen entries while this module shipped, cross-built and tested twenty. They arrived as experimental in v1.53.0 — netbsd/amd64 reviving a port that had been broken for years, freebsd/386 replacing a stale, effectively untested SQLite 3.41 transpile, and freebsd/arm entirely new — and were deliberately kept out of that table until they had accumulated real-world exposure, with promotion promised once "a period of broader real-world testing … elapses without surprises". That period has elapsed: all three have been in the builder test matrix and in make build_all_targets since v1.53.0, all three pass the full test suite on this release's commit alongside the seventeen platforms already listed, and no open issue reports a defect in any of them. The two netbsd/amd64 build failures filed before the revival, GitLab issue #202 and GitLab issue #234, no longer reproduce at this commit: Xsqlite3_is_interrupted is present in the sources that target selects, and the mu.enter/mu.leave symbols that broke the build are gone. Documentation only — the transpiled sources under lib/ are byte-for-byte what v1.56.0 shipped, and nothing about how these targets behave changes.

• 2026-08-03 v1.56.0:

  • Re-vendor the transpiled SQLite sources, picking up modernc.org/libsqlite3's fix for an upstream data-corruption bug in SQLite 3.53.3's journal rollback. The SQLite version is unchanged at 3.53.3; what changes is that the amalgamation is now patched before it is transpiled. 3.53.3 reworked readSuperJournal() to return the super-journal name through a char** out-parameter, and pager_playback() now tests that pointer where it used to test zSuper[0]. A crash during the commit of a multi-database (ATTACH) transaction can leave the super-journal name and its checksum zeroed while the name length and the trailing magic survive; the checksum is a plain byte sum, so an all-zero name still validates and readSuperJournal() hands back a non-NULL pointer to an empty string. pager_playback() then calls sqlite3OsAccess(pVfs, "", SQLITE_ACCESS_EXISTS), gets ENOENT, and deletes the hot journal without playing it back — leaving the database corrupted. This is not a transpilation artifact: a plain gcc build of the stock 3.53.3 amalgamation fails on the same bytes while 3.53.2 recovers them, and it is what has been making upstream's own test/crash.test fail intermittently, in roughly 2% of runs, on every platform. The patch restores the pre-3.53.3 behaviour of reporting a (nul) super-journal name and will be dropped once upstream ships its own fix. Every supported target carries it.
  • Two targets change beyond that patch. On linux/s390x the regenerated transpile allocates C bit-fields MSB-first, as the big-endian platform ABI requires, rather than LSB-first; this comes from modernc.org/cc/v4 v4.29.1 and touches bit-field accesses throughout the SQLite core, s390x being this module's only big-endian target. On linux/riscv64 the transpile was regenerated on a host running GCC 11.4.0 where the previous one used GCC 13.3.0, which drops a handful of unexported compiler-predefined macro constants (the __FLT16_* family, __DBL_IS_IEC_60559__ and friends) and changes the COMPILER=gcc-13.3.0 entry PRAGMA compile_options reports to COMPILER=gcc-11.4.0; no SQLite code generation differs. Every other target's generated code is byte-identical to v1.55.0 apart from the journal-rollback patch above.
  • Bump the pinned modernc.org/libc to v1.74.4, and the remaining dependencies to their current releases. v1.74.2 and v1.74.3 are retracted upstream — a freeaddrinfo lock leak that deadlocks name resolution — and v1.74.4 is the fix. As always, downstream modules must pin the exact modernc.org/libc version this module's go.mod pins (see GitLab issue #177).
  • Documentation sweep. openbsd/amd64 and openbsd/arm64 join the supported platforms table in the package documentation: both have been in the builder test matrix since January and are cross-built by make build_all_targets, but had never been listed. The vfs DSN query parameter — which names a VFS registered with SQLite, such as one returned by vfs.New — is now documented alongside the other DSN parameters on Driver.Open. The "Debug and development versions" section no longer describes a GO_GENERATE environment variable and a go generate that this repository has not had since generator.go moved to modernc.org/libsqlite3; it now points at that repository and make vendor instead, and the stale //go:generate directive naming the removed file is dropped with it. modernc.org/sqlite/vec and modernc.org/sqlite/vfs gained the package doc comments they were missing, so both finally carry a synopsis on pkg.go.dev. Documentation only; no behavior changes.
  • Add NewConnector, returning a database/sql/driver.Connector for use with sql.OpenDB. It opens the same connections sql.Open("sqlite", dsn) does, from the same registered driver, so every function, collation, connection hook and virtual table module registered through this package applies to them. It exists for callers that need to interpose on the physical connections database/sql opens — tracing, metrics, connection-scoped setup — which sql.Open gives no access to: such a caller can embed the returned Connector, override Connect, and pass its own wrapper to sql.OpenDB. Previously the only way to reach the registered driver was the db, _ := sql.Open("sqlite", ""); drv := db.Driver(); db.Close() idiom, which works only because sql.Open does not connect and this driver does not implement driver.DriverContext; and the only way to get a wrapper into a *sql.DB was sql.Register, which is process-global, panics on a name it has already seen, and cannot be undone, so a library had to invent a unique driver name per configuration. sql.OpenDB registers nothing. Constructing a &sqlite.Driver{} is not an alternative — its fields are unexported, so it carries none of the registrations. NewConnector checks the DSN only as far as it can without opening a database — a query string that does not parse, and conflicting vfs parameters; everything else continues to be validated when the connection is opened, so an unknown parameter or an out-of-range value is reported by Connect rather than at construction. Nothing about the existing sql.Open path changes: *Driver deliberately still does not implement driver.DriverContext, so sql.Open remains lazy and DSN errors continue to surface where they always have. A runnable sample is in examples/connector. Resolves GitLab issue #253, thanks Alessandro Segala (@ItalyPaleAle)!
  • Document that a caller-constructed sqlite.Driver is not the driver this package registers as "sqlite". Its fields are unexported, so it starts with no functions, collations or connection hooks and the only way to give it any is its own RegisterConnectionHook method; the package-level Register* functions always apply to the registered driver. Connections such a Driver opens therefore run without the package-level functions and collations — and because a registered function silently replaces a SQLite built-in of the same name, a Driver you construct can evaluate upper(x), date(x) and the like differently from one opened through sql.Open. Virtual table modules are the one exception: they are held process-globally and reach every Driver. Constructing one remains supported for the private-hook pattern — a driver registered under a name of its own with sql.Register so its connection hooks apply only to its own connections — and is otherwise best avoided in favour of sql.Open or NewConnector. Documentation only; no behavior changes.

• 2026-07-20 v1.55.0:

  • Add github.com/mattn/go-sqlite3-compatible shorthand DSN query parameters to ease migration from that driver: _busy_timeout/_timeout, _foreign_keys/_fk, _journal_mode/_journal, _synchronous/_sync, _auto_vacuum/_vacuum, and _query_only, each setting the correspondingly named PRAGMA. Values are validated against the same set mattn/go-sqlite3 accepts (case-insensitive) and an unrecognized value fails the connection with an error, so a typo such as _synchronous=fu1l or _foreign_keys=yes_please is reported rather than silently downgrading durability or dropping foreign-key enforcement. The keys are applied in a fixed order independent of their order in the DSN — _busy_timeout and _auto_vacuum before any _pragma values (auto_vacuum must be set before the database is first written), the rest after, and _query_only last — and where a key and its alias are both supplied the alias wins, matching mattn/go-sqlite3; selection is by presence rather than by value, so supplying the alias empty (_foreign_keys=on&_fk=) suppresses the PRAGMA rather than deferring to the primary key, again matching that driver. Behavior change to note: prior releases ignored these keys entirely, so a DSN carried over from a mattn/go-sqlite3 setup changes in two ways. A recognized key that previously did nothing now takes effect — _foreign_keys=on begins enforcing constraints against data that may already violate them, _journal_mode=wal persistently converts the database file, and _query_only=1 makes the connection read-only. And a value outside the accepted set now fails the connection with an error where the same DSN previously opened successfully — for example a duration-style _busy_timeout=5s or _timeout=5000ms, neither of which is the integer that key requires. Review such DSNs before upgrading. _pragma is unchanged and no pre-existing parameter changes meaning, though see the following entry for a change in when all of them are validated.
  • See GitLab merge request #134, thanks Toni Spets (@beeper-hifi) and Ian Chechin!
  • Validate every DSN query parameter before applying any of them. Parameters were previously checked as each was reached, so a DSN whose later parameter was rejected had already executed the PRAGMAs ahead of it. Because PRAGMA journal_mode and PRAGMA auto_vacuum are persistent changes to the database file, a DSN such as file:x.db?_journal_mode=wal&_synchronous=bogus failed the connection and yet left x.db converted to WAL. A failed Open now leaves the database as it found it. This covers the pre-existing _txlock, _timezone, _time_format, _time_integer_format, _inttotime and _texttotime parameters as well as the shorthand keys above: all of them were validated only after the _pragma list had already run, so the same DSN shape — a valid _pragma=journal_mode=wal alongside a misspelled _txlock — converted the file before reporting the error. Only the values accepted for each parameter are unchanged; a DSN that opened successfully before still opens, and one that failed still fails with the same error. _pragma remains the sole exception, since its values are executed verbatim and cannot be checked in advance: a malformed _pragma is still rejected by SQLite as it runs, after any earlier _pragma in the list has taken effect.

• 2026-07-15 v1.54.0:

  • Upgrade to SQLite 3.53.3. This also bumps the pinned modernc.org/libc to v1.74.1; as always, downstream modules must pin the exact same modernc.org/libc version this module's go.mod pins (see GitLab issue #177).
  • Under the opt-in _texttotime DSN parameter, best-effort parse date-shaped TEXT values from columns SQLite reports with an empty declared type — aggregates and expressions over a date column (MAX(d), COALESCE(d, ...), upper(d), d || ''), subqueries, and typeless real columns (CREATE TABLE t(x)) — into time.Time, instead of delivering them as a raw string that Scan cannot store into a *time.Time. The existing declared DATE/DATETIME/TIME/TIMESTAMP path is unchanged; this only adds the empty-decltype case. The conversion is strictly best-effort: a value that does not parse as a time falls through to the original string, so no Scan that worked before can newly fail. ColumnTypeScanType continues to report string for empty-decltype columns, since the declared type cannot prove the column is temporal. Without _texttotime the behavior is byte-for-byte unchanged. Resolves GitLab issue #248.
  • See GitLab merge request #133, thanks Ian Chechin!

• 2026-06-21 v1.53.0:

  • Add experimental netbsd/amd64 support, resolving the long-standing build break in GitLab issue #246. This target is intentionally not yet listed among the supported platforms in the package documentation: the port had been broken for years and is only now revived, and there is as yet no real-world experience running it under production workloads. Green CI is not the same as battle-tested — so while the full test suite (including the pcache and vec packages and the -race concurrency test) passes on NetBSD 10.1 / Go 1.26.3, and the entire upstream toolchain (libc, cc, ccgo, libz, libtcl8.6, libsqlite3, libsqlite_vec) is green on the NetBSD CI builder, the target is offered for evaluation only. If you run NetBSD, please exercise it with your own workloads and report back via #246; the intent is to promote it to a fully supported platform after a period of broader real-world testing (on the order of a month) elapses without surprises.
  • Implementation notes: the previously shipped lib/sqlite_netbsd_amd64.go was a stale old-generator transpile that no longer compiled (the mu.enter/mu.leave break in #246); it is replaced by a fresh new-generator transpile consistent with every other platform, and modernc.org/sqlite/vec (sqlite-vec) is vendored and auto-registers on netbsd. Correct operation requires the matching pinned modernc.org/libc, which carries two NetBSD-specific fixes found during this work: the mmap(2) PAD-argument ABI (without it, concurrent WAL access faults with SIGBUS in the WAL-index shared memory) and a working abort(3) (the prior stub left SQLite's crash-recovery writecrash test unable to terminate by signal). As usual, downstream modules must pin the exact modernc.org/libc version this module's go.mod pins.
  • See GitLab merge request #82, thanks Leonardo Taccari (@iamleot) and Thomas Klausner (@wiz)!
  • Add experimental freebsd/386 and freebsd/arm support. As with the netbsd/amd64 target above, these two 32-bit FreeBSD ports are intentionally not yet listed among the supported platforms in the package documentation: freebsd/386 previously shipped a stale, effectively untested SQLite 3.41 transpile, and freebsd/arm is entirely new, so neither has real-world production mileage yet. Both are now freshly transpiled at SQLite 3.53.2 consistent with every other platform, build cleanly, and pass the full test suite (core, WAL/concurrency, and the vec package) on the FreeBSD CI builders; they are offered for evaluation only. If you run 32-bit FreeBSD, please exercise these targets with your own workloads and report back — the intent is to promote freebsd/386, freebsd/arm, and netbsd/amd64 to fully supported platforms in a future release cycle, once a period of broader real-world testing elapses without surprises.
  • Implementation notes: correct operation on freebsd/arm requires the matching pinned modernc.org/libc (v1.73.4), which fixes the per-arch mmap(2) off_t encoding for 32-bit FreeBSD; without it the WAL shared-memory mapping faults with SIGBUS under concurrent access, the same class of bug found on the netbsd port. As usual, downstream modules must pin the exact modernc.org/libc version this module's go.mod pins.
  • See GitLab merge request #119, thanks Olivier Cochard-Labbé (@ocochard)!
  • Add a Go-facing wrapper for SQLITE_CONFIG_PCACHE2. PageCache is the factory and Cache the per-database instance, both idiomatic Go interfaces; Page exposes the raw Buf and Extra pointers that SQLite reads through the C pcache contract. RegisterPageCache and MustRegisterPageCache install the module process-globally before the first sql.Open; subsequent Open calls are gated through a one-shot Xsqlite3_config(SQLITE_CONFIG_PCACHE2) so a too-late Register returns ErrPageCacheTooLate rather than silently falling through to the built-in pcache1. The binding owns the sqlite3_pcache_page stub and re-consults the implementation on every Fetch, reusing the stub only when the returned Page value is unchanged, which keeps a bounded/evicting purgeable cache safe by construction.
  • See GitLab merge request #126, thanks Ian Chechin!
  • Add modernc.org/sqlite/pcache, the reference page-cache implementation that accompanies the #126 SQLITE_CONFIG_PCACHE2 wrapper. pcache.New returns a *Pool satisfying the PageCache interface; register it once with sqlite.MustRegisterPageCache(pcache.New()) and every connection opened afterwards draws its pages from it. Each Pool.Create mints a fresh per-database Cache: a bounded, LRU-evicting page store that honours the PRAGMA cache_size soft cap and releases the least-recently-unpinned page when it must make room. Page memory — the Buf and Extra buffers SQLite reads through — is allocated with libc.Xmalloc/libc.Xcalloc and therefore lives off the Go heap, which keeps SQLite's interior pointer arithmetic on the page extras from tripping the race detector's checkptr enforcement. Pool.Stats reports aggregate lifetime counters (hits, misses, allocs, evictions, rekeys, truncates, caches) across every cache a Pool has created, so hit/miss/eviction behaviour is observable without instrumenting individual caches. Cross-connection page sharing is out of scope for now; each Create returns an independent per-database cache.
  • Validated end-to-end against the #126 stress workload (cache_size=16, 4000 BLOB rows with DELETE and incremental_vacuum, integrity_check clean under -race) and benchmarked for the memory-utilization goal tracked in GitLab issue #204.
  • See GitLab merge request #127, thanks Ian Chechin!
  • Tighten the modernc.org/sqlite/pcache reference implementation per cznic's !127 review follow-ups. Adds Stats.EasyRefusals, a per-Pool counter for the cases where FetchCreateEasy returns nil at cap; SQLite reacts to a refusal by spilling dirty pages and retrying with FetchCreateForce, so the new field is a direct proxy for the I/O pressure the strict Easy contract imposes vs pcache1's recycle-without-spill behavior. BenchmarkPoolEvictionChurn was reworked to drive a rotating-residue DELETE (k % 3 = i % 3) and re-insert a matching batch each cycle so the spill pressure recurs and easy-refusals/op scales with b.N instead of capping at the seed's one-time first-cycle cost; both existing benchmarks now report easy-refusals/op alongside the page-allocs/evictions metrics. Stats.Evictions documentation was tightened to match the actual behavior (counts LRU eviction, Unpin(discard=true), Shrink releases, and Unpin(discard=false) trimming back to target after a FetchCreateForce overcommit; bulk frees from Truncate, Rekey collisions, and Destroy are not counted). The TestPoolRoundTripIntegrity comment claiming the workload exercises xRekey ~15 times has been corrected; the SQL surface does not reliably emit xRekey here, and that codepath is covered by the unit tests instead.
  • See GitLab merge request #130, thanks Ian Chechin!
  • Make modernc.org/sqlite/pcache -race-clean under SQLite's cache=shared mode. The pool already runs correctly under shared-cache because every callback into a given Cache is serialised internally by SQLite's sqlite3BtreeEnter on the BtShared mutex; verified empirically with a lock-free in-flight probe (max-in-flight = 1 on the canonical two-connection workload, 4 on a positive control with goroutines hitting the cache directly). However the Go race detector does not recognise SQLite's libc mutex as a happens-before edge and reports false-positive races on Fetch vs Unpin reads/writes of the per-cache state, which surfaces as DATA RACE failures for any user who registers the pool and runs their suite under -race. A sync.Mutex on the cache type is now taken on every public method (SetSize, PageCount, Fetch, Unpin, Rekey, Truncate, Destroy, Shrink), always. On the common non-shared-cache path the lock is uncontended (one atomic CAS per Lock/Unlock pair, negligible next to the SQLite work it bookends); on the shared-cache path it just rubber-stamps the order SQLite's BtShared mutex already established. A new e2e_test.go TestSharedCacheTwoConns_Integrity drives two sql.Conn against the same cache=shared URI with concurrent writers and asserts PRAGMA integrity_check = ok under -race; passes cleanly with the lock, would surface the false-positive without it. Design notes live in pcache/sharing.go.
  • See GitLab merge request #131, thanks Ian Chechin!
  • Add a Go wrapper for sqlite3_db_status, the per-connection runtime counters (cache hit/miss/write/spill rates, schema and prepared-statement memory, lookaside usage, deferred foreign keys). DBStatus is an interface implemented by the driver connection and reached through the database/sql escape hatch (*sql.Conn).Raw(), mirroring the existing FileControl surface; DBStatusOp is a distinct typed enum of the SQLITE_DBSTATUS_* verbs so a counter from a different op family will not compile in its place. Status(op, reset) returns the (current, high) pair and optionally resets the counter. This also lets modernc.org/sqlite/pcache measure real I/O instead of the EasyRefusals proxy: the new BenchmarkPoolSpillIO reads the pager-level SQLITE_DBSTATUS_CACHE_SPILL/_CACHE_WRITE counters, which the pager maintains identically for pcache1 and the pool, making the pcache1-vs-pool comparison cznic raised on the !127 review a genuine apples-to-apples measurement. On the rotating-residue eviction-churn workload at cache_size=16 the pool spills ~3.5x more than pcache1 (cache-spill/op 31.96 vs 8.96) for ~3% more page writes (cache-write/op 450 vs 436) at identical hit/miss, quantifying the I/O cost of the strict Easy contract that EasyRefusals only proxied.
  • See GitLab merge request #132, thanks Ian Chechin!
  • Add an opt-in _dqs DSN query parameter that disables SQLite's double-quoted string literal compatibility quirk on a per-connection basis. When _dqs=0 (or any strconv.ParseBool false value) is supplied, the driver calls sqlite3_db_config with SQLITE_DBCONFIG_DQS_DDL and SQLITE_DBCONFIG_DQS_DML set to off before any statement is prepared, so a double-quoted identifier that fails to resolve raises a parse error instead of silently falling back to a string literal. Absence of the parameter, or _dqs=1, leaves SQLite's default behavior unchanged; existing DSNs continue to work byte-for-byte. Resolves GitLab issue #61.
  • See GitLab merge request #128, thanks Ian Chechin!
  • Add an opt-in _error_rc DSN query parameter for clearer error reporting on open-time failures. When _error_rc=1 (or any strconv.ParseBool true value) is supplied, error strings synthesised from a (rc, db) pair only append sqlite3_errmsg(db) when sqlite3_extended_errcode(db) is consistent with the operation rc (full match first, primary code &0xff as fallback). On mismatch the canonical sqlite3_errstr(rc) is used alone, so an open-time SQLITE_CANTOPEN no longer carries the temporary handle's stale "out of memory" errmsg. Absence of the parameter, or _error_rc=0, preserves the legacy "errstr: errmsg" form byte-for-byte; existing callers that parse error strings are unaffected. The driver's *Error.Code() returns the same SQLite result code in both modes. Parsed before sqlite3_open_v2 so open-time errors are covered. Resolves GitLab issue #230.
  • See GitLab merge request #129, thanks Ian Chechin!

• 2026-06-06 v1.52.0:

  • Upgrade to SQLite 3.53.2.
  • Add Backup.Remaining and Backup.PageCount, thin wrappers around the existing sqlite3_backup_remaining and sqlite3_backup_pagecount C symbols. Together they expose the per-Step progress counters that the underlying backup object already maintains, enabling progress reporting during online backups without dropping to modernc.org/sqlite/lib directly.
  • See GitLab merge request #122, thanks Ian Chechin!
  • Drop the redundant second copy in (*conn).columnText, the path that backs every Rows.Scan into a Go string for a TEXT column. The value's bytes are still copied once out of SQLite-owned memory into a fresh Go buffer; that buffer is then reinterpreted as the result string with unsafe.String rather than copied a second time by the implicit string([]byte) conversion. This removes one allocation per TEXT value per row and roughly halves the bytes allocated on that path; on the new BenchmarkColumnTextScan cases it is ~13–20% faster for payloads of 256 B and larger, with no measurable change for very short strings. Purely internal: no API or behavioral change, and the returned string never aliases SQLite's buffer.
  • See GitLab merge request #123, thanks Ian Chechin!
  • Cache each result column's declared type once per result set in newRows instead of recomputing it on every row. The TEXT branch of Rows.Next calls ColumnTypeDatabaseTypeName for every TEXT column on every row (independent of any DSN flag), which previously did a libc.GoString + strings.ToUpper each time; that lookup is now a single index into a cached, pre-uppercased []string, and ColumnTypeScanType reads the same cache and drops its per-call strings.ToLower. The declared type is fixed for the lifetime of a prepared statement, so the C round-trip is paid once per column rather than once per column per row, removing exactly 1 alloc + 8 B per TEXT column per row from the Next hot path. The new BenchmarkTextToTimeScan cases show ~7% faster on a 1000-row DATETIME SELECT under _texttotime=1. Purely internal: ColumnTypeDatabaseTypeName and ColumnTypeScanType return identical values, no API or behavioral change.
  • See GitLab merge request #124, thanks Ian Chechin!
  • Cache, per result column, the parseTimeFormats index that first parsed a TEXT-stored DATE/DATETIME/TIMESTAMP value, and try that format first on later rows instead of re-walking the list from the top. (*conn).parseTime previously ran time.Parse down the format list on every such row; for the canonical SQLite TEXT datetime format every row paid two failed time.Parse attempts — each allocating a *time.ParseError — before the match. On a 1000-row DATETIME TEXT SELECT this cuts ~50% of allocs/op and ~57% of B/op and is ~37% faster. The fall-through chain is preserved exactly: the seven formats are mutually exclusive, so the cached hint can never select a different match than the in-order scan, and the parsed driver.Value is identical to before. Purely internal: no API or behavioral change.
  • See GitLab merge request #125, thanks Ian Chechin!

• 2026-05-28 v1.51.0:

  • Pool the []driver.Value slice passed to scalar/aggregate UDF callbacks and to vtab Filter/Insert/Update callbacks, eliminating the dominant per-row allocation on UDF-heavy queries. Benchmarks on a 1000-row, 3-arg noop scalar UDF show ~40% fewer bytes/op and ~15% fewer allocs/op.
  • Document the matching "arguments are not valid past return" contract on vtab.Cursor.Filter and vtab.Updater.Insert/Update, consistent with the existing rule for FunctionImpl.Scalar / AggregateFunction.Step / WindowInverse.
  • Resolves GitLab issue #226. See GitLab merge request #114, thanks Ian Chechin!
  • Add FileControl.FileControlDataVersion, a wrapper around SQLITE_FCNTL_DATA_VERSION for observing pager-cache data-version changes, including those made on the same connection. Useful as a primitive for application-level cache invalidation.
  • Exposed via the idiomatic database/sql escape hatch (*sql.Conn).Raw(), consistent with the existing FileControlPersistWAL.
  • See GitLab merge request #115, thanks Ian Chechin!
  • Fix a regression where in-memory connections (:memory:, file::memory:, shared-cache memory URIs) were discarded by database/sql after a context-cancelled query, taking the entire in-memory store with them. The fix for #198 had added an sqlite3_is_interrupted check to the connection validator that mistakenly applied to in-memory connections too, re-introducing the bug originally fixed by !74. File-backed connections keep the existing behaviour and are still discarded after an interrupt.
  • Resolves GitLab issue #196. See GitLab merge request #116, thanks Ian Chechin!
  • Add an opt-in FunctionImpl.VolatileArgs flag that hands TEXT and BLOB arguments to scalar and aggregate UDF callbacks as zero-copy views (unsafe.String/unsafe.Slice) over SQLite's own value buffers, eliminating the per-argument libc.GoString/make([]byte) copy that the #226 slice-pooling left as the remaining per-row allocation. On the same 1000-row, 3-arg (INTEGER/TEXT/BLOB) noop scalar UDF this removes a further ~35% of allocs/op and ~11% of bytes/op on top of #226.
  • The views are valid only for the duration of the callback and must not be retained past return or across rows; a callback that needs to keep a value must copy it. With VolatileArgs unset (the default) arguments keep the existing copied, caller-owned semantics, so the flag is fully backward compatible; it has no effect on integer, float, time, or NULL arguments.
  • See GitLab merge request #120, thanks Ian Chechin!
  • Extend the opt-in VolatileArgs zero-copy TEXT/BLOB argument access from #120 to the virtual-table Cursor.Filter (xFilter) and Updater.Insert/Update (xUpdate) callbacks. A vtab.Module opts in by implementing the new optional vtab.VolatileArgsOpter interface (VolatileArgs() bool); the flag is read once at module registration and shared by every table created from it. On a vtab call carrying one TEXT and one BLOB argument this removes 2 allocs/op (one libc.GoString, one make([]byte)) on each of the Filter and Update paths.
  • The same safety contract as #120 applies: the views are valid only for the duration of the callback and must not be retained past return or across rows; a callback that needs to keep a value must copy it. Modules that do not implement VolatileArgsOpter (the default for all existing modules) are byte-for-byte unchanged, and the flag has no effect on integer, float, time, or NULL arguments.
  • See GitLab merge request #121, thanks Ian Chechin!

• 2026-05-10 v1.50.1:

  • Upgrade to SQLite 3.53.1.

• 2026-04-24 v1.50.0:

  • Upgrade to sqlite-vec v0.1.9.
  • Introduce ColumnInfo, enabling dynamic query builders and ORMs to retrieve underlying SQLite C-API metadata (OriginName, TableName, DatabaseName, and DeclType).
  • This feature is exposed via the idiomatic database/sql escape hatch (*sql.Conn).Raw(), avoiding custom statement handles and keeping the standard library workflow intact.
  • See GitLab merge request #113, thanks Josh Bleecher Snyder!

• 2026-04-17 v1.49.1: Documentation only: update the per-platform SQLite version table in doc.go to 3.53.0, the version shipped since v1.49.0. No code changes.

• 2026-04-17 v1.49.0: Upgrade to SQLite 3.53.0.

  • Added -DSQLITE_ENABLE_DBPAGE_VTAB to the transpilation. See "The SQLITE_DBPAGE Virtual Table" for details.

• 2026-04-06 v1.48.2:

  • Fix ABI mapping mismatch in the pre-update hook trampoline that caused silent truncation of large 64-bit RowIDs.
  • Ensure the Go trampoline signature correctly aligns with the public sqlite3_preupdate_hook C API, preventing data corruption for high-entropy keys (e.g., Snowflake IDs).
  • See GitLab merge request #98, thanks Josh Bleecher Snyder!
  • Fix the memory allocator used in (*conn).Deserialize.
  • Replace tls.Alloc with sqlite3_malloc64 to prevent internal allocator corruption. This ensures the buffer is safely owned by SQLite, which may resize or free it due to the SQLITE_DESERIALIZE_RESIZEABLE and SQLITE_DESERIALIZE_FREEONCLOSE flags.
  • Prevent a memory leak by properly freeing the allocated buffer if fetching the main database name fails before handing ownership to SQLite.
  • See GitLab merge request #100, thanks Josh Bleecher Snyder!
  • Fix (*conn).Deserialize to explicitly reject nil or empty byte slices.
  • Prevent silent database disconnection and connection pool corruption caused by SQLite's default behavior when sqlite3_deserialize receives a 0-length buffer.
  • See GitLab merge request #101, thanks Josh Bleecher Snyder!
  • Fix commitHookTrampoline and rollbackHookTrampoline signatures by removing the unused pCsr parameter.
  • Aligns internal hook callbacks accurately with the underlying SQLite C API, cleaning up the code to prevent potential future confusion or bugs.
  • See GitLab merge request #102, thanks Josh Bleecher Snyder!
  • Fix checkptr instrumentation failures during go test -race when registering and using virtual tables (vtab).
  • Allocate sqlite3_module instances using the C allocator (libc.Xcalloc) instead of the Go heap. This ensures transpiled C code can safely perform pointer operations on the struct without tripping Go's pointer checks.
  • See GitLab merge request #103, thanks Josh Bleecher Snyder!
  • Fix data race on mutex.id in the mutexTry non-recursive path.
  • Ensure consistent atomic writes (atomic.StoreInt32) to prevent data races with atomic loads in mutexHeld and mutexNotheld during concurrent execution.
  • See GitLab merge request #104, thanks Josh Bleecher Snyder!
  • Fix resource leak in (*Backup).Commit where the destination connection was not closed on error.
  • Ensure dstConn is properly closed when sqlite3_backup_finish fails, preventing file descriptor, TLS, and memory leaks.
  • See GitLab merge request #105, thanks Josh Bleecher Snyder!
  • Fix Exec to fully drain rows when encountering SQLITE_ROW, preventing silent data loss in DML statements.
  • Previously, Exec aborted after the first row, meaning INSERT, UPDATE, or DELETE statements with a RETURNING clause would fail to process subsequent rows. The execution path now correctly loops until SQLITE_DONE and properly respects context cancellations during the drain loop, fully aligning with native C sqlite3_exec semantics.
  • See GitLab merge request #106, thanks Josh Bleecher Snyder!
  • Fix "Shadowed err value (stmt.go)".
  • See GitLab issue #249, thanks Emrecan BATI!
  • Fix silent omission of virtual table savepoint callbacks by correctly setting the sqlite3_module version.
  • See GitLab merge request #107, thanks Josh Bleecher Snyder!
  • Fix vfsRead to properly handle partial and fragmented reads from io.Reader.
  • Replace f.Read with io.ReadFull to ensure the buffer is fully populated, preventing premature SQLITE_IOERR_SHORT_READ errors on valid mid-stream partial reads. Unread tail bytes at EOF are now efficiently zero-filled using the built-in clear function.
  • See GitLab merge request #108, thanks Josh Bleecher Snyder!
  • Refactor internal error formatting to safely handle uninitialized or closed database pointers.
  • Prevent a misleading "out of memory" error message when an operation fails and the underlying SQLite database handle is NULL (db == 0).
  • See GitLab merge request #109, thanks Josh Bleecher Snyder!
  • Fix error handling in database backup and restore initialization (sqlite3_backup_init).
  • Ensure error codes and messages are accurately read from the destination database handle rather than hardcoding the source or remote handle. This prevents swallowed errors or mismatched "not an error" messages when a backup or restore operation fails to start.
  • See GitLab merge request #111, thanks Josh Bleecher Snyder!
  • Fix database handle and C-heap memory leaks when sqlite3_open_v2 fails.
  • Ensure sqlite3_close_v2 is called on the partially allocated database handle during a failed open, and explicitly close libc.TLS in newConn to prevent resource leakage.
  • Prevent misleading "out of memory" error messages on failed connections by correctly extracting the exact error string from the allocated handle before it is closed.
  • See GitLab merge request #112, thanks Josh Bleecher Snyder!

• 2026-04-03 v1.48.1:

  • Fix memory leaks and double-free vulnerabilities in the multi-statement query execution path.
  • Ensure bind-parameter allocations are reliably freed via strict ownership transfer if an error occurs mid-loop or if multiple statements bind parameters.
  • Fix a resource leak where a subsequent statement's error could orphan a previously generated rows object without closing it, leaking the prepared statement handle.
  • See GitLab merge request #96, thanks Josh Bleecher Snyder!

• 2026-03-27 v1.48.0:

  • Add _timezone DSN query parameter to apply IANA timezones (e.g., "America/New_York") to both reads and writes.
  • Writes will convert time.Time values to the target timezone before formatting as a string.
  • Reads will interpret timezone-less strings as being in the target timezone.
  • Does not impact _inttotime integer values, which will always safely evaluate as UTC.
  • Add support for _time_format=datetime URI parameter to format time.Time values identically to SQLite's native datetime() function and CURRENT_TIMESTAMP (YYYY-MM-DD HH:MM:SS).
  • See GitLab merge request #94 and GitLab merge request #95, thanks Josh Bleecher Snyder!

• 2026-03-17 v1.47.0: Add CGO-free version of the vector extensions from https://github.com/asg017/sqlite-vec. See vec_test.go for example usage. From the GitHub project page:

  • Important: sqlite-vec is a pre-v1, so expect breaking changes!
  • Store and query float, int8, and binary vectors in vec0 virtual tables
  • Written in pure C, no dependencies, runs anywhere SQLite runs (Linux/MacOS/Windows, in the browser with WASM, Raspberry Pis, etc.)
  • Store non-vector data in metadata, auxiliary, or partition key columns
  • See GitLab merge request #93, thanks Zhenghao Zhang!

• 2026-03-16 v1.46.2: Upgrade to SQLite 3.51.3.

• 2026-02-17 v1.46.1:

  • Ensure connection state is reset if Tx.Commit fails. Previously, errors like SQLITE_BUSY during COMMIT could leave the underlying connection inside a transaction, causing errors when the connection was reused by the database/sql pool. The driver now detects this state and forces a rollback internally.
  • Fixes GitHub issue #2, thanks Edoardo Spadolini!

• 2026-02-17 v1.46.0:

  • Enable ColumnTypeScanType to report time.Time instead of string for TEXT columns declared as DATE, DATETIME, TIME, or TIMESTAMP via a new _texttotime URI parameter.
  • See GitHub pull request #1, thanks devhaozi!

• 2026-02-09 v1.45.0:

  • Introduce vtab subpackage (modernc.org/sqlite/vtab) exposing Module, Table, Cursor, and IndexInfo API for Go virtual tables.
  • Wire vtab registration into the driver: vtab.RegisterModule installs modules globally and each new connection calls sqlite3_create_module_v2.
  • Implement vtab trampolines for xCreate/xConnect/xBestIndex/xDisconnect/xDestroy/xOpen/xClose/xFilter/xNext/xEof/xColumn/xRowid.
  • Map SQLite’s sqlite3_index_info into vtab.IndexInfo, including constraints, ORDER BY terms, and constraint usage (ArgIndex → xFilter argv[]).
  • Add an in‑repo dummy vtab module and test (module_test.go) that validates registration, basic scanning, and constraint visibility.
  • See GitLab merge request #90, thanks Adrian Witas!

• 2026-01-19 v1.44.3: Resolves GitLab issue #243.

• 2026-01-18 v1.44.2: Upgrade to SQLite 3.51.2.

• 2026-01-15 v1.44.1: Bump the pinned modernc.org/libc to v1.67.6. Its memory-safe strlen fixes the AddressSanitizer global-buffer-overflow in the transpiled window-function name arrays reported in GitLab merge request #88, which was closed in favour of that upstream fix, thanks Matt Topol!

• 2026-01-13 v1.44.0: Upgrade to SQLite 3.51.1.

• 2026-01-08 v1.43.0:

  • Add IsReadOnly(schema string) (bool, error) on the driver connection, a wrapper around sqlite3_db_readonly, reachable through (*sql.Conn).Raw(). Resolves GitLab issue #242.
  • Honour _time_integer_format=unix_micro and unix_nano when reading integer time values back with _inttotime=true; previously only the seconds/milliseconds heuristic applied. Resolves GitLab issue #240.
  • Make the mutex implementation in lib/mutex.go more robust.
  • Re-vendor the transpiled library from modernc.org/libsqlite3 v1.11.0; the SQLite version stays 3.50.4.
  • Add openbsd/amd64 and openbsd/arm64 to the targets exercised by the builders (see GitLab issue #237).

• 2025-12-28 v1.42.2: Fix a TOCTOU race in query interruption: a cancelled context could interrupt an unrelated, later query on the same connection. Resolves GitLab issue #241. See GitLab merge request #86, thanks Josh Bleecher Snyder!

• 2025-12-27 v1.42.1: Revert to the v1.41.0 state (SQLite 3.50.4) and retract v1.42.0 in go.mod.

• 2025-12-27 v1.42.0: Upgrade to SQLite 3.51.1. Retracted the same day as accidentally broken, see the retract directive in go.mod; v1.42.1 reverts it and v1.44.0 carries the working 3.51.1 upgrade.

• 2025-12-08 v1.41.0:

  • Add the modernc.org/sqlite/vtab subpackage, a Go virtual-table API (Module, Table, Cursor, IndexInfo, vtab.RegisterModule) bridged to the engine and validated by an in-tree dummy module. See GitLab merge request #84, thanks Adrian Witas!
  • Optimize prepared statements: reuse the prepared SQLite statement handle across executions instead of preparing it again. Updates GitLab issue #236.
  • Fix the build on 32-bit targets.

• 2025-11-12 v1.40.1: Close the Rows of a query whose context was cancelled after the statement had already produced them; such Rows were previously leaked, which kept a read transaction open, starved WAL checkpoints and let the WAL grow without bound. See GitLab merge request #81, thanks Silvio Moioli!

• 2025-10-17 v1.40.0: Add pre-update, commit and rollback hooks: RegisterPreUpdateHook, RegisterCommitHook and RegisterRollbackHook on the driver connection, plus the SQLitePreUpdateData type (Depth, Count, Old, New), reachable through (*sql.Conn).Raw(). Updates GitLab issue #125. See GitLab merge request #80, thanks Walter Wanderley!

• 2025-10-10 v1.39.1: Upgrade to SQLite 3.50.4.

• 2025-08-11 v1.39.0:

  • Add the _time_integer_format DSN parameter (unix, unix_milli, unix_micro, unix_nano): time.Time values are bound as integers in that unit instead of as text, so timestamps fit INTEGER columns, STRICT tables included. Resolves GitLab issue #216. See GitLab merge request #77, thanks Guénaël Muller!
  • Add the _inttotime DSN parameter: when enabled, INTEGER values read from columns declared DATE, DATETIME or TIMESTAMP are converted to time.Time (seconds, or milliseconds for magnitudes above 1e12), matching mattn/go-sqlite3, so scanning them into a *time.Time no longer fails. Resolves GitLab issue #214. See GitLab merge request #76, thanks Harald Albrecht!
  • Add a VFS unit test. See GitLab merge request #78, thanks Harald Albrecht!

• 2025-07-28 v1.38.2: Upgrade to SQLite 3.50.3, this time including the regenerated library that v1.38.1 lacked.

• 2025-07-25 v1.38.1: Update dependencies. Intended as the upgrade to SQLite 3.50.3, but the regenerated library was not included, so the shipped SQLite version is still 3.50.1. Superseded by v1.38.2.

• 2025-06-09 v1.38.0: Upgrade to SQLite 3.50.1.

• 2025-02-26 v1.36.0: Upgrade to SQLite 3.49.0.

• 2024-11-16 v1.34.0: Implement ResetSession and IsValid methods in connection

• 2024-07-22 v1.31.0: Support windows/386.

• 2024-06-04 v1.30.0: Upgrade to SQLite 3.46.0, release notes at https://sqlite.org/releaselog/3_46_0.html.

• 2024-02-13 v1.29.0: Upgrade to SQLite 3.45.1, release notes at https://sqlite.org/releaselog/3_45_1.html.

• 2023-12-14: v1.28.0: Add (*Driver).RegisterConnectionHook, ConnectionHookFn, ExecQuerierContext, RegisterConnectionHook.

• 2023-08-03 v1.25.0: enable SQLITE_ENABLE_DBSTAT_VTAB.

• 2023-07-11 v1.24.0: Add (*conn).{Serialize,Deserialize,NewBackup,NewRestore} methods, add Backup type.

• 2023-06-01 v1.23.0: Allow registering aggregate functions.

• 2023-04-22 v1.22.0: Support linux/s390x.

• 2023-02-23 v1.21.0: Upgrade to SQLite 3.41.0, release notes at https://sqlite.org/releaselog/3_41_0.html.

• 2022-11-28 v1.20.0: Support linux/ppc64le.

• 2022-09-16 v1.19.0: Support frebsd/arm64.

• 2022-07-26 v1.18.0: Add support for Go fs.FS based SQLite virtual filesystems, see function New in modernc.org/sqlite/vfs and/or TestVFS in all_test.go

• 2022-04-24 v1.17.0: Support windows/arm64.

• 2022-04-04 v1.16.0: Support scalar application defined functions written in Go. See https://www.sqlite.org/appfunc.html

• 2022-03-13 v1.15.0: Support linux/riscv64.

• 2021-11-13 v1.14.0: Support windows/amd64. This target had previously only experimental status because of a now resolved memory leak.

• 2021-09-07 v1.13.0: Support freebsd/amd64.

• 2021-06-23 v1.11.0: Upgrade to use sqlite 3.36.0, release notes at https://www.sqlite.org/releaselog/3_36_0.html.

• 2021-05-06 v1.10.6: Fixes a memory corruption issue (https://gitlab.com/cznic/sqlite/-/issues/53). Versions since v1.8.6 were affected and should be updated to v1.10.6.

• 2021-03-14 v1.10.0: Update to use sqlite 3.35.0, release notes at https://www.sqlite.org/releaselog/3_35_0.html.

• 2021-03-11 v1.9.0: Support darwin/arm64.

• 2021-01-08 v1.8.0: Support darwin/amd64.

• 2020-09-13 v1.7.0: Support linux/arm and linux/arm64.

• 2020-09-08 v1.6.0: Support linux/386.

• 2020-09-03 v1.5.0: This project is now completely CGo-free, including the Tcl tests.

• 2020-08-26 v1.4.0: First stable release for linux/amd64. The database/sql driver and its tests are CGo free. Tests of the translated sqlite3.c library still require CGo.

• 2020-07-26 v1.4.0-beta1: The project has reached beta status while supporting linux/amd64 only at the moment. The 'extraquick' Tcl testsuite reports

• 2019-12-28 v1.2.0-alpha.3: Third alpha fixes issue #19.

• 2019-12-26 v1.1.0-alpha.2: Second alpha release adds support for accessing a database concurrently by multiple goroutines and/or processes. v1.1.0 is now considered feature-complete. Next planed release should be a beta with a proper test suite.

• 2019-12-18 v1.1.0-alpha.1: First alpha release using the new cc/v3, gocc, qbe toolchain. Some primitive tests pass on linux_{amd64,386}. Not yet safe for concurrent access by multiple goroutines. Next alpha release is planed to arrive before the end of this year.

• 2017-06-10: Windows/Intel no more uses the VM (thanks Steffen Butzer).

• 2017-06-05 Linux/Intel no more uses the VM (cznic/virtual).