1.2.0 is a minor version release containing potentially breaking behavioral changes:

• Client error codes previous returned with the XX prefix are now more specific and match Postgres behavior, e.g. 44 prefix for write contention errors.
• The auth.db file, containing users and grants, is not replicated to a secondary during cluster replication. Previously replicas always had the default user and password (postgres, password)

Clients relying on the old error codes may need to update application code. Replication customers should ensure that readers to replicas are configured with the correct user name and password.

Merged PRs

dolt

• 11534: go: remotesrv: http: Authenticate incoming requests against the correct remote peer address.
  For authentication, this previously defaulted to localhost, while still applying authorization on the true remote peer address.
• 11528: branch: fix -d when the upstream does not resolve
  Deleting a branch with dolt branch -d could fail with branch not found when deleted from a remote. Dolt now compares against the local snapshot of the upstream, not the remote repository, and falls back to the current branch when its copy is gone.
  • Add env.UpstreamRef to resolve a branch's upstream local ref from BranchConfig.
  • Add upstreamOrHead to decide when to use the upstream branch or switch to current working branch.
  • Upstream is searched by the configured merge ref in the BranchConfig to support different names.
  • Merged predicate reads the remote-tracking ref instead of the live remote state.
  • RemoteDbProvider parameter dropped because it goes unused.
    Fix dolthub/dolt#11450
• 11486: Bug fixes for binlog replication with virtual columns
  Adds support for correctly handling virtual columns and functional indexes in binlog replication, both when Dolt is acting as the replication primary and when Dolt is acting as a replica.
  Fixes: dolthub/dolt#11475
• 11481: go: store/nbs,dprocedures_backup: Add dolt backup sync{,-url} --prune-with-grace-period.
  Backups into a destination store land new table files there before they update the manifest of the destination store to refer to the files. If the backup fails partially through, it can leave behind temporary files and unreferenced table files.
  --prune-with-grace-period is a new flag which can be used when calling backup sync and backup sync-url. It will cause all unreferenced table files in the destination store to be removed, before the sync starts, if no files in the destination store have an mtime more recent than the provided grace period.
  In general, --prune-with-grace-period should be larger than the amount of time it takes to land the backup, but smaller than the period of time between backup attempts.
  Fixes #11361
• 11448: refactored auth replication to support doltgres
  This change generalizes auth persistence to allow doltgres to participate in it with its own file type. The GRPC methods are unchanged, since they already take an opaque blob payload.
  Doltgres side: #3057

doltgresql

• 3132: {go.mod,go.sum,utils}: update builder to match method signature
• 3128: Fix array_to_string() panic on int2vector/oidvector values
  Caused by ArrayBaseType() returning the wrong element type.
  Fixes: #3108
• 3127: Fix IN clause failing on schema-qualified dolt system tables
  Dolt system tables without the dolt_ prefix were not getting wrapped with GMSCast elements correctly, which caused incorrect execution.
  Fixes: #3115
• 3104: new skipped test
• 3089: automatically translate hard-coded oids in the regression tests
  Some regression tests use hard-coded OIDs for entities created during tests. Since doltgres assigns its own OIDs, these need to be translated for these tests to pass.
• 3087: more function support
  Functionality required to stop special-casing psql commands, which now all work natively.
  In addition to some missing functions, this PR also implements a few other features / fixes required for full psql support:
  • Correctly report triggers from pg_class
  • Correctly compare OIDs when one is NULL
  • Support for WITH ORDINAL for table functions
• 3081: Handle apd.Decimal types in limit clauses
  Depends on: dolthub/go-mysql-server#3701
• 3073: audited error codes
  This PR audits error code processing to ensure that all common error codes typically examined by clients are correctly returned by the server. In particular, some clients interpret an XX000 error code by terminating the connection to the server, so we need to be careful to only return that code on a catastrophic, unexpected error.
  Also adds new client library tests that ensure clients behave as expected in response to these codes.
• 3059: Added CREATE OPERATOR and CREATE AGGREGATE
  This adds support for the CREATE OPERATOR and CREATE AGGREGATE statements, which we need to support additional emulated extensions.
• 3057: replicate the auth.db file to secondaries
  This closes a significant gap in replication.
  See related changes in dolthub/dolt#11448
• 3055: bug fix for index miss on dolt_diff_* tables

go-mysql-server

• 3703: Fix panic for LIKE expression with empty ESCAPE character
  Despite what MySQL documentation says, empty ESCAPE character actually escapes the NUL character (\0).
  fixes: dolthub/dolt#11518
• 3702: fix panic in format()
  fixes: dolthub/dolt#11517
• 3700: fix panic on rand(null)
  fixes: dolthub/dolt#11494
• 3697: fix inet6 type assertions
  fixes: dolthub/dolt#11493
• 3696: cap initial heapsize
  This PR adds a cap to the initial heap size of the maxRowHeap within TopNRowsIter, to prevent overflows.
  Additionally, unnecessarily large order by ... limit x clauses would result wasted time allocating space.
  heap.Push appends to the internal slice, so we'll rely on golang to manage the memory.
  fixes: dolthub/dolt#11503
• 3695: Allow NULL and other numerical swap flags for BIN_TO_UUID
  Fixes dolthub/dolt#11457
• 3693: fix panic for GeometryFromWKT and various GeoSpatial bugs
  We added support for "pure empty geometry forms", which MySQL doesn't support, except for "GEOMETRYCOLLECTION EMPTY".
  Changes:
  • fix panics for pure empty geometry forms
  • fix panics for invalid axis-order options
  • fix ordering for axis-order options
  • expose st_latitude function
  • fix geospatial srid for st_x() and st_y() functions
    Fixes: dolthub/dolt#11492
• 3692: When serializing bytes for a MySQL client, correctly handle wrapped values, such as adaptive encoded columns
  Currently, when a sql.Value contains a wrapped value, but the value's type doesn't implement sql.ValueType, we fail to properly serialize it when sending the bytes to the MySQL client.

Closed Issues

• 3108: array_to_string() panic with int2vector
• 3115: IN clause fails with dolt_branches
• 11517: Dolt panics on a numeric FORMAT locale from a window expression
• 11518: Dolt panics on empty LIKE ... ESCAPE '' within window expressions
• 11494: Dolt panics on RAND(NULL) on a nil seed.
• 11493: Dolt panics on INET6_ATON on a non-string SQL argument
• 11361: dolt backup sync: an interrupted sync leaves a complete, unreferenced archive in the destination forever — the archive is published before the manifest and there is no prune
• 11024: Incorrect expression evaluation when NULL appears in a tuple
• 11456: MySQL Binary Prepared Statement (COM_STMT_EXECUTE) truncates TCP packet payload on JSON columns containing multi-line formatted strings
• 11475: INSERT INTO a table with functional index enabled panics when binlog replication is active on dolt sql-server
• 11450: dolt branch -d does not find branch that exists when doing dolt branch [-a]
• 11503: ORDER BY + huge LIMIT (≥ 2^62) panics
• 11457: BIN_TO_UUID panics for a NULL swap flag
• 11492: Dolt panics on empty POLYGON WKT

1.2.0 Latest

Latest

Compare

Choose a tag to compare

1.1.0 has a breaking change in transaction semantics. Previous transaction semantics did not correctly emulate Postgres behavior.

• In 1.0.0 and earlier, multiple statements in a single Query message were each committed separately, as if they had arrived in individual Query messages. In 1.1.0, all statements in a single message are treated as an implicit transaction, with all statements being committed or rolled back atomically.
• In 1.0.0 and earlier, statements received via the extended query protocol (parse, bind, execute, etc.) used by many libraries were all committed independently unless wrapped in an explicit transaction with BEGIN. In 1.1.0, such statements are committed atomically when a Sync message is received.
• In 1.0.0 and earlier, statements that produced an error were treated as independent errors and additional statements in the same transaction could still succeed. In 1.1.0, any error in a transaction causes all subsequent statements to fail until a COMMIT or ROLLBACK is received.

This change is unlikely to break most customers, but in accordance with our release policy this change in behavior requires a minor version increment.

Merged PRs

doltgresql

• 3070: bug fix: implicit transactions
  Postgres's rules for handling implicit transactions are subtle and diverge dramatically from MySQL's. This PR implements the correct semantics for implicit transactions:
  • Multiple statements in a single query message execute in an implicit transaction
  • Queries using the extended query protocol all take place in an explicit transaction, with Sync triggering a commit
    There are lots of additional subtleties in this behavior, fully documented at https://www.postgresql.org/docs/current/protocol-flow.html
• 3069: Decorrelate EXISTS / NOT EXISTS queries
  Fix Doltgres's TypeSanitizer analyzer rule casting every *plan.ExistsSubquery node before GMS's unnestExistsSubqueries rule runs, which hid the node from that rule's pattern match and silently disabled EXISTS/NOT EXISTS decorrelation (semi/anti join rewrite) for every Doltgres query using them.
  Manually perf testing results (10k rows, indexed): a simple correlated EXISTS query dropped from ~18s to ~20ms, and a nested EXISTS/NOT EXISTS query that hadn't finished after 90s completed in ~2s, roughly three orders of magnitude on the query shapes this fixes.
• 3064: Fix goroutine deadlock from unlocked mutex during panic
  Fixes a deadlock in DoltgresType.IoOutput/ConvertToType where a panic while resolving a type's cast or output function (e.g. an index-out-of-range on an unexpected empty ResolvedTypes()) left the type's mutex locked
  forever, hanging every later query that touched the same type.
  Also adds a step in the CI workflow to check for signs of a timeout in the logs, and if so, errors with a message about a deadlock timeout.
• 3063: Accept transaction isolation levels, but always use REPEATABLE READ
  Follow Dolt's existing behavior to allow clients to specify a transaction isolation level, but still always use REPEATABLE READ.
  Also includes an update to the Doltgres SQL parser, to allow us to differentiate between specified isolation levels in the future.
• 3054: bug fix: SHOW support global settings
  A bug prevented SHOW from working with global settings.
• 3049: Fix panic in test query converter
  Fix a panic in the Doltgres' test query converter when processing DEFAULT values in ON DUPLICATE KEY UPDATE / UPDATE ... SET clauses.
  Fixes: #3045
• 3041: bug fix for nonlocal tables to work in Doltgres
  Depends on: dolthub/dolt#11406
  Fixes: #3035

Closed Issues

• 3036: Window functions returning the wrong results
• 3045: Doltgres panics on INSERT...ON DUPLICATE KEY UPDATE <col> = DEFAULT query
• 3035: non-local tables broken

1.1.0

Compare

Choose a tag to compare

This is the 1.0 release of Doltgres.

1.0 means:

1. Forward Storage Compatibility: all 1.x releases will be compatible with no data migrations.
2. Production Performance: Fast enough for your production workloads.
3. Postgres Compatibility: Your Postgres tools, libraries and queries will just work.
4. Stable Version Control Interface: Version control queries you write will work for all 1.x releases.

For more detail, read up on the release here:

https://dolthub.com/blog/2026-08-06-doltgres-1-0/

Merged PRs

doltgresql

• 3031: fix unknown type array input for any expr
• 3030: Fix RANGE window frame panics and bugs
  Fixes panics and silent mishandling when the frame boundary offset is numeric or an INTERVAL over a DATE/TIMESTAMP column.
  Depends on: dolthub/go-mysql-server#3665
• 3029: Add fast path for binary CompiledFuncs
  We can avoid memory allocations if we know the number of arguments we're going to process.
• 3026: Add native ntile(), nth_value(), and cume_dist() window functions
  New window functions for ntile(), nth_value(), and cume_dist() implemented in the Doltgres layer, matching Postgres's exact return types and semantics.
• 3024: Homogenize RootObject and Collection storage
  This simplifies root object collections by tying them all to use the same underlying storage, and also fixes a few long-standing issues:
  • We were aggressive with cache eviction, meaning collections were often cached for only a single statement, since we didn't have good guarantees that separate statements would properly invalidate a stale cache.
  • Simply loading a collection would write a different value into the root, as it went from an "unwritten collection" to an "empty collection", which is a material difference as far as the root is concerned. We tried to mitigate this, but it was a per-collection fix.
  • Concurrent branch modifications could cause a session to get into an error loop which we couldn't exit in the session, as we'd never clear the error from the previous statement and we'd resurface that error before we could even do anything
• 3023: Revamped extensions to use emulation
  This removes the old logic of looking at the locally-installed Postgres instance and directly using those DLLs, and replaces it with emulated versions. So far only uuid-ossp really worked in the previous version, and we're successfully emulating it in Go code now. We can follow this same model and implement the other extensions as well.
• 2986: Create a SequenceTracker for Postgres sequences.
  This is the Doltgres side of dolthub/dolt#11337
  The purpose of this PR is for Doltgres to have the same behavior for sequences and SERIAL columns that Dolt has for AUTO INCREMENT tables.
  In order to accomplish this, we introduce a SequenceState type that mirrors the AutoIncrementState in Dolt. The GlobalState has trackers for both that manage a state for each Sequence object.

Closed Issues

• 3034: 0.57.3 arm64 Docker image ships the 0.57.2 binary (dolt_version() returns 0.57.2)
• 3042: (withdrawn)
• 2240: support for xmin and other hidden system columns

1.0.0

Compare

Choose a tag to compare

Merged PRs

doltgresql

• 3021: avoid reallocating slice and sorting during typecoll name resolution
• 3020: use postgres error code
• 3018: Add regression tests for bindvar type inference with #>> operator
  Fixes: #3012
  Depends on: dolthub/go-mysql-server#3663
• 3015: Add support for ALL()
  The ALL() expression allows comparing a value with the results from a subquery or from an array, and testing how the value compares to all the values in that array or subquery result set.
  Fixes: #3013
• 3010: added OUT params to dolt_* system functions
  Currently, all dolt_* functions return []text as their output, e.g.:

  select * from dolt_checkout('-b', 'new-branch');
  dolt_checkout
  ---------------------------------------
  {0,"Switched to branch 'new-branch'"}
  (1 row)

  This is also true of functions that return only a single scalar results, e.g. {0}.
  This PR changes this behavior to give all dolt procedures named OUT params so that they return column multiple column values or a single scalar, as the case may be.

  select * from dolt_add('.');
  status
  --------
  0
  (1 row)
  select status from dolt_add('.');
  status
  --------
  0
  (1 row)
  select dolt_add('.');
  dolt_add
  ----------
  0
  (1 row)

  This is a breaking change, but will enable customers who want to inspect the output of dolt_ functions to be able to do so much more cleanly. It also matches the behavior of built-in postgres functions with multi-value return semantics, such as pg_stat_file. Finally, it enables us to allow customers to write postgres-native SELECT constructs such as select status from dolt_checkout()..., which is not possible with the []text return type.
  Examples with multiple return values:

  postgres=> select * from dolt_checkout('-b', 'new-branch2');
  status |             message
  --------+----------------------------------
  0 | Switched to branch 'new-branch2'
  (1 row)
  postgres=> select message from dolt_checkout('-b', 'new-branch3');
  message
  ----------------------------------
  Switched to branch 'new-branch3'
  (1 row)
  select dolt_checkout('-b', 'new-branch4');
  dolt_checkout
  ----------------------------------------
  (0,"Switched to branch 'new-branch4'")
  (1 row)

  This should also work but doesn't yet, and can be added in a followup:

  postgres=> select (dolt_checkout('-b', 'new-branch3')).message;
  ERROR:  column "message" not found in data type record (errno 1105) (sqlstate HY000)

• 3009: bug fix for joins on set-returning functions
• 3008: fixed tests that were missing pgcatalog info
• 3007: use given schema if exists as composite type for table parsed as column
• 3005: pointer comparison for DoltgresType.Equals()
  Compare the pointers to the DoltgresTypes before comparing their Serialized bytes, because they are the same most of the time.
• 3004: added missing functions for psql queries
• 2999: table statistics support
• 2995: commit generated files so doltgresql can be consumed as a dependency
  postgres/parser/build.sh generates five Go files that were gitignored, so they never reached the published module zip. Any consumer importing postgres/parser/... failed to build. This is a blocker for building a doltgresapi go service in DoltHub. This PR aims to make doltgres consumable as a library.
• 2972: support OUT parameter in functions
  VARIADIC mode is still not supported yet.

Closed Issues

• 3011: Querying a missing table returns SQLSTATE XX000 with a MySQL-shaped message instead of undefined_table (42P01)
• 3012: Parameter compared to jsonb #>> text[] is inferred as json; PostgreSQL infers text
• 3013: ALL (array) not supported — the standard cycle-check idiom in recursive CTEs fails
• 2979: Server panic in kvexec lookup-join when the join key is not the first non-PK column (0.57.0 + 0.57.1)

0.57.3

Compare

Choose a tag to compare

Merged PRs

doltgresql

• 2990: Flesh out remaining pg_catalog tables
• 2988: Regression test for bug fix in lookup join when using extended types
  Fixes: #2979
  Depends on: dolthub/dolt#11368
• 2987: Add regression tests for bug fix in CASE support for ExtendedType
  Depends on: dolthub/go-mysql-server#3656
  Related to: #2980
• 2983: Re-enable replication tests on MacOS
  GitHub MacOS runners can take a long time to do DNS lookups, even for localhost, which was the root cause of the races on that platform. All are now changed to the loopback IP.
  Also fixes an unrelated bug: connections that default to a DB with the same name as a user are now rejected if that DB doesn't exist.
• 2982: Allow aliasing columns in a table function
  Allows aliasing the column names of a table function, such as: SELECT * FROM generate_series(1,3) AS s(r).
  Also enables regproc to work with schema-qualified procedures.
  Both of these gaps were found while fixing a gap to enable DBeaver to work correctly with Doltgres.
  Depends on:
  • dolthub/vitess#477
  • dolthub/go-mysql-server#3652
• 2977: Hidden column support
  This adds partial support for hidden columns, in that they return empty values, but don't have the full internal mechanics backing them at this time.

Closed Issues

• 2979: Server panic in kvexec lookup-join when the join key is not the first non-PK column (0.57.0 + 0.57.1)
• 2980: CASE with mixed numeric/integer branches resolves to text — SUM errors, MAX returns silently wrong values (0.57.1)
• 2973: Silent data corruption: int4 parameter bound to BIGINT column is misread as int8 when the same statement contains an untyped NULL parameter (extended query protocol)

0.57.2

Compare

Choose a tag to compare