Merge upstream#7
Draft
adamcrown wants to merge 236 commits into
Draft
Conversation
Without the `USE...` in `TestExecuteMultiple` the test panics as the
result in `conn_test.go:108` is nil.
```
$ go test
----------------------------------------------------------------------
PANIC: conn_test.go:92: connTestSuite.TestExecuteMultiple
... Panic: runtime error: invalid memory address or nil pointer dereference (PC=0x438592)
/usr/lib/golang/src/runtime/panic.go:884
in gopanic
/usr/lib/golang/src/runtime/panic.go:260
in panicmem
/usr/lib/golang/src/runtime/signal_unix.go:837
in sigpanic
conn_test.go:108
in connTestSuite.TestExecuteMultiple.func1
conn.go:273
in Conn.ExecuteMultiple
conn_test.go:101
in connTestSuite.TestExecuteMultiple
/usr/lib/golang/src/reflect/value.go:370
in Value.Call
/home/dvaneeden/go/pkg/mod/github.com/pingcap/check@v0.0.0-20190102082844-67f458068fc8/check.go:836
in suiteRunner.forkTest.func1
/home/dvaneeden/go/pkg/mod/github.com/pingcap/check@v0.0.0-20190102082844-67f458068fc8/check.go:730
in suiteRunner.forkCall.func1
/usr/lib/golang/src/runtime/asm_amd64.s:1598
in goexit
----------------------------------------------------------------------
FAIL: conn_test.go:132: connTestSuite.TestExecuteSelectStreaming
conn_test.go:170:
c.Assert(err, IsNil)
... value *mysql.MyError = &mysql.MyError{Code:0x416, Message:"No database selected", State:"3D000"} ("ERROR 1046 (3D000): No database selected")
----------------------------------------------------------------------
FAIL: conn_test.go:86: connTestSuite.TestFieldList
conn_test.go:88:
c.Assert(err, IsNil)
... value *mysql.MyError = &mysql.MyError{Code:0x416, Message:"No database selected", State:"3D000"} ("ERROR 1046 (3D000): No database selected")
----------------------------------------------------------------------
FAIL: conn_test.go:38: connTestSuite.TearDownSuite
conn_test.go:43:
s.testExecute_DropTable(c)
conn_test.go:83:
c.Assert(err, IsNil)
... value *mysql.MyError = &mysql.MyError{Code:0x416, Message:"No database selected", State:"3D000"} ("ERROR 1046 (3D000): No database selected")
OOPS: 18 passed, 3 FAILED, 1 PANICKED
--- FAIL: Test (0.46s)
FAIL
exit status 1
FAIL github.com/go-mysql-org/go-mysql/client 0.465s
```
Support LOCAL INFILE request
Switch zstd library to allow build on arm64 and amd64
Bumps [golang.org/x/text](https://github.com/golang/text) from 0.3.6 to 0.3.8. - [Release notes](https://github.com/golang/text/releases) - [Commits](golang/text@v0.3.6...v0.3.8) --- updated-dependencies: - dependency-name: golang.org/x/text dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
client: wait connection producer exit in pool.Close()
…odules/golang.org/x/text-0.3.8 Bump golang.org/x/text from 0.3.6 to 0.3.8
client: fix ci
client: add close method for client pool
Signed-off-by: lance6716 <lance6716@gmail.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
…ql-org#1103) Bumps [filippo.io/edwards25519](https://github.com/FiloSottile/edwards25519) from 1.1.0 to 1.1.1. - [Commits](FiloSottile/edwards25519@v1.1.0...v1.1.1) --- updated-dependencies: - dependency-name: filippo.io/edwards25519 dependency-version: 1.1.1 dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Signed-off-by: lance6716 <lance6716@gmail.com>
* driver: export Connector for structured sql.OpenDB * driver: harden Connector option parsing * export Conn Signed-off-by: lance6716 <lance6716@gmail.com> --------- Signed-off-by: lance6716 <lance6716@gmail.com>
* Extends TableColumn with attribute flags for DEFAULT_GENERATED and ON UPDATE * schema: test DEFAULT_GENERATED and ON UPDATE flags --------- Co-authored-by: lance6716 <lance6716@gmail.com>
* Adds OnTableNotFound() error handler to Canal * Apply suggestions from code review Co-authored-by: Daniël van Eeden <github@myname.nl> --------- Co-authored-by: lance6716 <lance6716@gmail.com> Co-authored-by: Daniël van Eeden <github@myname.nl>
* driver: improve RE matching case Signed-off-by: lance6716 <lance6716@gmail.com> * driver: address review comments * driver: gofumpt * driver: clarify compat DSN comment --------- Signed-off-by: lance6716 <lance6716@gmail.com>
* Fix: incorrect offset in parseBinlogDumpGTID * add: replication_test.go * Update server/replication_test.go --------- Co-authored-by: lance6716 <lance6716@gmail.com>
…-mysql-org#1117) decodeValue now accepts an isUnsigned flag and returns the appropriate unsigned Go type (uint8/uint16/uint32/uint64) for TINY/SHORT/INT24/LONG/ LONGLONG columns. decodeImage derives the flag per column from the TableMapEvent UnsignedMap so values above the signed maximum are no longer misrepresented as negative numbers.
…_TYPE_DECIMAL (go-mysql-org#1119) * replication: fix IsNumericColumn to include MYSQL_TYPE_YEAR and MYSQL_TYPE_DECIMAL MySQL's has_signedess_information_type() writes a SignednessBitmap bit for MYSQL_TYPE_YEAR (and MYSQL_TYPE_DECIMAL) in TABLE_MAP_EVENT, but go-mysql's IsNumericColumn() did not include these types. This caused UnsignedMap() to silently skip the YEAR/DECIMAL bitmap bit, shifting every subsequent numeric column to the wrong signedness value (off-by-one bitmap desync). Fix: add MYSQL_TYPE_YEAR and MYSQL_TYPE_DECIMAL to IsNumericColumn() to match MySQL's has_signedess_information_type() definition from sql/field_common_properties.h. Add TestUnsignedMapWithYearColumn to reproduce the exact scenario described in the bug report (INT UNSIGNED → YEAR → INT UNSIGNED → DECIMAL UNSIGNED), and update TestTableMapHelperMaps to reflect that the t_year column now correctly appears in the UnsignedMap. Fixes: YEAR column between numeric columns causes all following columns to report wrong signedness in UnsignedMap(). Made-with: Cursor * chore: remove comment * fix: resolve review comments
…void duplicate domain id error (go-mysql-org#1122) * fix: MariadbGTIDSet.AddSet() replace server entry on same domain to avoid duplicate domain id error * refactor: remove explanatory comment from MariadbGTIDSet.AddSet() and TestMariaDBGTIDSetEqual() * docs: add clarifying comments to MariadbGTIDSet and AddSet --------- Co-authored-by: Sharath L P <sharathlp@microsoft.com>
…y per domain) (go-mysql-org#1124) * refactor: flatten MariadbGTIDSet to map[uint32]*MariadbGTID (one entry per domain) Fixes go-mysql-org#1123. MariaDB tracks one GTID position per domain_id. The nested map[domainID]map[serverID]*MariadbGTID allowed multiple server entries per domain, which is incorrect. Flatten to map[domainID]*MariadbGTID and use forward() to handle server ID changes within a domain. * docs: add comprehensive comments and tests for MariaDB GTID position semantics
…#1128) SetReadDeadline unblocks ReadPacket so onStream notices ctx cancellation. Some transports (notably SSH tunnels) refuse deadlines with an error, in which case close the underlying connection to force ReadPacket to return. Otherwise wg.Wait below parks forever when KILL also fails to reach server (e.g. thread already reaped, "ERROR 1094 unknown thread id") Co-authored-by: lance6716 <lance6716@gmail.com>
…on) (go-mysql-org#1125) * mysql: Parsing MySQL GTIDs with tags (second alternative implementation) Co-authored-by: lance6716 <lance6716@gmail.com> * Update based on review * Update based on review * Fix formatting * Move NewMysqlGTIDSet() next to type definition * Fix FuzzTag_MarshalBinary * Updates based on review * Updates based on review * mysql: preallocate slices in GTID set hot paths Preallocate slice capacity where the final size is known, to avoid incremental growslice reallocations and copy work in Clone, Encode, String, Normalize, and DecodeMysqlGTIDSet. Also adds an Equal test covering same-UUID/different-tag-count for full branch coverage. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * Simplify IntervalSlice.Equal() based on review * mysql: make IntervalSlice.Contain robust to unsorted sub The previous implementation walked s with a monotonic forward index, silently relying on sub also being sorted. An unsorted sub produced false negatives even when every interval was contained. Switch to a per-sub[i] binary search over s. s still must be sorted and normalized (the type's other operations enforce that invariant); sub may now be in any order. O(m log n), no allocation. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: lance6716 <lance6716@gmail.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Run go fix ./... * Add modernice CI check
* Enable revive * Use more idiomatic Go for builtin function overlap
* Update dependencies * Run go fix ./... * Remove Go 1.24 from CI * Fix issues from linters
* CHANGELOG: Update for v1.15 * Update CHANGELOG.md Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> * Add note for breaking changes * Update CHANGELOG.md Co-authored-by: lance6716 <lance6716@gmail.com> --------- Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: lance6716 <lance6716@gmail.com>
…#1138) MySQL 8.0.26 renamed rpl_semi_sync_master_enabled to rpl_semi_sync_source_enabled and the @rpl_semi_sync_slave session variable to @rpl_semi_sync_replica, keeping the old names as aliases. 8.4.0 removed the old aliases entirely, so semi-sync replication failed to negotiate against 8.4+ servers. Match either spelling in SHOW VARIABLES and set both session user variables, mirroring the approach already used for @master_binlog_checksum/@source_binlog_checksum and the heartbeat variables. Refs go-mysql-org#866 Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Co-authored-by: lance6716 <lance6716@gmail.com>
* replication: add RenderJSONAsMySQLText flag for type-preserving JSON output
Existing JSON binary decoding goes through an intermediate Go value
(float64, shopspring/decimal, map[string]any, ...) and then json.Marshal,
which silently drops JSONB type tags:
* JSONB_DOUBLE 1.0 -> float64(1.0) -> json.Marshal -> "1"
(re-parsed by MySQL as JSON INTEGER)
* JSONB_OPAQUE/NEWDECIMAL 1.0
-> Go string "1.0" -> json.Marshal -> "\"1.0\""
(re-parsed by MySQL as JSON STRING)
Replaying either of those into a MySQL JSON column changes the stored
binary representation, which breaks CRC32-style row checksums in binlog
replication / online schema change tools.
This adds an opt-in BinlogSyncerConfig.RenderJSONAsMySQLText flag that
routes JSON values through a renderer (replication/json_mysql_text.go)
which emits MySQL-text-compatible JSON directly from the JSONB byte
stream, preserving each value's original type tag.
Mirrors the existing UseDecimal and UseFloatWithTrailingZero opt-in
flags: default behaviour is unchanged, callers feeding the decoded
string back into a MySQL JSON column opt in.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* change implementation to use json_binary
* replication: satisfy modernize rangeint lint
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* replication: apply ignoreDecodeErr uniformly across both modes
Address PR review: drop the mysqlTextMode guard so IgnoreJSONDecodeError
behaves consistently for top-level short payloads regardless of the
rendering mode. Also fix a misleading DATE comment ("emitted" -> "omitted").
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* incorporate reviewer feedback
* incorporate reviewer feedback
* replication: clarify test-helper doc and tidy test name
- Rename TestNewTransactionPayloadEvent_InheritsParent to drop the
underscore, per review feedback.
- Reword the renderJSONAsMySQLText helper comment to be explicit that
production code never calls it.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* replication: drop underscores from new test function names
Per review feedback, rename all new tests in this PR to the no-underscore
TestSubjectVerbObject style used elsewhere in the package.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* refactor: replace interface{} with any (Go 1.18+)
* fix: apply gofumpt formatting to example_client_test.go
Resolves all merge conflicts while preserving fork-specific features: - MCE-18894: Load binlog position from Redis (MasterInfoLoader interface) - MCE-18519: Retriable sync with handleRetriableSync and syncErrorCh channel - MCE-18274: FsInfoLoader for file-based master info persistence - Custom column type handling (TYPE_TEXT, TYPE_BLOB) - instructure/mc-go-mysql import paths throughout Upstream changes incorporated: - Security dependency updates (edwards25519, klauspost/compress, etc.) - Replaced DataDog/zstd and siddontang/* with modern equivalents - New utils.Now() for testability, slog-based structured logging - New server authentication handler/provider architecture - MySQL 8.4 semi-sync variable support, MariaDB GTID improvements - EventCacheCount and FillZeroLogPos config options - Removed failover package (upstream deleted) - Various bug fixes: unsigned int decoding, binlog GTID parsing, etc. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is a basic merge of upstream, but the merge conflicts were significant, so I had Claude handle the conflicts. I'm leaving this as a draft until I have a chance to review and test it.
Conflicts resolved (45 files):
Fork features preserved:
Upstream improvements incorporated: