# Changelog

All versions are tagged by the major Postgres version, plus an individual semver for this library itself.

## 17-6.0.0   2024-09-27

* Upgrade to Postgres 17
* Deparser:
  - Add support for deparsing `JSON_TABLE`, `JSON_QUERY`, `JSON_EXISTS`, `JSON_VALUE`
  - Add support for deparsing `JSON`, `JSON_SCALAR`, `JSON_SERIALIZE`
  - Add support for deparsing `COPY ... FORCE_NULL(*)`
  - Add support for deparsing `ALTER COLUMN ... SET EXPRESSION AS`
  - Add support for deparsing `SET STATISTICS DEFAULT`
  - Add support for deparsing `SET ACCESS METHOD DEFAULT`
  - Add support for deparsing `... AT LOCAL`
  - Add support for deparsing `merge_action()`
  - Add support for deparsing `MERGE ... RETURNING`
  - Add support for deparsing `NOT MATCHED [ BY TARGET ]`

## 16-5.1.0   2024-01-08

* Add support for compiling on Windows
  - In order to build on Windows when using MSVC, use the new "Makefile.msvc"
    with nmake, or directly compile all .c files in the src folder into a library
  - If compiling directly, add src/postgres/include/port/win32 to the include path,
    and when using MSVC also add src/postgres/include/port/win32_msvc
* Add support for compiling on 32-bit systems
  - The relevant code is enabled at compile time by checking the pointer
    size (__SIZEOF_POINTER__ == 4)
* Move internal headers and included .c files to src/include folder
  - This avoids having any .c files in the top-level src/ folder that can't
    be directly compiled, and thus lets us simplify the logic for defining
    which source units are to be compiled.
* Remove strnlen fallback implementation
* Avoid use of mmap, asprintf and strndup to improve portability
* Improve compatibility with non-POSIX systems and C89 compilers

## 16-5.0.0   2023-12-22

* Update to Postgres 16.1
* Drop support for arbitrary trailing junk on integer literals
  - Support for parsing junk after parameters, e.g. `$1OR` is retained
* Deparser:
  - Fix deparsing of `SYSTEM_USER`
  - Add support for deparsing `STORAGE` mode
  - Add support for deparsing `REVOKE ... CASCADE`
  - Rework a_expr/b_expr/c_expr deparsing to match gram.y structure
  - Add support for deparsing `COMPRESSION` option for columns
  - Add support for deparsing `NULLS NOT DISTINCT` in unique constraints
  - Add support for deparsing new SQL/JSON functionality

## 15-4.2.4   2023-12-20

* Scanner: Add token `ASCII_36` ("$") to support queries like "SELECT $identifier" [#211](https://github.com/pganalyze/libpg_query/pull/211), [#219](https://github.com/pganalyze/libpg_query/pull/219)
  - Whilst these queries are not valid SQL and would fail parsing, this token can show up when using `pg_query_scan` or `pg_query_split_with_scanner` directly
* Normalize: Fix incorrect type cast [#223](https://github.com/pganalyze/libpg_query/pull/223)
* Deparser:
  - Fix some `BooleanTest` cases [#206](https://github.com/pganalyze/libpg_query/issues/206)
  - Fix `ALTER TABLE ... ENABLE TRIGGER ALL` [#222](https://github.com/pganalyze/libpg_query/pull/222)
  - Add parens around type casts in `IndexElem` [#214](https://github.com/pganalyze/libpg_query/pull/214)
  - Treat type casts in `func_expr_windowless` as functions [#214](https://github.com/pganalyze/libpg_query/pull/214)
* Support changing parse mode and config settings affecting the parser [#216](https://github.com/pganalyze/libpg_query/pull/216)
  - Alternate parse modes are useful for parsing PL/pgSQL expressions, as well as type names
  - Additionally, you can now change config settings that affect parsing, like `standard_conforming_strings`
  - To pass options, use the new methods ending in `_opts`, e.g. `pg_query_parse_opts`

## 15-4.2.3   2023-07-07

* Fix builds when compiling with `glibc >=  2.38` [#203](https://github.com/pganalyze/libpg_query/pull/203)
* Deparser: Add support for COALESCE and other expressions in LIMIT clause [#199](https://github.com/pganalyze/libpg_query/pull/199)

## 15-4.2.2   2023-07-07

* Deparser:
  - Add support for multi-statement CREATE PROCEDURE definitions [#197](https://github.com/pganalyze/libpg_query/pull/197)
  - Correctly quote identifier in ALTER TABLE ... ADD CONSTRAINT [x] [#196](https://github.com/pganalyze/libpg_query/pull/196)
  - Add support for index fillfactor within CREATE TABLE, fix SHOW ALL [#193](https://github.com/pganalyze/libpg_query/pull/193)
  - Move to dedicated file for easier inclusion in third-party projects [#192](https://github.com/pganalyze/libpg_query/pull/192)

## 15-4.2.1   2023-05-17

* Deparser: Handle INTERVAL correctly when used in SET statements [#184](https://github.com/pganalyze/libpg_query/pull/184)
* Deparser: Ensure index names are quoted as identifiers [#182](https://github.com/pganalyze/libpg_query/pull/182)
* Remove limits.h from pg_query_deparse.c [#181](https://github.com/pganalyze/libpg_query/pull/181)

## 15-4.2.0   2023-02-06

* Update copyright notice years and authors [#175](https://github.com/pganalyze/libpg_query/pull/175)
* Allow trailing junk in numeric literals [#177](https://github.com/pganalyze/libpg_query/pull/177)
  - Allows parsing queries like `SELECT * FROM a WHERE b=$1ORc=$2`
* NetBSD support [#172](https://github.com/pganalyze/libpg_query/pull/172/)

## 15-4.1.0   2022-12-28

* Add `Boolean` fingerprinting
  - `Boolean` nodes are now output during fingerprinting
* Fix parsing issue on 32-bit big endian machines
  - Now we use `size_t` to indicate the protobuf message size

## 15-4.0.0   2022-11-29

* Update to Postgres 15.1
* Add support for `MERGE` statements
* Add support for `ALTER TABLE ALL IN TABLESPACE ...` statements
* Add support for publication objects
  - E.g. `CREATE PUBLICATION foo FOR TABLES IN SCHEMA CURRENT_SCHEMA`
* Deparser now attempts to deparse `COPY` statements first using the old
  Postgres 8.4-style syntax (e.g. `COPY foo FROM STDIN FREEZE CSV`).

## 14-3.0.0   2022-11-17

Special thanks to @wolfgangwalther and @tlisanti for most of the work done on this release.

* Update to Postgres 14.6
* Drop support for `?` parameter syntax
* Update `fingerprint.json` to include newly added tests, regeneate tests

## 13-2.2.0   2022-11-02

* Update to Postgres 13.8 patch release [#156](https://github.com/pganalyze/libpg_query/pull/156)
* Backport Xcode 14.1 build fix from upcoming 13.9 release [#156](https://github.com/pganalyze/libpg_query/pull/156)
* Fingerprinting version 3.1 [#155](https://github.com/pganalyze/libpg_query/pull/155)
  - Fixes issue with "SELECT DISTINCT" having the same fingerprint as "SELECT"
    (fingerprints for "SELECT DISTINCT" will change with this revision)
  - Group additional DDL statements together that otherwise generate a lot of
    unique fingerprints (ListenStmt, UnlistenStmt, NotifyStmt, CreateFunctionStmt,
    FunctionParameter and DoStmt)
* Normalize additional DDL statements [#155](https://github.com/pganalyze/libpg_query/pull/155)
  - Normalizes arguments to CreateFunctionStmt, DoStmt,
    CreateSubscriptionStmt, AlterSubscriptionStmt, CreateUserMapping and
    AlterUserMapping.
  - Note that this is different from pg_stat_statements itself, which does
    not normalize utility statements at all today.


## 13-2.1.2   2022-06-28

* Add support for analyzing PL/pgSQL code inside DO blocks [#142](https://github.com/pganalyze/libpg_query/pull/142)
* Fix memory leak in pg_query_fingerprint error handling [#141](https://github.com/pganalyze/libpg_query/pull/141)


## 13-2.1.1   2022-05-03

* PL/pgSQL parser
  - Add support for Assert [#135](https://github.com/pganalyze/libpg_query/pull/135)
  - Add support for SET, COMMIT, ROLLBACK and CALL [#130](https://github.com/pganalyze/libpg_query/pull/130)
* Add support for parsing more operators that include a `?` character (special cased to support old pg_stat_statements query texts)
  - ltree extension [#136](https://github.com/pganalyze/libpg_query/pull/136)
  - promscale extension [#133](https://github.com/pganalyze/libpg_query/pull/133)
* Deparser improvements
  - Prefix errors with "deparse", and remove some asserts [#131](https://github.com/pganalyze/libpg_query/pull/131)
  - Fix potential segfault when passing invalid protobuf (RawStmt without Stmt) [#128](https://github.com/pganalyze/libpg_query/pull/128)


## 13-2.1.0   2021-10-12

* Normalize: add funcname error object [#121](https://github.com/pganalyze/libpg_query/pull/121)
* Normalize: Match GROUP BY against target list and re-use param refs [#124](https://github.com/pganalyze/libpg_query/pull/124)
* PL/pgSQL: Setup namespace items for parameters, support RECORD types [#123](https://github.com/pganalyze/libpg_query/pull/123)
  - This significantly improves parsing for PL/pgSQL functions, to the extent
    that most functions should now parse successfully


## 13-2.0.7   2021-07-16

* Normalize: Don't modify constants in TypeName typmods/arrayBounds fields (#118)
  - This matches how pg_stat_statement behaves, and avoids causing parsing
    errors on the normalized statement
* Don't fail builds on systems that have strchrnul support (FreeBSD)


## 13-2.0.6   2021-06-29

* Normalize: Don't touch "ORDER BY 1" expressions, keep original text [#115](https://github.com/pganalyze/libpg_query/pull/115)
  - This avoids obscuring the semantic meaning of integers in the ORDER BY
    clause, which is to reference a particular column in the target list.


## 13-2.0.5   2021-06-24

* Update to Postgres 13.3 patch release [#114](https://github.com/pganalyze/libpg_query/pull/114)
* Add optional Makefile target to build as shared library [#100](https://github.com/pganalyze/libpg_query/pull/100)
* Normalize: Don't touch "GROUP BY 1" type statements, keep original text [#113](https://github.com/pganalyze/libpg_query/pull/113)
  -  This avoids obscuring the semantic meaning of integers in the GROUP BY
    clause, which is to reference a particular column in the target list.
* Fingerprint: Cache list item hashes to fingerprint complex queries faster [#112](https://github.com/pganalyze/libpg_query/pull/112)
  - This was exhibiting quite bad runtime behaviour before, causing both an
    explosion in memory, as well as very high CPU runtime for complex queries.
  - The new approach ensures we don't calculate the hashes for a particular
    list more than once, which ensures that we roughly have quadratic runtime
    instead of exponential runtime.
* Deparser: Emit the RangeVar catalogname if present [#105](https://github.com/pganalyze/libpg_query/pull/105)
* Fix crash in pg_scan function when encountering backslash escapes [#109](https://github.com/pganalyze/libpg_query/pull/109)
* Integrate oss-fuzz fuzzer [#106](https://github.com/pganalyze/libpg_query/pull/106)


## 13-2.0.4   2021-04-05

* Deparser: Fix crash in CopyStmt with HEADER or FREEZE inside WITH parens
  - The parse tree does not contain an explicit argument in those cases,
    but does when specified in the legacy mode without the wrapping WITH.
  - With this change we only output the "1" argument when the original tree
    also had this, to ensure parse tree comparisons match. Note the intent
    here is technically the same, which is to enable these options.


## 13-2.0.3   2021-04-02

* Normalize: Fix handling of two subsequent DefElem elements [#96](https://github.com/pganalyze/libpg_query/pull/91)
  - We were incorrectly adding too many DefElem locations to the recorded
    constant values, causing a crash when more than a single DefElem
    is present in a utility statement.


## 13-2.0.2   2021-03-30

* srcdata/nodetypes.json: Avoid bogus values accidentally parsed from inside comments


## 13-2.0.1   2021-03-30

* Fix ARM builds: Avoid dependency on cpuid.h header
* Simplify deparser of TableLikeClause [#91](https://github.com/pganalyze/libpg_query/pull/91) [Lele Gaifax](https://github.com/lelit)
* Fix asprintf warnings by ensuring _GNU_SOURCE is set early enough


## 13-2.0.0   2021-03-18

* Update to PostgreSQL 13 parser (13.2 release)
* Changes to JSON output format
  - WARNING: These JSON format changes are incompatible with prior releases.
  - New top-level result object that contains the Postgres version number the
    parser is based on
  - Node type names are only output when the field is a generic field (Node*),
    but not when the field always has the same type. This matches how the
    Postgres source looks like, and ensures the JSON and (new) Protobuf format
    match in their structure. You can utilize the `srcdata/struct_defs.json`
    file as needed to get the necessary context on field types.
  - Whitespace between control characters in JSON is no longer added
  - "<" and ">" characters are escaped to avoid browser HTML injections
  - Enum values are output with the value's name, instead of the integer value
* Introduce new Protobuf parse tree output format
  - Up until now, this library relied on JSON to pass the parse result back
    to the caller, which has a number of downsides, most importantly that
    we don't have a readily available parser for JSON thats not tied to a
    running Postgres server. That in turn makes it hard to provide
    cross-language features such as deparsing directly in this library
    (which would require reading back a parse tree that gets passed in).
  - Protobuf isn't perfect, but its straightforward enough to generate the
    schema definitions for the parse tree nodes, and output the tree using
    a bundled C protobuf library, which has a small enough SLOC count (~3k)
    to not be noticeable in the big picture.
* Add support for returning Postgres scanner result
  - This allows utilizing pg_query for use cases that need the raw token
    information, instead of a parse tree. Due to additional modifications
    to the Postgres source, this also contains information about comments
    in the query string, and their location.
* Add deparsing functionality that turns parse tree back into a SQL query
  - This is based on the deparser that was written over multiple years for
    the pg_query Ruby library, and is now accessible for all bindings through
    this new API and implementation.
* Fingerprinting: Introduce v3 version and 64-bit XXH3 hash
  - See full details in the wiki page here: https://github.com/pganalyze/libpg_query/wiki/Fingerprinting#version-30-based-on-postgresql-13
* Add new pg_query_split_with_scanner/pg_query_split_with_parser functions to
  split up multi-statement strings
  - Naively one could assume that splitting a string by ";" is sufficient,
    but it becomes tricky once one takes into consideration that this
    character can also show up in identifier, constants or comments.
  - We provide both a parser-based split function and a scanner-based split
    function. Most importantly when splitting statements in a file that may
    contain syntax errors that cause a parser error, but are accepted by the
    scanner. Otherwise the parser-based split function is recommended
    due to better accuracy.
* Add experimental Protobuf C++ outfuncs, converge JSON output to match Protobuf
  mapped output
* Extract source with USE_ASSERT_CHECKING enabled
  - This ensures we have the necessary functions to compile an
    assert-enabled build if necessary. Note that this doesn't mean that
    asserts are enabled by default (they are not, you need to explicitly
    use DEBUG=1).
* Ensure codebase has a clean Valgrind run
* PL/pgSQL: Output NEW/OLD variable numbers, record dno fields [Ethan Resnick](https://github.com/ethanresnick)
* Makefile: Allow passing in customized CFLAGS/PG_CONFIGURE_FLAGS/TEST_* [Ethan Resnick](https://github.com/ethanresnick)


## 10-1.0.5   2021-02-17

* Update to latest Postgres 10 patch release (10.16)
* Free Postgres top-level memory context on thread exit / with function
  - Previously there was no way to free the top-level Postgres memory context,
    causing threaded programs that churn through a lot of threads to leak
    memory with each newly initialized thread-local top-level memory context.
  - Instead, this uses a newly introduced cleanup method to free the memory
    when a pthread exits (note this causes a pthread dependency to be added
    to this library). In addition, primarily for memory testing purposes, add a
    new method "pg_query_exit" that performs the same cleanup on demand.
* Resolve correctness issues and possible memory leak in PL/pgSQL parser
* Add arch-ppc.h for PPC architectures [#80](https://github.com/pganalyze/libpg_query/pull/80) [@pkubaj](https://github.com/pkubaj)


## 10-1.0.4   2020-12-27

* Update to latest Postgres 10 patch release (10.15)
* PL/pgSQL parsing: Handle asprintf failures (and prevent compiler warning)


## 10-1.0.3   2020-11-07

* Update to latest Postgres 10 patch release (10.14)
* Add support for ARM builds by explicitly copying ARM header file
* Ignore return value of asprintf without compiler warnings [@herwinw](https://github.com/herwinw)
* Free stderr_buffer when parsing plpgsql [@herwinw](https://github.com/herwinw)


## 10-1.0.2   2018-05-18

* Avoid compiler warning due to unused result in pg_query_parse_plpgsql


## 10-1.0.1   2017-10-31

* Fix fingerprint tests


## 10-1.0.0   2017-10-30

* First release based on PostgreSQL 10.0
  - Parse tree output may have changed in backwards-incompatible ways!
  - Fingerprint base version bumped to "02" to reflect the change in parse trees


## 9.5-1.7.1    2017-10-26

* Allow "$1 FROM $2" to be parsed
  - This is new with Postgres 10 output of pg_stat_statements, so we should
    treat this the same as "? FROM ?" in earlier versions.


## 9.5-1.7.0    2017-10-20

* Update to Postgres 9.5.9
* Support gcc versions earlier than 4.6.0
* Export version information in pg_query.h directly


## 9.5-1.6.2    2017-07-30

* Fingerprinting Version 1.3
  * Attributes to be ignored:
    * RangeVar.relname (if node also has RangeVar.relpersistence = "t")
  * Special cases: List nodes where parent field name is valuesLists
    * Follow same logic described for fromClause/targetList/cols/rexpr


## 9.5-1.6.1    2017-07-29

* Fingerprinting Version 1.2
  * Ignore portal_name in DeclareCursorStmt, FetchStmt and ClosePortalStmt


## 9.5-1.6.0    2017-07-29

* Change normalization methods to output $1/$2 .. $N instead of ? characters
  * BREAKING CHANGE in pg_query_normalize(..) output
  * This matches the change in the upcoming Postgres 10, and makes it easier
    to migrate applications to the new normalization format


## 9.5-1.5.0    2017-05-26

* Update to Postgres 9.5.7
* Disable 128-bit integer support (they are not used), to support 32-bit archs [@herwinw](https://github.com/herwinw)
* Cleanup efforts [@herwinw](https://github.com/herwinw)
  * Improve concurrency tests
  * Make sure we have a valid proc_source
  * Normalized whitespace in pg_query_parse_plpgsql
  * Move inclusion of stdio.h in plpgsql parser


## 9.5-1.4.2    2016-12-06

* Cut off fingerprints at 100 nodes deep to avoid excessive runtimes/memory
* Fix warning on Linux due to missing asprintf include


## 9.5-1.4.1    2016-06-26

* Automatically call pg_query_init as needed to ease threaded usage


## 9.5-1.4.0    2016-06-26

* Clean up includes to avoid dependency on stdbool.h and xlocale.h
* Change PL/pgSQL input to be the full CREATE FUNCTION statement
  * This is necessary for parsing, since we need the argument and return types
* Fingerprinting Version 1.1
  * Only ignore ResTarget.name when parent field name is targetList *and*
    we have a SelectStmt as a parent node (fixes UpdateStmt fingerprinting)
* Normalize the password in ALTER ROLE ... PASSWORD '123' statements
* Make library thread-safe through thread-local storage [#13](https://github.com/pganalyze/libpg_query/issues/13)


## 9.5-1.3.0    2016-05-31

* Extract source code using LLVM instead of manually compiling the right objects
  * This speeds up build times considerably since we don't download the Postgres
    source anymore, instead shipping a partial copy created as part of a release.
* Experimental support for parsing PL/pgSQL source code (output format subject to change)


## 9.5-1.2.1    2016-05-17

* Make sure we encode special characters correctly in JSON output ([@zhm](https://github.com/zhm) [#11](https://github.com/pganalyze/libpg_query/pull/11))


## 9.5-1.2.0    2016-05-16

* Fix stack overflow when parsing CREATE FOREIGN TABLE ([#9](https://github.com/pganalyze/libpg_query/issues/9))
* Update to PostgreSQL 9.5.3


## 9.5-1.1.0    2016-04-17

* Add pg_query_fingerprint() method that uniquely identifies SQL queries,
  whilst ignoring formatting and individual constant values
* Update to PostgreSQL 9.5.2


## 9.5-1.0.0    2016-03-06

* First release based on PostgreSQL 9.5.1
* Make JSON_OUTPUT_V2 the default and remove outfuncs patch
  * NOTE: This is a backwards incompatible change in the output parsetree format!


## 9.4-1.0.0    2016-03-06

* First tagged release based on PostgreSQL 9.4.5