255 commits

Author	SHA1	Message	Date
Preston Thorpe	231282a5be	Merge 'Simulator Roadmap' from Alperen Keleş ... This PR is a working doc on a roadmap for the simulator. @pedrocarlo @LeMikaelF please take a look. Reviewed-by: Preston Thorpe <preston@turso.tech> Closes #3954	2025-12-05 13:27:57 -05:00
PThorpe92	a9fc469b3a	Update COMPAT.md	2025-12-02 15:42:41 -05:00
Nikita Sivukhin	b08374be46	support table_xinfo pragma	2025-11-20 17:38:10 +04:00
alpaylan	4aac076400	wip: add simulator roadmap	2025-11-13 14:25:17 -05:00
PThorpe92	b443b09516	Remove VRollback and VCommit as they are unused opcodes in sqlite	2025-11-09 11:27:09 -05:00
PThorpe92	40f7ddb28e	Add VBegin, VRename, VRollback and VCommit opcodes to COMPAT.md	2025-11-09 11:07:42 -05:00
Jussi Saurio	ca70df21ac	Update COMPAT.md	2025-10-28 13:11:12 +02:00
Jussi Saurio	171bcd83ec	COMPAT.MD: note about WINDOW functions	2025-10-13 10:16:13 +03:00
PThorpe92	b40e784903	Update COMPAT.md, add fk related opcodes	2025-10-07 16:22:15 -04:00
Preston Thorpe	d44a1f8bca	Merge 'Implement json_tree' from Mikaël Francoeur ... This PR implements the `json_tree` table-valued function. It's not 100% compatible with SQLite, because SQLite does some iffy things with the `key` and `path` columns. I started a [thread](https://www.sqlite.org/forum/forumpost/48f5763d8c) on their forum and I linked it to the disabled tests in `json.test`. Reviewed-by: Preston Thorpe <preston@turso.tech> Closes #3256	2025-09-25 16:45:44 -04:00
Pavan-Nambi	49d5141f2d	Merge remote-tracking branch 'origin/main' into cdc_fail_autoincrement	2025-09-24 18:06:02 +05:30
PThorpe92	7dccff0bee	Update COMPAT.md, add sequence and sequencetest opcodes	2025-09-23 22:32:22 -04:00
Mikaël Francoeur	8cc206e748	compat	2025-09-23 14:23:42 -04:00
Mikaël Francoeur	3e915d9868	implement json_tree	2025-09-23 14:22:02 -04:00
Pavan-Nambi	0de92cd7fb	update compat.md that we have memmax	2025-09-20 10:54:59 +05:30
Piotr Rzysko	b911e80607	Add AggValue instruction ... Adds the AggValue instruction, which computes the current aggregate result and writes it to a dedicated destination register. Unlike AggFinal, it does not overwrite or clear the accumulator register. This makes it possible to retrieve aggregate results multiple times—needed when processing window functions—while preserving the accumulator state.	2025-09-13 10:49:14 +02:00
Piotr Rzysko	867bef55d8	Add ResetSorter instruction ... This instruction isn't used yet, but it will be needed for window functions, since they heavily rely on ephemeral tables.	2025-09-13 10:44:56 +02:00
Piotr Rzysko	ea9599681e	Add OpenDup instruction ... The instruction isn’t used yet, but it’ll be needed for window functions, since they heavily rely on ephemeral tables.	2025-09-13 10:35:33 +02:00
Pekka Enberg	c620a15a55	Merge 'core: Support ceiling modifier in datetime' from Ceferino Patino ... Resolves #2677 - Implements the modifier for Floor on the datetime object. - Implements the modifier for Ceiling on the datetime object. - Includes additional testing changes in testing/scalar-functions- datetime.test to test the sql functionality. Consolidation PR of #2678 and #2679 since the functions ended up being much more complicated than I initially thought and had to be done in a single PR. Closes #2757	2025-09-10 14:46:07 +03:00
Mikaël Francoeur	e6d3d6ea54	1-arg json_each implementation	2025-09-05 14:47:40 -04:00
Pekka Enberg	12cf4d2e72	core: Make strict schema support experimental ... It's not tested properly so let's mark it as experimental for now. Fixes #2775	2025-09-02 16:40:02 +03:00
PThorpe92	761da801e8	Add ON CONFLICT DO to COMPAT.md	2025-08-29 20:58:44 -04:00
PThorpe92	8257496411	Update COMPAT.md to remove CREATE INDEX default disabled	2025-08-29 20:44:09 -04:00
Pekka Enberg	44ed4d562f	core: Initial pass on synchronous pragma ... This adds support for "OFF" and "FULL" (default) synchronous modes. As future work, we need to add NORMAL and EXTRA as well because applications expect them.	2025-08-28 16:02:41 +03:00
C4 Patino	9bda897118	chore: marked datetime ceiling modifier as completed in COMPAT.md	2025-08-23 22:59:08 -05:00
Jussi Saurio	d7186c7d7b	Merge 'Add support for unlikely(X)' from bit-aloo ... Implements the unlikely(X) function. Removes runtime implementations of likely(), unlikely() and likelihood(), replacing them with panics if they reach the VDBE. Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com> Closes #2559	2025-08-14 10:56:27 +03:00
Glauber Costa	5ab6f78f6b	Implement views ... Views (non materialized) are relatively simple, since they are just query aliases. We can expand them as if they were subqueries.	2025-08-13 14:14:03 -05:00
bit-aloo	71525c90e5	update compat.md	2025-08-12 16:27:58 +05:30
bit-aloo	e165deb09d	update compact md	2025-08-11 10:03:58 +05:30
Pekka Enberg	0f9d0cf519	Merge branch 'main' into 2025-08-07-add-query-only-pragma	2025-08-08 07:41:38 +03:00
bit-aloo	ee483ce4e9	update compat.md with query_only pragma	2025-08-07 23:50:18 +05:30
Glauber Costa	071330a739	implement the JournalMode vdbe instruction ... We do this already, but not through any opcode. Move it to an opcode for compatibility reasons.	2025-08-06 19:30:19 -05:00
Glauber Costa	f36974f086	implement the MaxPgCount opcode ... It is used by the pragma max_page_count, which is also implemented.	2025-08-06 13:20:15 -05:00
Jussi Saurio	aaa9ed1d9f	Merge 'Add regexp capture' from bit-aloo ... This PR adds RegExp Capture to regexp module Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com> Closes #2465	2025-08-06 12:13:19 +03:00
Glauber Costa	d1be7ad0bb	implement the collseq bytecode instruction ... SQLite generates those in aggregations like min / max with collation information either in the table definition or in the column expression. We currently generate the wrong result here, and properly generating the bytecode instruction fixes it.	2025-08-05 13:49:04 -05:00
bit-aloo	eff096273c	update compat.md	2025-08-05 20:54:55 +05:30
Glauber Costa	0506da70ed	more compat police ... * Affinity is already present * InsertInt is not a thing * String is never generated directly, it is a second-execution optimization for String8 so the size doesn't have to be recomputed, but we always store the size anyway.	2025-07-31 10:46:12 -05:00
Glauber Costa	9e8ba5263b	Implement the AddImm opcode ... It is a simple opcode. The hard part was finding a sqlite statement that uses it =)	2025-07-31 08:08:07 -05:00
PThorpe92	07137c7aaf	Merge 'Implement the Cast opcode' from Glauber Costa ... Our compat matrix mentions a couple of opcodes: ToInt, ToBlob, etc. Those opcodes do not exist. Instead, there is a single Cast opcode, that takes the affinity as a parameter. Currently we just call a function when we need to cast. This PR fixes the compat file, implements the cast opcode, and in at least one instance, when explicitly using the CAST keyword, uses that opcode instead of a function in the generated bytecode. Reviewed-by: Preston Thorpe (@PThorpe92) Reviewed-by: Preston Thorpe (@PThorpe92) Closes #2352	2025-07-30 22:32:09 -04:00
Glauber Costa	caec3f7c51	remove non-existent opcode ... $ egrep -rI "define OP" sqlite3.c | grep Cookie sqlite3.c:#define OP_ReadCookie 99 sqlite3.c:#define OP_SetCookie 100	2025-07-30 20:50:00 -05:00
Glauber Costa	4bd1582e7d	Implement the Cast opcode ... Our compat matrix mentions a couple of opcodes: ToInt, ToBlob, etc. Those opcodes do not exist. Instead, there is a single Cast opcode, that takes the affinity as a parameter. Currently we just call a function when we need to cast. This PR fixes the compat file, implements the cast opcode, and in at least one instance, when explicitly using the CAST keyword, uses that opcode instead of a function in the generated bytecode.	2025-07-30 20:44:54 -05:00
bit-aloo	6f45cb9770	update vector slice to adhere to start..end and not start, length	2025-07-30 14:10:43 +05:30
bit-aloo	96a99ca48a	rename subvector to vector_slice	2025-07-30 13:34:49 +05:30
bit-aloo	b5d46803b5	update compact.md with vector concat and subvector listings	2025-07-30 09:51:08 +05:30
Pekka Enberg	a2a31a520f	Remove CreateTable from COMPAT.md ... There is no such opcode.	2025-07-27 09:12:31 +03:00
Pekka Enberg	6bf6cc28e4	Merge 'Implement the Returning statement for inserts and updates' from Glauber Costa ... They are very similar. DELETE is very different, so that one we'll do it later. Closes #2276	2025-07-27 09:11:16 +03:00
PThorpe92	60c37eb036	Merge 'implement the pragma encoding' from Glauber Costa ... Do not allow setting it. That ship has sailed around 2005. Reviewed-by: Preston Thorpe (@PThorpe92) Closes #2282	2025-07-26 21:31:27 -04:00
PThorpe92	22df76d37d	Merge 'compat police' from Glauber Costa ... All those opcodes seem implemented. Reviewed-by: Preston Thorpe (@PThorpe92) Closes #2284	2025-07-26 21:28:28 -04:00
Glauber Costa	421ab44cf2	compat police ... All those opcodes seem implemented.	2025-07-26 20:06:21 -05:00
Glauber Costa	f0b1abd567	remove upsert statement from COMPAT.md ... There's no such statement in SQLite.	2025-07-26 19:39:32 -05:00