This PR fixes two bugs in `CREATE VIRTUAL TABLE` (see individual commits
for details).
I added a test in `extensions.py` instead of the TCL test suite because
it's currently more convenient - there’s no existing framework for
loading extensions in TCL tests. However, I believe the test should
eventually be moved to TCL, as it verifies behavior expected to be
compatible with SQLite and is independent of any specific extension.
I've seen a PR proposing to migrate TCL tests to Rust, so please let me
know if moving this test to TCL would still be valuable.
Reviewed-by: Preston Thorpe (@PThorpe92)
Closes#1471
While debugging another issue for #1469, I noticed that the symbols
aren't being properly referenced for `getError` for either limboRows or
limboStmt types.
Closes#1470
After reading #1123, I wanted to see what optimizations I could do.
Sqlite optimizes `count` aggregation for the following case: `SELECT
count() FROM <tbl>`. This is so widely used, that they made an
optimization just for it in the form of the `COUNT` opcode.
This PR thus implements this optimization by creating the `COUNT`
opcode, and checking in the select emitter if we the query is a Simple
Count Query. If it is, we just emit the Opcode instead of going through
a Rewind loop, saving on execution time.
The screenshots below show a huge decrease in execution time.
- **Main**
<img width="383" alt="image" src="https://github.com/user-
attachments/assets/99a9dec4-e7c5-41db-ba67-4eafa80dd2e6" />
- **Count Optimization**
<img width="435" alt="image" src="https://github.com/user-
attachments/assets/e93b3233-92e6-4736-aa60-b52b2477179f" />
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com>
Closes#1443
closes#1449
A quick `explain` will show the root of the problem:
vs `sqlite3`
We process the columns in the order they need to be to create the
record, and although we store the index of the value meant to be
inserted into that column based on the order they are bound, we still
lose the original ordering.
## EDIT: Fixed
### Multi-row insert:
## Solution:
We just needed to traverse the insert values beforehand and create an
array of N integers each representing the index of each
`ast::Expr::Variable`: then later on we can lookup the `value_index` in
that array, and grab the index of matching integer to get the
parameters index to be bound to the relevant variable.
(yes I know that sounds like a leetcode problem)
Thanks to @jnesss for the write-up for this bug
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com>
Closes#1459
Fixes a runtime panic caused by failing to halt the virtual machine
after executing CREATE VIRTUAL TABLE IF NOT EXISTS.
Previously resulted in:
thread 'main' panicked at core/vdbe/mod.rs:408:52:
index out of bounds: the len is 3 but the index is 3
This commit adds suport for DROP INDEX.
Bytecode produced by this commit differs from SQLITE's bytecode, main
reason we don't do autovacuum or repacking of pages like SQLITE does.
Closes#1280Closes#1444
## Changes
Since limbo now supports UPDATE and DELETE, I'm planning to implement
the rest of the features for integrating JDBC and limbo.
Since limbo's [total_changes() function seems to work differently in
compared to the sqlite's](https://discord.com/channels/12586588262579610
20/1321869557459058778/1368830400985563176), let's remove the
`@Disabled` annotation from
`execute_update_should_return_number_of_updated_elements` test after
the issue is handled.
## Reference
https://github.com/tursodatabase/limbo/issues/615Closes#1451
This pull request includes a small change to the
`bindings/wasm/package.json` file. The change adds a `types` field
pointing to the TypeScript declaration file for the package (`[bindings/
wasm/package.jsonR15](diffhunk://#diff-
1f41234b939ba148924b9bbfedd557853ebcd22351a9c300e568ce7af0291babR15)`).
Closes#1460
This PR implements basic transaction support in the Limbo Go adapter by
adding the required methods to fulfill the `driver.Tx` interface.
## Changes
- Add `Begin()` method to `limboConn` to start a transaction
- Add `BeginTx()` method with context support and proper handling of
transaction options
- Implement `Commit()` method to commit transaction changes
- Implement `Rollback()` method with appropriate error handling
- Add transaction tests
## Implementation Details
- Uses the standard SQLite transaction commands (BEGIN, COMMIT,
ROLLBACK)
- Follows the same pattern as other SQL operations in the adapter
(prepare-execute-close)
- Maintains consistent locking and error handling patterns
## Limitations
- Currently, ROLLBACK operations will return an error as they're not yet
fully supported in the underlying Limbo implementation
- Only the default isolation level is supported; all other isolation
levels return `driver.ErrSkip`
- Read-only transactions are not supported and return `driver.ErrSkip`
## Testing
- Added basic transaction tests that verify BEGIN and COMMIT operations
- Adjusted tests to work with the current Limbo implementation
capabilities
These transaction methods enable the Go adapter to be used in
applications that require transaction support while providing clear
error messages when unsupported features are requested. I'll add to it
when Limbo supports ROLLBACK and/or additional isolation levels.
Closes#1435
Probably this is larger than it should, but this PR:
- Enhance error handling (instead of `unwraps`, napi's `GenericFailure`
with custom message)
- Add missing methods and properties for Database and Statements to turn
explicit what should be done
- Implement `all()` and `run()`* methods for `Statement`
- Bump napi version to the latest stable
- Implement `DatabaseStorage::sync()` for DatabaseFile
- Add conversion from js values to limbo values
\* `run()` isn't 100% compatible yet for two reasons:
1. The function API isn't variadic -- I chatted with napi's maintainer
in Discord and he said that this isn't possible using only napi (in this
version), and we could normalize "[...] parameters at the JavaScript
layer before passing them to Rust". Something similar to [this](https://
github.com/rolldown/rolldown/blob/main/packages/rolldown/src/utils/bindi
ngify-input-options.ts), which I plan to do in another PR.
2. better-sqlite version returns a result object, which isn't currently
supported by core (AFAIK), I also plan to do this in another PR.
Closes#1464
PR #1442 added support for using time.Time as query parameters.
Scanning time.Time values from query results still fails:
`sql: Scan error on column index 4, name "mod_time": unsupported Scan,
storing driver.Value type string into type *time.Time`
This change modifies the `toGoValue` function to detect RFC3339
formatted datetime strings and convert them back to time.Time objects
when reading from the database. This provides complete roundtrip support
for time.Time values.
Also added boolean support by converting Go bool values to integers (0
for false, 1 for true) when binding parameters, following SQLite's
convention for representing boolean values.
Reviewed-by: Preston Thorpe (@PThorpe92)
Closes#1465
Go 1.23.9 introduced a change with to it's linker that caused a
`duplicate symbol` error with purego 1.82
this is the recommended fix per
https://github.com/golang/go/issues/73617Closes#1466
