Pekka Enberg comments

Results 454 comments of


                                            Pekka Enberg

Preparing a statement is slower than with SQLite

I assume the `Parser::new()` associate function does all of the parsing so replacing that line means replacing the parser.

JSON function support

@madejejej I think it's best to first look at how SQLite handles that and copy if possible :)

Simulator improvements

Duplicate of the newer https://github.com/tursodatabase/limbo/issues/1844

Make I/O completion explicit

@arpit-saxena You absolutely can. Feel free to drop by our Discord if you want to chat about it

`ORDER BY` is much slower compared to SQLite

Likely https://github.com/penberg/limbo/issues/203

`ORDER BY` is much slower compared to SQLite

No significant difference as of e1c8c676ca21490ec36ff13fb921db808861cd00: ```limbo % /usr/bin/time sqlite3 testing/testing.db "SELECT id FROM users ORDER BY zipcode" >/dev/null 0.01 real 0.01 user 0.00 sys limbo % /usr/bin/time target/release/tursodb --output-mode...

SQLite C API improvements

@misterclayt0n, nice! I think things like backup API, user defined functions, and things like that are probably best tackled later. What we need is basics like `sqlite3_prepare()`, `sqlite_bind_*()`, `sqlite_column_*()`, and...

`BEGIN CONCURRENT` support

Me and @psarna did some work on doing this with SQLite itself which might be helpful https://github.com/penberg/libsql/tree/mvcc. The MVCC implementation itself is based on Hekaton here: https://github.com/penberg/tihku