gwenn

@t-kalinowski Could you please explain why https://github.com/rusqlite/rusqlite/blob/master/examples/loadable_extension.rs#L21-L29 is too much ? Considering that we cannot impose the `sqlite3_extension_init` name.

https://github.com/rusqlite/rusqlite/issues/1270#issue-1498727620 > the compiled version of SQLite inside of Windows is only meant to be used by projects that are bundled as part of the Windows operating system ( https://github.com/microsoft/win32metadata/issues/824#issuecomment-1067220882),...

@davidbarsky, @thomcc Could you please advise ?

Conflict with https://github.com/rusqlite/rusqlite/pull/1362/files And as far as i know, we don't want a MSRV (see https://github.com/rusqlite/rusqlite/issues/566)

> @gwenn I looked at the #1362 - that PR has far greater scope than a simple 2021 edition bump. I think this PR is much easier to merge first,...

> @gwenn sure, this is an "upkeep" PR. It is small to review but makes a fairly big change in terms of potentially breaking things (2021 edition) - so if...

Could you please try (if you don't have already) to force bindings generation with this features: `buildtime_bindgen` ? Thanks.

Are you sure that Thread 1 transaction is committed when Thread 2 sends a `SELECT` ? This should be the case if you are **not** using WAL mode because Thread...

See [official Tcl extension](https://github.com/sqlite/sqlite/blob/master/src/tclsqlite.c#L127-L140). `rusqlite` does the same with cached statements (but uses an LRU map instead of linked list).

> I have made an attempt at improving the ergonomics of the `Statement` type. 🙂 > > #1375 which actually doesn't solve the original issue.