Nuno Cruces

I'm interested in this for my SQLite driver. May solve some issues with unnecessary allocs and type conversions due to the dynamic nature of SQLite. > If the driver "knows"...

Doesn't that depend on the Go version it's compiled with? I mean the driver may implement `RowsColumnScanner` with the expectation that it will be called, and so cheat and simply...

Well it could receive a `src` value that is whatever the driver put in `dest[i]` on `database/sql/Rows.Next(dest []Value)`. But my expectation is that drivers that bother implementing this, are likely...

On Linux my SQLite driver uses OFD locks, which showed up in Linux 3.15 (2017, all currently supported versions have it). Given the complexity of supporting [POSIX advisory locks](https://github.com/sqlite/sqlite/blob/b74eb0/src/os_unix.c#L1073-L1161), this...

It's possible the SQLite driver is to blame. The performance depends on the JIT compiler which only supports amd64 and arm64. See the support matrix: https://github.com/ncruces/go-sqlite3/wiki/Support-matrix

As for WASI being heavily inspired by POSIX just reading the [README](https://github.com/WebAssembly/WASI/blob/main/README.md): > WASI started with launching what is now called Preview 1, an API using the witx IDL, and...

@sirily11 does the posted solution work for you? I can stand up a PR adding that to the README, or you can put one up yourselves.

This spec change is about allowing already existing instructions to be used in places where only constants are allowed. To tell you the truth, I'm not sure exactly what this...

@karloscarweber I did https://github.com/wazero/wazero/tree/const-expr but please don't feel discouraged as I don't have much spare time. Use it to guide your effort, or not at all. Running the new `spec_test.go`...

So the compiler still doesn't work on OpenBSD. Maybe someone wants to tackle that.