WebAssembly version

[hjerabek]

Are there any plans to provide this database engine as a WebAssembly/WASM module? It could then be used with the same bindings in any environment that provides a WebAssembly runtime, like browsers, Node.js, Deno, Bun etc. Instead of a file, the input would be a continuous piece of memory the module can operate on. If the environment allows it, it could be (memory-)mapped to whatever underlying persistent storage is available...

[erthink]

The support of WASM is in the TODO-list.

However, I do not develop using WebAssemble (via Emscripten). So there is a lot that is unclear to me, especially everything related to use-cases and necessary features. For instance, I don't understand the value of running a database without durable persistence and/or only in shared memory, because all such (or like the same) can be done using JS.

On the technical side, the most important point is to access the native system API, including POSIX mutexes in shared memory and undocumented Win32 functions from NTDLL. Without this, many features of libmdbx become unavailable, or unexpected limitations arise (for example, the database cannot be used simultaneously from the WebAssembly environment and a native application).

So, I've just added Emscripten/WebAssembly support to TODO and am dealing with more demanding tasks, especially what sponsors need.

If someone needs Emscripten support, then I would like to see what is required by looking at a particular project and seeing the willingness of assist efforts.

[hjerabek]

The use-case I was thinking about is related to the relatively new browser support for an Origin Private File System (OPFS) which, according to the spec, "provides access to a special kind of file that is highly optimized for performance and offers in-place write access to its content". I thought this could be used to run high-performance file-based database engines like LMDB or libmdbx in the browser, and to no longer have to rely on the archaic low-performance stores like IndexedDB, localStorage etc. However, in contrast to you, I am not all too familiar with the intricacies of libmdbx, and the whole thing is probably much more complex than it appears to be. I still thank you for putting it on the todo list.

PS: Should I mark this issue as closed?

[erthink]

Should I mark this issue as closed?

I don't think you need to worry about anything. Let it remain in question and answer status.

[erthink]

Conclusion:

1. It is a lot that is unclear to me, especially everything related to use-cases and necessary features. For instance, I don't understand the value of running a database without durable persistence and/or only in shared memory, because all such (or like the same) can be done using JS.
2. Before do something I would like to see what is required by looking at a particular project and seeing the willingness of assist efforts.