IMPORTANT - this release of PyScript uses a new version of Pyodide that has different behaviour relating to the handling of null / None. THIS MAY BREAK YOUR CODE. Read the notes (especially regarding the pyscript.ffi.is_none feature), to understand how PyScript supports you in this transition.

• All fetch operations now produce readable error messages when these happen: pyscript/polyscript#146
• PyScript's package caching mechanism is now based on latest Pyodide lockFileContents: pyscript/polyscript#144
• Update MicroPython to its preview-386 version: pyscript/polyscript#143
• Update Pyodide to its 0.28.1 version: https://pyodide.org/en/stable/project/changelog.html#version-0-28-1 and added pyscript.ffi.is_none(renference) to ease-out the null VS None distinction.
• Add an explicit pyscript.fs.remove(path) option to both unmount that path and erase its content from the database: #2368
• Change pyscript.WebSockets default to arrayBuffer to retrieve directly memory views: #2366
• Still on WebSockets, asynchronous listeners are now supported: #2366
• PyWorker args are now more user friendly: #2366

The most relevant Pyodide changes (more Pyodide context here):

• There is a new pyodide.ffi.jsnull reference which is the default type for JavaScrpt null: goodbye null to None implicit conversion, we are providing a pyscript.ffi.is_none(reference) to ease-out this distinction that works regardless of the runtime.
• If jsobj is a JavaScript proxy for a Python dict then jsobj[name] and jsobj.name now works seamlessly, allowing a JS object representing a Python dict to be used in way more scenarios than before. For more information, see the Pyodide docs.