Attributes/props of custom elements with names starting with "on" are shadowed

[mspears-alaska]

Describe the bug

When using custom elements as a child of a svelete component in svelte 5, attributes/props starting with 'on' get shadowed by the implicit remapping of on:foo to onfoo. For example, as seen in the repl, a property like oneworld which is expected to be a bool can't be used as a shorthand and returns null as a value even when explicitly setting it as true.

Reproduction

https://svelte.dev/playground/8e97714826aa4a0fb17c5aafef24bdc5?version=5.14.0

Logs

This is the error shown if you try to use the boolean property shorthand

Event attribute must be a JavaScript expression, not a string
https://svelte.dev/e/attribute_invalid_event_handler

e.g. <custom-element oneworld foo={false}></custom-element>

System Info

System:
    OS: macOS 15.1.1
    CPU: (11) arm64 Apple M3 Pro
    Memory: 64.33 MB / 18.00 GB
    Shell: 5.9 - /bin/zsh
  Binaries:
    Node: 20.18.1 - ~/.nvm/versions/node/v20.18.1/bin/node
    npm: 10.8.2 - ~/.nvm/versions/node/v20.18.1/bin/npm
  Browsers:
    Chrome: 131.0.6778.140
    Edge: 131.0.2903.99
    Safari: 18.1.1
  npmPackages:
    svelte: ^5.11.0 => 5.12.0

Severity

blocking an upgrade

[Ocean-OS]

Perhaps Svelte could treat on-props as regular props for any non-standard elements?
Components don't treat on-props any differently than any other prop, so I don't see why this should be different for custom elements (or maybe instead of treating any attribute starting with on as an event handler, Svelte could have an array of valid HTML events and look for those instead).

[dummdidumm]

While this is true, the general advice - outside of Svelte, too - is to not use properties/attributes starting with on. All JSX-based frameworks (like React etc) also rely on this convention.

[JustinSometimesWrong]

While this is true, the general advice - outside of Svelte, too - is to not use properties/attributes starting with on. All JSX-based frameworks (like React etc) also rely on this convention.

Do you have references for this statement?