new encodingStyle attribute

Cantor, Scott cantor.2 at osu.edu
Mon Sep 17 11:26:27 EDT 2018


> The reason I ask is that changes to the MDQ spec might be needed. In
> particular, the question mark seems to be a MUST-encode character
> since there are entityIDs with query strings in the wild.

That doesn't follow at all. All we're doing is applying a reasonable/well-regarded encoder to the entityIDs when they appear in different segments of a URL, whatever place that happens to be based on the deployer's instructions.

Java's default behavior is old and very unaware of context, so it's a bad choice to use with very little pragmatic experience that doesn't date back to the 90s I would guess. That's all this is about.

No matter which one we used, a question mark would be encoded, that's just one of the ones everybody knows about. Whether it's in the path, fragment, or query string, you can't leave that one alone.

-- Scott




More information about the dev mailing list