MIME type parsing, parameter values

[annevk]

This is probably worth discussing here, if not just to have a reference. Parameter values can be unquoted or quoted. When they're quoted they can contain a couple more characters, such as ";". Inside a quoted value you escape \ and ", by preceding them with a \. You can also escape other code points, but that's a no-op.

The data model I created effectively stores the values as given. "\g\b\k" does become "gbk", but "\"" stays "\"". gbk stays gbk. This means that you need an algorithm to extract the "raw value".

The reason I did this is because browsers currently mostly pass through MIME types, and don't normalize them to all have quoted (or unquoted) parameters. We could decide on something else, but it seems that would be a bigger change from the status quo.

cc @bzbarsky @sleevi @mcmanus

[annevk]

It's worth mentioning here that we'll need quoted values sometimes as you cannot represent ";" otherwise for instance. Given that, I think this data model is good enough. It just complicates a couple consumers in the platform a little as they have to invoke a "get the parameter value" algorithm to get the actual value to operate on.