A user has pointed out that while Wget regards spaces within href attributes (and the like) as user error, browsers such as Firefox simply strip them out.

Strict HTML syntax would interpret any whitespace within the attribute as part of the attribute's value. Whitespace is not legal as part of a URI (but then, neither are non-ASCII byte values, which have been accepted for some time); apparently a common way to deal with this is to simply strip it all out.

We should "do what other browsers do" by default, and provide an option to make Wget stricter (similar to how we currently do SGML comments).