You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
define new ValueValue kinds for new directive values which are more complex than just a boolean. At a glance, you probably want a special Value kind for tt-policy-name, but the others look like they can be handled as booleans on the directive.
add helpers to allow reading and writing their attributes, and ensure that updating keeps their internal values field up to date (by using the addValue and removeValueIgnoringCase helpers, mainly)
define new ValueValue kinds for new directive values which are more complex than just a boolean. At a glance, you probably want a special Value kind for tt-policy-name, but the others look like they can be handled as booleans on the directive.
add helpers to allow reading and writing their attributes, and ensure that updating keeps their internal values field up to date (by using the addValue and removeValueIgnoringCase helpers, mainly)
The main challenges in my experience is figuring out what the spec is trying to say and whether that's actually what browsers do. For example, it says the syntax is defined by an ABNF, which means keywords are case-insensitive ("ABNF strings are case insensitive"). So per spec, 'ALLow-DUPlicATES' is a legal way to write 'allow-duplicates'. Is that actually intentional? Is that what browsers do? Gotta check. Similarly, are policy names case-sensitive or not? That affects the representation and especially the manipulation of those Values.
The text was updated successfully, but these errors were encountered:
It would be nice to have handling for: the
trusted-types
trusted-types
andrequire-trusted-types-for directives.require-trusted-types-for
This advice was provided pre-fork:
The process will be basically
ValueValue kinds for new directive values which are more complex than just a boolean. At a glance, you probably want a special Value kind fortt-policy-name
, but the others look like they can be handled as booleans on the directive.DirectivesDirective kind for both new directives (trusted-types
trusted-types
) and (require-trusted-types-for)require-trusted-types-for
values
field up to date (by using theaddValue
andremoveValueIgnoringCase
helpers, mainly)ValueValue kinds for new directive values which are more complex than just a boolean. At a glance, you probably want a special Value kind fortt-policy-name
, but the others look like they can be handled as booleans on the directive.DirectivesDirective kind for both new directives ("trusted-types" and ("require-trusted-types-for")values
field up to date (by using theaddValue
andremoveValueIgnoringCase
helpers, mainly)"Should Trusted Type policy creation be blocked by Content Security Policy?"Should Trusted Type policy creation be blocked by Content Security Policy?
, though these aren't so important.update the parser to handle themThe main challenges in my experience is figuring out what the spec is trying to say and whether that's actually what browsers do. For example, it says the syntax is defined by an ABNF, which means keywords are case-insensitive ("ABNF strings are case insensitive"). So per spec,
'ALLow-DUPlicATES'
is a legal way to write'allow-duplicates'
. Is that actually intentional? Is that what browsers do? Gotta check. Similarly, are policy names case-sensitive or not? That affects the representation and especially the manipulation of those Values.The text was updated successfully, but these errors were encountered: