Allow wildcards in whitelist attributes

[foo4u]

HTML5 allows the use of data-foo, data-foo-bar, etc to specify information on elements. These are relatively harmless and should only contain text.

Currently, each data- attribute needs to be specified explicitly on a whitelist so that it's not removed by Jsoup.clean(). Can we add support for either:

1. Wildcard attributes, e.g. Whitelist.relaxed().addAttributes("a", "data-*") or
2. A new function, like Whitelist.relaxed().allowDataAttributes("a")

[foo4u]

Also wanted to add that I'm willing to contribute code to support this. Before doing so, I just want to make sure this change is acceptable and determine the best way to support it (options 1 or 2 above, or something entirely different). Thanks!

[jhy]

Hi @foo4u. Sorry for the late reply. I like option 2 (because I can't think of another case which it would be helpful for). Would be great if you write that.

[remisbaima]

I guess it would be more flexible if you implemented option 1 with regex patterns instead of only wildcards. E.g.: Whitelist.relaxed().addAttributes("a", "data-.*") Think about e.g. https://angularjs.org code that has attributes starting with "ng-". Also almost every second ;-) a new JS framework appears and these might require new attributes prefixes. With option 1 with regex support you would be more future-proof.

[jhy]

OK, handling examples like that makes sense. I'd be OK with either a prefix or a regex matcher. The prefix match seems simple and unlikely to let anyone shoot themselves in the foot.

[foo4u]

Ok, will try to get a PR for prefix matching sent in a few weeks.

[jhy]

(Closing out old, dormant bugs. If you are still impacted by this, please reopen & vote.)

[swapab]

I am trying out jsoup to validate html pages. Works great so far. Would have been awesome, if wildcards were possible with jsoup.

[promiselaoliu]

Ok, will try to get a PR for prefix matching sent in a few weeks.

Hi @foo4u , have you ever prepared a PR?

[promiselaoliu]

I have prepared a PR: https://github.com/jhy/jsoup/pull/1871

[jhy]

(Reopening as mentioned in earlier close, there is renewed interest here.)

[irandamay]

Is there any update on this... it looks like the changes requested in the PR were made back in February?

I have been watching these for some months now because we have a need to not strip out aria-* attributes.