[JSOUP-2224] add wildcards #2225
Conversation
It is common for applications to use custom attributes, e.g., 'aria-*', and in fact with HTML 5(?) this is offically recognized with 'data-*' wildcards. It is often impossible to list all possible custom attributes, yet we want to be more selective than using ':all'. This patch adds support for both per-tag and global attribute wildcards. The wildcard must be recognized by `java.text.Pattern`, e.g., `data-.+`. The patch will quietly add the '^' and '$' delimiters to the provided pattern.
Fixed removeTags() - it hadn't removed the wildcard attributes. Also changed the Map so the key is a TagName, not a String, in order to avoid any uncertainty regarding capitalization.
|
I'll check the error... After thinking about it overnight I think it makes sense to remove the explicit 'wildcard' methods and use I also realized that I need to add corresponding functionality to the add/remove protocol methods. On a related note the 'All' value should be exposed. That would be used for global wildcards. |
|
I almost forgot - the All should probably be extended with FormControl, Media, and possibly Visible. These would be used with the standard event attributes - it's shorthand for multiple related tags. |
Added wildcards attributes for both tags and global scope. This allows the user to specify an unknown set of attributes, e.g.,
data-*, using a wildcard pattern. This is different from:allsince it is selective.This pattern must be recognized by
java.util.regex.Pattern, e.g.,data-.+. The pattern is case-insensiitve and must match the entire string since '^' and '$' are quietly added.A variant approach is to extend the existing
addAttributes()andremoveAttributes()methods to recognize patterns, e.g., via a marker like enclosing braces. (hence{data-.+}). This approach could use the same implementation as below - the existing methods would simply strip the braces and call the now-private methods.