* empty lines are always ignored instead of separating groups
* the "user-agent" line after a rule starts a new group
* groups matching the same user agent are now merged
* the rule with the longest match wins instead of the first matching rule
* in case of equal matches, the “Allow” rule wins over “Disallow”
* special characters “$” and “*” are now supported in rules
* prefer full match for user agent

Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>

…eError (#146407)

When Levenshtein-based suggestions find no match for an AttributeError
on list, str, or dict, check a static table of common method names from
JavaScript, Java, C#, and Ruby.

For example, [].push() now suggests .append(), "".toUpperCase() suggests
.upper(), and {}.keySet() suggests .keys().

The list.add() case suggests using a set instead of suggesting .append(),
since .add() is a set method and the user may have passed a list where
a set was expected (per discussion with Serhiy Storchaka, Terry Reedy,
and Paul Moore).

Design: flat (type, attr) -> suggestion text table, no runtime
introspection. Only exact builtin types are matched to avoid false
positives on subclasses.

Discussion: https://discuss.python.org/t/106632

Co-authored-by: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

…atch documentation and improve clarity (#144983)

…#149384)

Co-authored-by: Savannah Ostrowski <savannah@python.org>

#149375)

Co-authored-by: Savannah Ostrowski <savannah@python.org>