I think we'll want some kind of isSetter(Symbol s) / isGetter(Symbol s) API as well.

cc @gbracha.
Added Library-Mirrors, Triaged labels.

Added Area-Library label.

I think we should add two getters to Symbol:

/// If this symbol is a setter name, return it. Otherwise, return the symbol corresponding to adding '=' to this symbol.
/// TBD: Throw an error if this symbol corresponds to an empty string.
Symbol get asSetterSymbol();
/// If this symbol is not a setter name, return it. Otherwise, return the symbol corresponding to removing a trailing '='.
Symbol get asGetterSymbol();

Set owner to @peter-ahe-google.
Added Accepted label.

sgtm

Issue #13587 has been merged into this issue.

This comment was originally written by @seaneagan

Symbols can represent not only getter/setter names, but also operator, class, and library names. And it can also be ambiguous as to which one of those a particular Symbol represents, even more so after minification. How would these work:

1. #+.asSetterSymbol()
2. (#my.lib.name).asGetterSymbol()
3. (#MyClass).asSetterSymbol()
4. and 2. could potentially throw, but I'm not sure if it is always possible to tell if a minified Symbol represented an operator / library name. Otherwise (and with 3. ), it would have to return a bogus Symbol, which will lead to incorrect behavior elsewhere in the program. Similarly isGetter etc. could lead to false positives:

(#MyClass).isGetter
(#my_lib).isGetter

Adding PropertyMirror and "properties" getters, and using those to lookup properties via the plain identifier Symbol, and retrieving the getter/setter from that seems like a good solution to me.