Most "redeclaration" errors are caught when the function is compiled: js> function f() { try { const a=1; const a=2; } catch(e) { print(e); } } typein:15: TypeError: redeclaration of const a: typein:15: function f() { try { const a=1; const a=2; } catch(e) { print(e); } } typein:15: ......................................^ But for |arguments|, it is caught only when the function is called: js> function f() { try { const arguments=null } catch(e) { print(e); } } js> f() typein:13: TypeError: redeclaration of var arguments This confuses an extension to jsfunfuzz, which expects that if X compiles, then try{X}catch(e){} does not throw. It also seems inefficient. Related to bug 355480?

(In reply to comment #0) > This confuses an extension to jsfunfuzz, which expects that if X compiles, then > try{X}catch(e){} does not throw. Object = function () { throw 1 }; try { throw 2 } catch (e) {} probably should throw 1 (but doesn't).

(In reply to comment #1) > (In reply to comment #0) > > This confuses an extension to jsfunfuzz, which expects that if X compiles, then > > try{X}catch(e){} does not throw. > > Object = function () { throw 1 }; > try { throw 2 } catch (e) {} > > probably should throw 1 (but doesn't). ECMA-262 Edition 3 is flawed in reflecting on Object and constructing an instance of it to make the scope object for the catch clause, and Edition 4 as proposed changes this part of the spec to use lexical scope (let, block scope). So you have to go back to MOZILLA_1_8_0_BRANCH, Firefox 1.5, to see the ES3 behavior: js> Object = function () { throw 1 }; function () { throw 1; } js> try { throw 2 } catch (e) { print(e) } uncaught exception: 1 In Firefox 2 and 3, JS1.7 and 1.8, we changed to match ES4: js> Object = function () { throw 1 }; function () { throw 1; } js> try { throw 2 } catch (e) { print(e) } 2 See bug 336379. But what does any of this have to do with comment 0? /be

No longer reproducible - Resolving as WFM.