Subj : Re: try catch blocks
To   : =?ISO-8859-1?Q?Georg_Maa=DF?= <georg@bioshop.de>
From : Brendan Eich <brendan@meer.net>
Date : Thu May 15 2003 01:07 pm

Georg Maaß wrote:

> Brendan Eich wrote:
>
>> SpiderMonkey supports an extension to ECMA: catch guards.  If you write
>>
>>  try {
>>    . . .
>>  } catch (e if e instanceof SQLException) {
>>    // here we know e is an SQLException
>>  } catch (e) {
>>    // this is the "catch-all" clause
>>  } finally {
>>    // code here runs no matter how control leaves the try-catch chain
>>  }
>>
>> I don't recall whether Rhino supports this extension.
>>
>> /be
>
>
> Will this awful hack remain in JS 2.0?


Why is this an "awful hack"?

In JS1, it's necessary unless you want to foist the requirement to have 
a catch-call clause that re-throws e on every programmer who uses 
catch.  Consider Amit's example, written using ECMA-262 Edition 3 only:

  try {
    /something complex that may throw arbitrary exceptions/;
  }
  catch (e) {
    if (e instanceof SQLException) {
      deal with the wanted exception;
    } else {
      throw e;
    }
  }
  finally {
    clean up;
  }

Now tell me why it's "awul" to save the programmer from the error-prone 
requirement to write the if-else-rethrow boilerplate.

> There it is not necesary, if it is implemented like known from C++ an 
> Java.


Yes, of course -- JS2 is quite a different language.  What made your 
comment relevant to JS1?

/be

.