DESCRIPTION¶
Evaluate the expressions. If there is no error, 0 is returned. If there is an error, the evaluation of the expressions stops at that point, and a string with the error message is returned.
System error messages start with a leading ‘*’, user-defined error values (other than 0) as given to throw(E) and raise_error(E) are returned as they are.
If at the time the catch(E) is encountered less than __CATCH_EVAL_COST__ eval ticks are left, a runtime error will be thrown from inside the catch(E) (and thus caught like every other error) and the expressions will not be executed. The ‘reserve’ modifier can be used to reserve a different amount of eval ticks.
The default behaviour of catch(E) can be changed using modifiers:
‘nolog’: Normally, the caught error will be logged in the error logs for easier debugging. With this modifier, the log is suppressed.
‘publish’: Normally,
runtime_error(M) is not called for a caught error. This modifier instructscatch(E) to call it nevertheless.‘reserve
expr‘: The expression has to evaluate to a number greater than 0 and is used to determine the amount of eval ticks to reserve, instead of the default of__CATCH_EVAL_COST__. The minimum required are 2 *__MASTER_EVAL_COST__.
catch(E) itself is not expensive as far as execution time is concerned: it is about the same as a intra-object function call.
throw(E) is not very expensive either, but does include the internal cleanup of several structures.
Real runtime errors on the other hand are expensive regardless of whether they are caught or not, as they include the generation of the stack backtrace.
catch(E) is not _really_ an efun but a compiler directive.
USAGE¶
This example demonstrates how to attempt to clone an object, but write a message if an error is encountered:
object obj;
string err;
if (err = catch(obj = clone_object("/foo/bar/baz")))
write("Cannot clone object, reason:"+err"+\n");