Where? Actually, there are two: CATCH function and [catch] function
attribute.
===The CATCH function
This function essentially enables us to immediately finish a complicated
evaluation.
Sample code
catch [
; doing complicated evaluation...
if done? [throw result1]
; more evaluations...
if done? [throw result2]
; more evaluations...
; now we are done for sure
result3
]
Similar effect can be achieved using ANY, ALL or CASE functions, but in
the general case CATCH is the most appropriate for the job.
When using "embedded" CATCH evaluation, we can use the CATCH/NAME
refinement to be able to specify the target for the desired results as
follows:
catch/name [
; ...
catch/name [
; ...
if done-a? [throw/name result 'a] ; this is meant for the 'a
CATCH (the "outer" CATCH)
; ...
if done-b? [throw/name result 'b] ; this is meant for the 'b
CATCH (the "inner" CATCH)
; ...
result-b
] 'b
; ...
if done-a? [throw/name result 'a]
; ...
result-a
]
===The [catch] function attribute
This attribute is meant to enable "above standard" error handling in
functions.
Example: let's suppose we need a HANDLE function to handle every integer
number except for 66. We consider it an error to call the HANDLE
function with number 66 as its argument. Code:
handle: func [
[catch]
argument1 [integer!]
argument2 [block!]
] [
if argument1 = 66 [throw make error! "number 66 not allowed as
argument1"]
do argument2
]
Now expression
handle 1 ["OK"]
yields "OK", while
handle 66 ["OK"]
is an error.
===Comparison
*the [catch] function attribute has a different purpose than the CATCH
function.
*there isn't a similar "mechanism" of discerning the meaning of THROW
for embedded functions
===Interference
An embedded [catch] - attribute function cannot discern the meaning of
THROWS and it catches all throws as far as they are throwing errors.
Example:
index: 1
block: copy ["OK"]
insert tail block make error! "error"
type? catch [
; ...
handle 1 [throw pick block index]
]
So far so good, but if we change the first line to
index: 2
, the interference will appear. We may ask whether the /NAME refinement
could help, but the answer is negative as Romano found out.
===Solution
I offered TFUNC functions to solve the throw issue (as well as some
other issues) for embedded functions earlier. (see
http://www.compkarori.com/vanilla/display/Function-Attributes).
The interference between [catch] and CATCH can be solved by using local
throws too, but some users may find this solution uncomfortable.
A simpler solution might use /NAME refinement for THROW, but that should
be extended to enable us to specify an appropriate catch name for
[catch] functions too. In that case I think, that it would be great if
the name comparison was using =? not just = to use the full power of Rebol.
The End
PS: I am curious, whether you prefer a "local-word" solution or a /NAME
solution, so consider this a user poll.
-Ladislav
--
To unsubscribe from the list, just send an email to
lists at rebol.com with unsubscribe as the subject.
