by rodd@[EMAIL PROTECTED]
(Rod Dorman)
May 4, 2008 at 04:56 AM
In article <FFQSj.209$KB3.87@[EMAIL PROTECTED]
>,
Duke Normandin <dukeofperl@[EMAIL PROTECTED]
> wrote:
>On 2008-05-02, Rod Dorman <rodd@[EMAIL PROTECTED]
> wrote:
>> ...
>> The danger in using ****d references comes from assuming that the
>> ****d indicator wont change when making subroutine or function
>> calls. It might work fine when you write and test it but break at some
>> point in the future if/when the subroutine is changed.
>>
>
>So the most prudent course to take would seem to be to refrain from
>using such potential booby-traps, would it not?
The choice is yours. I find if you use ****d references only when the
execution sequence is linear (i.e. no extrinsic function or subroutine
calls) it isn't a problem.
Note that this assumes you don't have a coding style that (shudder)
includes GOTO Tag+Offset
--
-- Rod --
rodd(at)polylogics(dot)com
