1. Re: Bind Problems ...
Boehme, Gabriel wrote:
>Roderick Jackson wrote:
>
>>Robert Craig wrote:
>[snip]
>>>In any redesign of bind, I would seriously consider making
>>>it 2-pass, so it could discard any routines, variables or
>>>constants that aren't used.
<snip>
>Later, Roderick Jackson wrote:
>
>>[snip]
>>I realize my argument against the idea sounds a bit self-serving, but then
>>again the impact would be somewhat farther reaching than just the few
>>things I've produced. We're talking about eliminating entire ways of
>>working in the language here, all for the sake of shrinking (compiled, not
>>source) code. I would rather see an effort to produce smaller, possibly
>>more numerous libraries rather than see it all come to that.
>
>Rod...I *REALLY* think you're getting all worked up over nothing here. First
>of all, I'm sure this discarding of unused routines/variables/constants
>would be in the bind program as a command-line *option*. Secondly, this idea
>makes a lot of sense for almost every situation, and it seems a bit silly to
>throw it away entirely. Thirdly, this would *only* happen with bound or
>shrouded code, so I don't see why dynamic include methods would have any
>problems. Finally, I'm sure Rob would make it so that the binder is *smart*,
>and only gets rid of stuff it definitely won't need.
Okay, I see your point. As a bind option, this would not bother me nearly
as much. Then, the situations that benefited from it could take advantage
of it, and those that wouldn't could leave it alone.
However, the dynamic include example addressed the idea of having unused
code removed in non-compiled code, where I have a hard time seeing it as
merely an option.
Mmm, I also think I see how the binder *might* be able to identify and keep
routines used arbitrarily in object-oriented code... provided (this seems to
keep popping up) that all routine_id calls were only passed literals.
>I for one think Rob's idea is wonderful -- it'll help cut down on the size
>of bound code, and I won't mind including a whole library just for one
>routine. (As things are now, I'm tempted just to code that one routine
>manually into my program and forget the include library...) Having it in the
>bind program as one of the command-line options should cover all situations
>quite nicely, IMHO.
I agree with you; as a command-line option, this would be useful. I
probably did overreact a bit; I hadn't stopped to consider the feature
might not be mandatory.
Rod Jackson
