2009/5/8 Christophe Rhodes <csr21@...>:
> "Nikodemus Siivola" <demoss@...> writes:
>> --- array.lisp 7 May 2009 11:11:05 -0000 1.86
>> +++ array.lisp 8 May 2009 19:08:07 -0000 1.87
>> @@ -328,19 +328,17 @@
>> ;;; vectors or not simple.
>> (macrolet ((def (name table-name)
>> - (defvar ,table-name)
>> + (defglobal ,table-name (make-array ,sb!vm:widetag-mask))
> I know it doesn't really really matter while widetags are x10, but
> this should probably be (make-array ,(1+ sb!vm:widetag-mask)),
> shouldn't it?
>> + (def !find-data-vector-setter **data-vector-setters**)
> I think for within sbcl itself, true magic like this is probably
> %%foo%% (like %%current-segment%% and %%current-vop%%).
>> (defmacro-mundanely defvar (var &optional (val nil valp) (doc nil docp))
>> - "Define a global variable at top level. Declare the variable
>> + "Defines a special variable at top level. Declare the variable
>> SPECIAL and, optionally, initialize it. If the variable already has a
>> value, the old value is not clobbered. The third argument is an optional
>> documentation string for the variable."
> I don't think I agree with this change of voice, but if you're doing
> this I think you should do it for all sentences in the docstring, not
> just the first.
I assume you mean the imperative to passive change? Initial imperative
sentences like that in documentation read oddly to me, but you are the
native speaker -- so restored the imperative.