Hi,
How do I clear the contents of an array or hash in
an extension? Setting the pointer to nil just doesn’t
seem to be right…
static VALUE t_reset (VALUE self)
{
VALUE arr;
arr = rb_iv_get(self, “@arr”);
RARRAY(arr)->ptr = ‘\0’; /* ??? */
return Qnil;
}
Thanks,
-mark.
What are you wanting to do? Make the array have 0 elements? You could use
rb_funcall to call the clear method on the array, or simply set the len
field to 0. Clearing a hash is less simple, so calling the clear method
seems best to me.
On Tue, 08 Oct 2002 05:54:30 +0900, Mark Probert wrote:
Hi,
How do I clear the contents of an array or hash in
an extension? Setting the pointer to nil just doesn’t
seem to be right…
Hi,
At Tue, 8 Oct 2002 07:32:59 +0900, Tim Hunter wrote:
or simply set the len
field to 0.
If you modify an object directly, it’s better to call
rb_check_frozen().
–
Nobu Nakada
Cheking whether frozen or not, in iteration or not, $SAFE permits
modifying the array or not. Then set len 0, and resize the underlying
bin if necessary.
That’s what done in rb_ary_clear(), implementing Array#clear and
is exported for C extension in intern.h 
I’ve look ruby-1.6.7 and -1.7.2. Perhaps more older version also have
the external declaration.
In message 200210080230.g982UiY10524@sharui.nakada.kanuma.tochigi.jp nobu.nokada@softhome.net writes:
or simply set the len
field to 0.If you modify an object directly, it’s better to call
rb_check_frozen().
–
kjana@dm4lab.to October 8, 2002
Out of sight, out of mind.
Thank you so much! I knew that there was a simple answer
in there some where.
As a general question, using rb_ary_clear(arr) seems clearer
than rb_funcall(arr, rb_intern(“clear”), 0). Is there any
downside (besides the fact that it is “undocumented” and
buried in intern.h ;-)?
-mark.
At 08:53 PM 10/8/2002 +0900, kjana@dm4lab.to wrote:
That’s what done in rb_ary_clear(), implementing Array#clear and
is exported for C extension in intern.h
Hello Mark,
Tuesday, October 08, 2002, 5:16:28 PM, you wrote:
At 08:53 PM 10/8/2002 +0900, kjana@dm4lab.to wrote:
That’s what done in rb_ary_clear(), implementing Array#clear and
is exported for C extension in intern.h
Thank you so much! I knew that there was a simple answer
in there some where.
As a general question, using rb_ary_clear(arr) seems clearer
than rb_funcall(arr, rb_intern(“clear”), 0). Is there any
downside (besides the fact that it is “undocumented” and
buried in intern.h ;-)?
of course! it can be overloaded, stolen or made private
so you must decide what you want - just clear the array or order array to
clear itself
–
Best regards,
Bulat mailto:bulatz@integ.ru
Hi,
At Tue, 8 Oct 2002 22:16:28 +0900, Mark Probert wrote:
As a general question, using rb_ary_clear(arr) seems clearer
than rb_funcall(arr, rb_intern(“clear”), 0). Is there any
downside (besides the fact that it is “undocumented” and
buried in intern.h ;-)?
Even if the arr is an instance of a subclass of Array which
overrides “clear” method, rb_ary_clear() always invokes the
original behavior. It may be expected behavior or not.
–
Nobu Nakada
Thanks for the info.
Regards,
-mark.
At 10:27 PM 10/8/2002 +0900, Nobu wrote:
At Tue, 8 Oct 2002 22:16:28 +0900, >Mark Probert wrote:
As a general question, using rb_ary_clear(arr) seems clearer
than rb_funcall(arr, rb_intern(“clear”), 0). Is there any
downside (besides the fact that it is “undocumented” and
buried in intern.h ;-)?Even if the arr is an instance of a subclass of Array which
overrides “clear” method, rb_ary_clear() always invokes the
original behavior. It may be expected behavior or not.