range->vector Marc Nieper-Wißkirchen (01 Sep 2020 11:29 UTC)
Re: range->vector John Cowan (01 Sep 2020 15:29 UTC)
Re: range->vector Marc Nieper-Wißkirchen (01 Sep 2020 15:45 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (01 Sep 2020 16:33 UTC)
Re: range->vector John Cowan (01 Sep 2020 17:12 UTC)
Re: range->vector Marc Nieper-Wißkirchen (01 Sep 2020 17:27 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (01 Sep 2020 17:34 UTC)
Re: range->vector Marc Nieper-Wißkirchen (01 Sep 2020 17:36 UTC)
Re: range->vector Arthur A. Gleckler (01 Sep 2020 17:37 UTC)
Re: range->vector Marc Nieper-Wißkirchen (01 Sep 2020 17:38 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (01 Sep 2020 17:46 UTC)
Re: range->vector John Cowan (01 Sep 2020 18:23 UTC)
Re: range->vector Arthur A. Gleckler (01 Sep 2020 18:40 UTC)
Re: range->vector John Cowan (01 Sep 2020 18:42 UTC)
Re: range->vector Marc Nieper-Wißkirchen (01 Sep 2020 18:52 UTC)
Re: range->vector Marc Nieper-Wißkirchen (01 Sep 2020 19:22 UTC)
Re: range->vector John Cowan (03 Sep 2020 00:15 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (03 Sep 2020 02:27 UTC)
Re: range->vector John Cowan (03 Sep 2020 03:35 UTC)
Re: range->vector Marc Nieper-Wißkirchen (03 Sep 2020 06:47 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (03 Sep 2020 18:04 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (03 Sep 2020 18:27 UTC)
Re: range->vector Marc Nieper-Wißkirchen (03 Sep 2020 19:10 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (03 Sep 2020 20:32 UTC)
Re: range->vector Wolfgang Corcoran-Mathe (03 Sep 2020 07:11 UTC)
Re: range->vector Marc Nieper-Wißkirchen (03 Sep 2020 07:14 UTC)

Re: range->vector Wolfgang Corcoran-Mathe 03 Sep 2020 02:27 UTC

On 2020-09-02 20:15 -0400, John Cowan wrote:
> So the safe implementation is to unpack all the elements of all the ranges
> into a single vector-style range, which makes string-append O(n) but
> preserves the guarantee of an O(1) range-ref in all cases.

ACK.  I'll update the sample implementation of range-append shortly.

> I added a note that the SRFI recommends that all indexers are O(1).

This means that range-reverse should also return a vector-style range.
Since the earliest days of SRFI 196, the sample implementation of this
function has used a composed indexer (which was the only option,
pre-vector-range).

--
Wolfgang Corcoran-Mathe  <xxxxxx@sigwinch.xyz>

"The intellectual advancement of man depends on how often he can
exchange an old superstition for a new truth." --Robert G. Ingersoll