Re: Bound index checks in cl-seq functions

[Tino Calancha <tino.calancha@gmail.com>]

Tino Calancha <tino.calancha@gmail.com> writes:

> Tino Calancha <tino.calancha@gmail.com> writes:
>
>> Consider following evaluations:
>>
>> (cl-position 2 (list 1 2 3) :start -999)
>> => -998
>> (cl-position 2 (list 1 2 3) :start 999)
>> => nil
>> (cl-position 2 (list 1 2 3) :start 1 :end 0)
>> => nil
>>
>> *) In the first two, the indices are out of range.
>> *) In the 3rd, the start index is higher than the end index.
>>
>> Guy L. Steele recommends to throw an error in those cases:
>> http://www.lispworks.com/documentation/HyperSpec/Issues/iss332_w.htm
>>
>> We have partially done the job, at least for `cl-subseq':
>> (cl-subseq (list 1 2 3) 999)
>> ; Signal: "Bad bounding indices: 999, nil"
>>
>> (cl-subseq (list 1 2 3) -999)
>> ; Signal: "Bad bounding indices: -999, nil"
>>
>> (cl-subseq (list 1 2 3) -1)
>> => (3) ; This is OK as documented in the docstring.
>>
>> See:
>> https://lists.gnu.org/archive/html/emacs-devel/2015-08/msg00182.html
>>
>> It would be nice if all the sequence functions in cl-seq.el
>> throw errors in those situations.
> Dear maintainers,
>
> do you agree with pushing to master the patch v4 for this thread ?
> Here:
> https://lists.gnu.org/archive/html/emacs-devel/2017-02/msg00191.html
>
> Any further suggestion/comment or something to address?
Dear John and Eli,

i am curious if you have any opinion on this thread.  Currently just
`cl-subseq' performs the bounding check index.  I don't see a
clear reason why just this function do it.
Are you in favour or against doing those checks in other functions
handling sequences?

Regards,
Tino