Matthew Woehlke <address@hidden> wrote:
> Disallowing it is fine, but if you disallow it, it seems reasonable to
> expect the program to say that it was disallowed :-). (I'm on the fence
> if this counts as a specification or implementation detail.)
The more specification details we can agree on, the less programming
time & effort will be wasted. But at a certain point, we need to write
them down somewhere else, otherwise it'll be hard to summarize them
easily.
> >Agreed (except s/nine/ten/).
>
> Erm... no, I think nine is right...
>
> line n -n
> 1 1 -9
> 2 2 -8
> 3 3 -7
> 4 4 -6
> 5 5 -5
> 6 6 -4
> 7 7 -3
> 8 8 -2
> 9 9 -1
>
> ...so '5,-5' in this case refers to one line... the fifth one.
Yes, you're right.
> So I guess it's important to specify if '-1' means "the last line", or
> "the n'th-to-last last line (n==1 in this example)" (in which case '-0'
> means something). Similarly, if 'n,+k' means "k line(s), starting at n"
> or "line n plus k line(s)".
I fail to see where your descriptions don't apply to the current
working mechanism (implying that I eventually don't fully grok your
wording).
> Possibly best to pick so that '-0' and '+0' both either make sense
> or do not make sense. I think I lean towards being all one-based.
I wouldn't want to use '+0' & '-0', because '+0' is basically a
no-op and assigning '-0' a meaning when '-1' does the same, seems
superfluous. We could emit some kind of outbound-error though.