zcurses move issue

zcurses move issue

[Ray Andrews]

This looks like a bug but maybe it's a feature:

I have a zcurses window and I want to write to it one line -- actually 
'element' because the lines might wrap -- at a time so I iterate inside 
a 'for' loop.  To write each line you need to grab the cursor position 
which is wherever the last write stopped, then you move to the next line:

     # 'array[1]' is mouse vertial position.
     zcurses position $win array

     zcurses move "$win" $(( array[1])) $3

... looks logical.  However the top line in the window is line '0', and 
if you use 'zcurses move' to move to line zero  all hell breaks loose.  
(It will start at line zero automatically but that's besides the point 
here, I need a for loop that starts at the beginning).  So in practice 
I've been doing this:

     (( line_count > 1 )) && zcurses move "$win" $(( array[1] + 1 )) $3

... and if I print 'line_count' against 'array[1]' I get:

1 - 0

2 - 0

3 - 1

4 - 2

... but I need a consistent relationship so I've been using this nasty 
little hack:

     integer aaa
     (( aaa = array[1] + 1 ))
     [[ "$line_count" > '1' ]] && (( aaa++ ))

... and the relationship vis. 'aaa' is:

1 - 1

2 - 2

3 - 3

.... All good, and it converts the mouse 'Y' value to 'one-based' which 
is fine, just so long as it's consistent.  However it would be better if 
'zcurses move' simply accepted '0' as input -- at the top of the loop 
the first move is 'no change' but so what? Perhaps it would be neater to 
simply convert from 'move, write' to 'write, move'.  Still, asking for a 
null move seems to me to be permissible -- move to where you already are.

Re: zcurses move issue

[Bart Schaefer]

On Tue, Jan 16, 2024 at 8:13 AM Ray Andrews <rayandrews@eastlink.ca> wrote:
>
> This looks like a bug but maybe it's a feature:

With very few exceptions the zcurses subcommands are directly
translated into calls to the ncurses library.  In this case, wmove().

I think it's important that we continue to faithfully reproduce the
library behavior, even if it seems inconvenient.

Re: zcurses move issue

[Ray Andrews]

On 2024-01-16 11:03, Bart Schaefer wrote:
> On Tue, Jan 16, 2024 at 8:13 AM Ray Andrews <rayandrews@eastlink.ca> wrote:
>> This looks like a bug but maybe it's a feature:
> With very few exceptions the zcurses subcommands are directly
> translated into calls to the ncurses library.  In this case, wmove().
>
> I think it's important that we continue to faithfully reproduce the
> library behavior, even if it seems inconvenient.

As you think best.  But it does seem buggy.  Can one not move to 
position zero?  Anyway my hack has served reliably for a long time so 
letting sleeping dogs lay isn't a a big deal.  Seems that ncurses could 
have used a bit more testing. I could have made myself useful back in 
the day.


BTW, I only got one copy of this post!  Did something change? It's my 
preference, still I myself didn't do anything.

>