Very weird shell & sysread behavior

Very weird shell & sysread behavior

[Sebastian Gniazdowski]

Hello,
the screenshot says it all:
http://psprint.blinkenshell.org/weird.png

In the while loop, $__tmp[-1] returns correct data.

After the loop, $__tmp[-1] nor $__tmp[@] contain any data.

More: when I provide the data (on the writing end) in following way:

        print -r -- "${(q)ZTSMAP[colsearch_pattern]}"$'\0'
        print -nr -- "${(j::)${__zts_hcw_found[@]/(#e)/--${num}-del--}}"

Then the issue occurs. But when I'll do it in following way:

        print -nr --
"${(q)ZTSMAP[colsearch_pattern]}"$'\0'"${(j::)${__zts_hcw_found[@]/(#e)/--${num}-del--}}"

I.e. through write of a single string – then the after-loop $__tmp
references contain correct data. What's going on?

-- 
Sebastian Gniazdowski
News: https://twitter.com/ZdharmaI
IRC: https://kiwiirc.com/client/chat.freenode.net:+6697/#zplugin
Blog: http://zdharma.org

Re: Very weird shell & sysread behavior

[Sebastian Gniazdowski]

Turned out the tricky output-parameter specifier, the '__tmp[(#__tmp +
1)]' isn't correct. Apparently because #__tmp in math context is
character-code. Regular __tmp_size variable and '__tmp[__tmp_size +
1]' works correctly, even when the data is printed in two print -r
instructions, not as the single string.
On Sat, 27 Oct 2018 at 22:35, Sebastian Gniazdowski
<sgniazdowski@gmail.com> wrote:
>
> Hello,
> the screenshot says it all:
> http://psprint.blinkenshell.org/weird.png
>
> In the while loop, $__tmp[-1] returns correct data.
>
> After the loop, $__tmp[-1] nor $__tmp[@] contain any data.
>
> More: when I provide the data (on the writing end) in following way:
>
>         print -r -- "${(q)ZTSMAP[colsearch_pattern]}"$'\0'
>         print -nr -- "${(j::)${__zts_hcw_found[@]/(#e)/--${num}-del--}}"
>
> Then the issue occurs. But when I'll do it in following way:
>
>         print -nr --
> "${(q)ZTSMAP[colsearch_pattern]}"$'\0'"${(j::)${__zts_hcw_found[@]/(#e)/--${num}-del--}}"
>
> I.e. through write of a single string – then the after-loop $__tmp
> references contain correct data. What's going on?
>
> --
> Sebastian Gniazdowski
> News: https://twitter.com/ZdharmaI
> IRC: https://kiwiirc.com/client/chat.freenode.net:+6697/#zplugin
> Blog: http://zdharma.org



-- 
Sebastian Gniazdowski
News: https://twitter.com/ZdharmaI
IRC: https://kiwiirc.com/client/chat.freenode.net:+6697/#zplugin
Blog: http://zdharma.org

Re: Very weird shell & sysread behavior

[Bart Schaefer]

On Sat, Oct 27, 2018 at 1:53 PM Sebastian Gniazdowski
<sgniazdowski@gmail.com> wrote:
>
> Turned out the tricky output-parameter specifier, the '__tmp[(#__tmp +
> 1)]' isn't correct. Apparently because #__tmp in math context is
> character-code. Regular __tmp_size variable and '__tmp[__tmp_size +
> 1]' works correctly, even when the data is printed in two print -r
> instructions, not as the single string.

You could also do __tmp[$#__tmp+1] if you want to avoid the extra variable.

Re: Very weird shell & sysread behavior

[Sebastian Gniazdowski]

On Sun, 28 Oct 2018 at 06:22, Bart Schaefer <schaefer@brasslantern.com> wrote:
>
> On Sat, Oct 27, 2018 at 1:53 PM Sebastian Gniazdowski
> <sgniazdowski@gmail.com> wrote:
> >
> > Turned out the tricky output-parameter specifier, the '__tmp[(#__tmp +
> > 1)]' isn't correct. Apparently because #__tmp in math context is
> > character-code. Regular __tmp_size variable and '__tmp[__tmp_size +
> > 1]' works correctly, even when the data is printed in two print -r
> > instructions, not as the single string.
>
> You could also do __tmp[$#__tmp+1] if you want to avoid the extra variable.

Thanks, didn't try this. The extra variable has additional positive
effect – when increased for each 0 return from sysread, it doesn't
compute array length each read. But for 65kB reads, it probably
doesn't matter, as there will not be many of reads at all (to exhaust
the input FD).

Blinkenshell isn't responding, a more permanent and responsive link:
https://imgur.com/YGWYTLe

-- 
Sebastian Gniazdowski
News: https://twitter.com/ZdharmaI
IRC: https://kiwiirc.com/client/chat.freenode.net:+6697/#zplugin
Blog: http://zdharma.org