Re: How do we get state of a flag in set -o ...

[alex xmb ratchev]

On Wed, Jul 12, 2023, 12:15 AM Greg Wooledge <greg@wooledge.org> wrote:

> On Tue, Jul 11, 2023 at 11:49:40PM +0200, alex xmb ratchev wrote:
> > set -- opt1 val1 opt2 val2 ; declare -A conf=( $@ ) ; declare -p conf
> >                                                      declare -A
> > conf=(["opt1 val1 opt2 val2"]="" )
> >
> > >
> > i dont get it
> > im suure i used this working maany times
> >
> > -- why doesnt it split
>
> The parser sees
>
>     conf=( $@ )
>
> Now, the parser is not very bright, so it tries to figure out whether
> this command should be parsed as
>
>     conf=( [key]=value )
>
> or as
>
>     conf=( key value )
>
> It makes this decision WITHOUT expanding $@.  It doesn't matter what you
> put in $@.  The parser has already decided.
>

it was "$@" initially .. same effect

however myy biig problem with that story is .. im 100 sure i used such
expandings many
so this all breaks a bit my head
i 100 remember way more -A .. $onevar than mixed multivars usage

unicorn:~$ unset -v hash; declare -A hash
> unicorn:~$ set -- '[key1]=meow' '[key2]=bowwow'
> unicorn:~$ hash=( $@ )
> unicorn:~$ declare -p hash
> declare -A hash=(["[key1]=meow [key2]=bowwow"]="" )
>

i wonder what a=( $@ '' ) say
prolly not the right either

As we can see here, the parser has decided that it's dealing with
>
>     conf=( key value )
>
> where "key" is whatever $@ expands to, and "value" is the empty string,
> because there was no second word after the $@ in the command as it was
> parsed.  (And since there were no square brackets.)
>

-A doesnt expand {1..3} or ,,,

>
> Returning to the original question for a moment, the OP's question was
> how to determine the state of shell options like 'vi' that have no
> single-letter equivalent.  My suggestion was to parse the output of
> "set -o".  Another answer was to use test -o vi.
>
> Pursuing my answer for a moment, we start by looking at the output of
> set -o:
>
> unicorn:~$ set -o
> allexport       off
> braceexpand     on
> emacs           off
> errexit         off
> [...]
> vi              on
> xtrace          off
>
> So, how would we parse this to determine whether the 'vi' option is on
> or off?  Here's one way:
>
>     read -r _ state < <(set -o | grep '^vi')
>
> This will store 'on' or 'off' in the state variable.
>
> The "test -o vi" answer may be better than mine.
>

the test is cool
the set -p parsing for completeness

All the talk of associative arrays and the   array=( key value ... )
> syntax is just a digression.
>

ye

>