bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

[積丹尼 Dan Jacobson]

for i do
    true
done
    : This line and the whole rest of the program misindented because emacs shallow knowlege of shell grammar.
    : Test with TAB. Emacs only knows about:
    for i
    do
	true
    done
    : If this was python, well, you would have to have the language change to accomidate emacs.
    : Also if we are colaborating on a big project we cannot just rip up line of other peoples code so
    : they are not misindeted for only us.
    : Yes, some emacs programmers do not know shell grammar and just assume... indeed the word "do" on the
    : first line is not even in a different color. But the word in right here in this line is! Three times!
    : Hmmm, [including this part of] the sh language is, maybe 50 years
    : old. Not something checkbashisms --extra --force --posix has an issue with. It is part
    : of sh, bash, dash, ksh, ...
    : I.e., not something invented before emacs...
    : What could be even older? Well, sh must have been written in C...
    : Yes, I reported this before. But it was closed by programmers who
    : do not know basic Bourne shell "for" statement grammar.
    : emacs-version "27.1"

Indeed, I was just reading in RISKS Digest,

> https://www.wired.com/story/ai-latest-trick-writing-computer-code/

> What fun -- being second-guessed in real time by software that doesn't
> understand my code...

bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

[Gregory Heytings]

>
> for i do
>    true
> done
>    : This line and the whole rest of the program misindented because emacs shallow knowlege of shell grammar.
>

for i in a b c; do
     true
done
# This line and the whole rest of the program correctly indented because Jidanni now knows the shell grammar

bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

[Glenn Morris]

Duplicate of https://debbugs.gnu.org/26217 (a normal report).
No-one need (or should) read jidanni's claptrap.

bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

[積丹尼 Dan Jacobson]

$ man bash
SHELL GRAMMAR
for name [ [ in [ word ... ] ] ; ] do list ; done

bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

[Gregory Heytings]

>
> $ man bash
> SHELL GRAMMAR
> for name [ [ in [ word ... ] ] ; ] do list ; done
>

So what?  Of course it it's not feasible to embed a complete Bourne shell 
interpreter (or for that matter the interpreter of any language) in Emacs 
to cope with every possible exotic construct of that language.  I note 
that

for i
do
     true
done

and

for i; do
     true
done

are both indented as you expect.

bug#48193: Youngsters in charge of emacs vs. basic Bourne shell "for" statement

[積丹尼 Dan Jacobson]

Just like if for some language,

if(1){
        bla;
}

was dealt with fine, but

if(1)
{
        bla;
}

messed up the parenthesis counter...