A modest indentation proposal

[Erann Gat]

In article <WBUN7.9250$FP6.336989 at atlpnn01.usenetserver.com>, "Steve
Holden" <sholden at holdenweb.com> wrote:

> Perhaps it wouldn't be "illegal". I think the point was that there may well
> be existing Python programs which are currently syntactically correct whose
> meaning would be altered by the adoption of your rule. If under present
> syntax I write
> 
>     def a(x,y):
>         stm1
>         stm2
>     stm3;
> 
> this is currently legal Python in which stm3 is not a part of the function
> body.  Would your rule treat this as a syntax error, or observe the
> indentation and accept it? Or something else?

It would, if the 'activate-paranoid-indentation-mode' flag were enabled,
signal a syntax error, or if that's unacceptable, issue a warning, e.g.
"Warning: the block structure defined by indentation does not match that
implied by punctuation.  Check your indentation."  Or something like that.

BTW, I am assuming that there's not a huge installed base of Python code
with trailing semicolons (since they're useless at the moment).

E.