Hello, all. With your permission I'd like to resurrect a topic that
seems to have died in 2002. At least, this is the most info I've
found. The question was, Why does modification of a constant produce
a warning and not an exception? Matz's answer, ruby-talk 48168, was
as follows (angle brackets added):
Hi,
>1. if someone attempts to modify a constant, why does ruby choose to
>emit a warning instead of a fatal error?If it is a fatal error, something like Ruby embedded editor will meet
serious problems, e.g.M-x eval-ruby-expression
Eval: Foo = 42<ret>
bang!>2. is there a beast called a 'global constant'? i tried defining $Foo
>and ruby allows me to rebind $Foo later.No. $Foo is a plain global variable.
matz.
Unfortunately, I'm not familiar with 'eval-ruby-expression', which
looks like some sort of Emacs command. So I still consider it a
fairly serious problem that modification of a constant produces
nothing more serious than a warning, which is useless without human
supervision. Especially because, even if an exception were raised
instead, a bypass system is still so easy to arrange:
class Module
def const_set!( name, val )
remove_const(name) if const_defined?(name)
const_set( name, val )
end
end
Is this 'embedded editor' problem the same as it was five years ago?
Could somebody who *does* know about embedded editors explain what it
is about them that makes constants so hard to deal with? Ruby-talk
48271, from the same thread:
Hi,
>> M-x eval-ruby-expression
>> Eval: Foo = 42<ret>
>> bang!
>
>What problems would there be with making it an exception?for example, eval-current-buffer may cause trouble if it redefines
contants.matz.
How do embedded editors normally handle exceptions?
Thanks to David Garamond for creating the prior thread.
- Jonathan
···
In message "warning modifying constant & 'global constant'" > on 02/08/26, David Garamond <davegaramond / icqmail.com> writes:
In message "Re: warning modifying constant & 'global constant'" > on 02/08/27, Paul Brannan <pbrannan / atdesk.com> writes: