How are exceptions actually implemented in assembly?

[Steven Majewski]

On 23 Jan 2002, Aahz Maruch wrote:

> Note carefully that the cost comes almost entirely during the handling
> of exceptions, not in setting up exception blocks.  So if as a
> percentage of instructions executing exceptions are rare, it's a very
> small penalty.

And that is the sense I've always thought of the term 'exception' -- i.e.
that it's the exceptional case that happens infrequently (relative
to the other cases). Most loops have a lot of repetitions but a
single exit. (And python builds up an exception block for 'for' loops
and uses an exception to exit the loop.) Tim, I think, was responsible
for the idiom (before there were defaults for dictionary .get()'s ):
	try:  return dict[key]
	except KeyError: return default
which, if missing keys are a relatively rare case, was more effecient
than testing if dict.has_key(key) for every access.

I don't think the problem is the term 'exception' so much as it is
that folks seem to think that they should use the same rules in
Python as they do in C++ or Java.

-- Steve