dictionary test + patch

[Brian T Rice]

On Wed, 21 Jan 2004, Lendvai Attila wrote:

>
> hi!
>
> i've made some tests for dictionary.slate, and fixed two bugs i've
> discovered with it.

Great! Thanks.

> i've also read the logs about breakOn: my intention with it is to
> quietly catch a signal and continue execution. just like the Exhaustion
> signal from streams... it makes code more readable. (altough i don't
> know if it works the way i've written... :) and also it adds a speed
> penalty i can't judge, but it's not a real issue. or maybe, it could
> even be cheaper if i loop on a zillion of stream elements and skip a
> zillion atEnd calls by using a signal handler, isn't it?

Yes, that's the benefit of good use of exceptions: the client library can
decouple their handling and detection of problems. (Much easier when you
just need an on:do:, and the library called takes care of itself; it knows
about itself better.)

> how is ensure similar to breakOn:?

Lee probably should explain that.

> sugar.slate is not included by default, nor is it used anywhere in the
> sources. should it be avoided, or it's just ignored until there's a
> bootstrapped system where you can use it everywhere?

It's mostly a question of syntax and Smalltalk right now. Those methods
represent things that you can't do in Smalltalk syntax, so we've kept it
separate so far. Generally, I'd like these to be avoided until we do have
a self-hosted IDE to make a decision about this (and automate that
decision). I'm also aiming for consistency here.

-- 
Brian T. Rice
LOGOS Research and Development
http://tunes.org/~water/