Ar an deichiú lá de mí na Nollaig, scríobh Stephen J. Turnbull:
Aidan Kehoe writes:
> Ar an deichiú lá de mí na Nollaig, scríobh Stephen J. Turnbull:
> > Now, not screwing up ASCII is more important than shouting Hrvoje's
> > name right by default.
> ASCII isn't screwed up.
Doing a search on a plain ASCII string in an all-ASCII buffer, and get
simple_string() instead of boyer_moore(), is screwed up.
No it’s not. ı is a valid case mapping for I, and always has been. Not to
mention that we should certainly, for interactive use, default to folding
accents as well as case, and basing such an approach on the current code
would have much the same behaviour.
There could be other such bugs.
> The bug is old.
Yeah, well, it still needs to be fixed if you want to load the
offending case tables by default.
The bug I’m talking about (the reverse search underrun) is something you’ve
fixed. I don’t view getting simple_string instead of boyer_moore as a bug.
An infelicity, perhaps, given that for the extant case tables it should be
possible to use boyer_moore; but not incorrect.
> > If I fix it, I will fix it by moving all case tables that
> > Mule charset boundaries to libraries that can be explicitly
> > `require'ed by users who want them.
> I will veto that.
Because it’s removing a feature for, to me, marginal to no benefit. How much
slower is simple_search than boyer_moore on representative data with modern
¿Dónde estará ahora mi sobrino Yoghurtu Nghé, que tuvo que huir
precipitadamente de la aldea por culpa de la escasez de rinocerontes?
XEmacs-Beta mailing list