bug#16493: 24.3.50; (setq search-invisible t) is useless, let's allow to turn visible-mode temporarily on

[Stefan Monnier]

> That's the wrong way of looking at the issue, if you want to talk
> about the aspects that I called "confusing".  I see no technical
> problems with using (t) instead of t, I just said that it confuses the
> heck out of me, because its semantics is not clear by just looking at
> the value.

I'm sorry, I misunderstood.  I don't see anything confusing about any of
this, I thought you were opposed to changing the default because you
thought it was a useful value (plus the usual backward compatibility
risks).

Regarding the possible new defaults and their possible confusing
qualities, then:

- could you explain what you find confusing about `(t)`?

- Personally I'm fine with either `(t)` and `nil` as default values.

- Currently an `invisible` property with value `t` "always" makes the
  text invisible without having to pay attention to
  buffer-invisibility-spec, so `(t)` is more backward compatible.  I put
  quotes around "always" because it's always possible for someone to
  come around and set buffer-invisibility-spec to nil by hand, and
  because until not too long ago, remove-from-invisibility-spec
  occasionally ended up setting buffer-invisibility-spec to nil.


        Stefan