Re: lynx-dev dev21: What broke "quit on anything other than y"?

[Bela Lubkin]

Kim DeVaughn wrote:

> On Tue, Mar 30, 1999, Doug Kaufman (address@hidden) said:
> |
> | On Tue, 30 Mar 1999, John Bley wrote:
> | >
> | > What broke the old behavior of quitting on anything other than y at the
> | > "Are you sure you want to quit? (y)" prompt?  Bouncing on "q" twice is in
> | > my blood now.
> |
> | Are you sure that your have QUIT_DEFAULT_YES set to TRUE in lynx.cfg?
> | The behavior is configurable.
> 
> That cfg var only changes the sense of the confirmation msg at q(uit)
> time (if TRUE, the default is y(es); if FALSE, it's n(o)).
> 
> Previously, one could hit q(uit) a 2nd time (in response to the confirm
> prompt), and you'd exit ... now you don't, as the 2nd "q" is ignored
> (probably anything other than "y" or "n", though I haven't checked).
> 
> This new, changed behavior is probably more technically correct, but I
> too certainly prefer the old behavior.

I also prefer the old behavior.

The stanza in lynx.cfg says:

# If QUIT_DEFAULT_YES is TRUE then when the QUIT command is entered, any
# response other than n or N will confirm.  It should be FALSE if you
# prefer the more conservative action of requiring an explicit Y or y to
# confirm.  The default defined here will override that in userdefs.h.

But this no longer matches Lynx's behavior.  The current behavior is
approximately:

  If QUIT_DEFAULT_YES is TRUE then when the QUIT command is entered, the
  resulting "are you sure" prompt shows "(y)" as the default, whereas if
  it is false, the prompt shows "(n)".  The prompt then accepts only
  "y", "n", or <Return> (which accepts the specified default).

IMHO, this is stupid.  What's apparently wanted (between all the
different factions) is a multivalued option.  The above choices would be
two of the values.  Another would be "'y' to quit, any other key to
return", and yet another would be "'n' to return, any other key to
quit".  Not that I really see the need for all of these.  The choices
*I* see as useful are:

  1. The conservative setting: accept *only* "y" or "n" [or the local
     language's equivalents, if we ever get to that] -- no default.

  2. The liberal setting: accept "n" [or the local equivalent] to mean
     "my mistake, do not quit"; anything else as "go ahead, quit".

and perhaps, for symmetry:

  3. The silly setting: accept "y" [or the local equivalent] to mean
     "go ahead, quit"; anything else as "my mistake, do not quit".

So change the name to QUIT_DEFAULT and give it three values: NONE, TRUE,
FALSE, respectively.

>Bela<