Re: Reading the current timeout value

[Sören Brinkmann]

On Sat, 2014-03-15 at 02:55PM -0400, Thomas Dickey wrote:
> On Thu, Mar 13, 2014 at 09:18:56PM -0700, Sören Brinkmann wrote:
> > Hi,
> > 
> > I hope this list is open and I don't need to subscribe to post.
> > 
> > Anyway, I was recently writing a program and at some point I needed to
> > change things like the timeout and cursor visibility for a code section.
> > 
> > For cursor visibility this is easy. curs_set() returns the current
> > visibility, so that SW can store and restore that value.
> > 
> > Unfortunately I couldn't find any interface to read the current value of
> > timeout() and was pretty disappointed to see that timeout() and
> > wtimeout() are void functions.
> > 
> > So, my question is, is there a way to read the current timeout and I
> > just didn't find it?
> 
> no - it was overlooked.  The value that you are interested in is the
> "_delay" member of the WINDOW structure.  If WINDOW is not opaque,
> then you could simply refer to it, e.g.,
>       win->_delay
> 
> (I'm assuming that WINDOW is opaque, otherwise you would not be
> asking this question).

I actually didn't check. This works. IMHO, not as clean having an
API-call but good enough.

> 
> However, with the opaque feature as currently implmented (normally
> not opaque except for the ncursest / ncursestw libraries - but I
> seem to recall some issue with some packagers renaming these...),
> the only access is a boolean.
> 
> > Would it be acceptable to change the prototypes of timeout() and
> > wtimeout() to return the current timeout (see patch below)?
> 
> no (these functions have well-known prototypes).

Yeah, that was to be expected.

> 
> I can add another function, for the opaque configuration.

Probably a good idea. Not being able to obtain the current value seems
like a limitation in some cases.

        Thanks,
        Sören