Curses ESC character and protected objects

Curses ESC character and protected objects

[Sandro Binetti]

Hi group.

My UnixWare 7.1.1 software is written in ADA, compiled by GNAT 3.13a1,
and uses a binding to Unix curses library for I/O.

The keyboard ESC key has a strange behaviour. Normally, when the ESC
key is pressed, after a delay of about 0.5 sec, it's correctly managed
by the program (in this 0.5 sec delay curses waits for some other
character, in order to correctly manage esc sequences coming from
other keyboard keys).

But since the moment I introduced a few protected objects to my
software, the behaviour of the ESC key is completely changed: I have
to press the ESC key twice to make it go right (it seems that the
original 0.5 sec delay has become an infinite timeout ...)

I don't search for a solution to my problem, I only have wish to
understand the possible reason of this strange thing.

TIA, Sandro.

Re: Curses ESC character and protected objects

[Warren W. Gay VE3WWG]

Sandro Binetti wrote:

> Hi group.
> 
> My UnixWare 7.1.1 software is written in ADA, compiled by GNAT 3.13a1,
> and uses a binding to Unix curses library for I/O.
> 
> The keyboard ESC key has a strange behaviour. Normally, when the ESC
> key is pressed, after a delay of about 0.5 sec, it's correctly managed
> by the program (in this 0.5 sec delay curses waits for some other
> character, in order to correctly manage esc sequences coming from
> other keyboard keys).
> 
> But since the moment I introduced a few protected objects to my
> software, the behaviour of the ESC key is completely changed: I have
> to press the ESC key twice to make it go right (it seems that the
> original 0.5 sec delay has become an infinite timeout ...)


It sounds like to me your Ada implementation does not grok the
character I/O timeout situation correctly. There is a V_TIME
parameter (IIRC) that curses will use for the 0.5 second
timeout, reading terminal input. Assuming you have no memory
corruption happening in your program, it sounds to me like
you have an Ada library implementation problem WRT the V_TIME
timeout handling.

Warren.