Re: towards 4.4

[Ethan M. <merritt@u.washington.edu>]

On Sunday 31 January 2010, Tatsuro MATSUOKA wrote:
> Hello
> 
> --- Ethan Merritt <merritt@u.washington.edu> wrote:
> 
> > I have this listed already as an un-fixed bug.
> > "pause mouse" should return immediately, perhaps with a warning,
> > if the there is  no mousable window present.
> > Here are my notes:
> > 
> >  - If an x11 plot window is killed, then a subsequent 'pause mouse'
> >    command should be rejected.  The only way I have managed to set this up
> >    so far is to have event_reset() set term_initialised = FALSE,
> >    and have pause_command() test for it.  This seems to work for x11, but
> >    I have no idea what havoc it might trigger on os2 or whatever else.
> 
> Sorry I have not noticed the above.
> Surely gnuplot with x11 on cygwin (in this case, cygwin can be included as a unixy system),
> 
> pause mouse" return immediately, perhaps with a warning, if the there is  no mousable window present.
> 
> gnuplot> pause mouse
>          Mousing not active

That is because you have not yet opened a plot window at all, right?
Therefore gnuplot_x11 has not yet been started.
The problem arises if there was previously a plot window in use,
but it was forcibly closed by the user.

> In x11.trm,
>     /* If ipc_back_fd is not open, we will never see any mouse events! */
>     else if (paused_for_mouse) {
> 	paused_for_mouse = 0;
> 	int_error(NO_CARET,"Mousing not active");

That test only confirms that there is an open pipe to gnuplot_x11, 
not that there is currently a plot window on the screen.

> Perhaps on windows terminal can be done this by 'graphwin.hWndGraph'.
> For wxt terminal, I do not have enough knowledge to do that.
> 
> BTW, does not wxt terminal on the linux have the same problem?

On linux it is always safe to hit "ctrl-c" in order to return to
the command prompt. As I understand it, this solution does not work
on windows (see other thread).