Re: LYNX-DEV Control-Z and SIGSTOP mysteries

[Klaus Weide]

On Tue, 20 May 1997, Foteos Macrides wrote:
> Klaus Weide <address@hidden> wrote:
>
> >try inserting a statement
> >
> >    signal(SIGTSTP, SIG_DFL)
> >
> >a) before the setup(terminal) call,
> >
> >OR
> >
> >b) before the mainloop() call.
> >
> >Does it change anything?
> 
>       I don't understand your logic for that suggestion.  

Just a suggestion for testing, not for a permanent change.
I looked at what ncurses does, it installs an interrupt handler for
SIGTSTP but only if it finds the action to be SIG_DFL when it is
initialized.

> Lynx doesn't do anything itself for SIGTSTP 

...except in the for-unix SLANG start_curses()...

>                                 unless the command line switch for
> disabling escapes to shell was included.  Including that signal()
> call wouldn't change anything normally, [..]

It might have (in the first location) if the initscr() (whatever gets
executed for Hiram's curses) depends on SIG_DFL for correct
initialization, and for some strange reason that wasn't set.

Anyway, I am not sure what exactly Hiram's symptoms are, I didn't really
follow his description.. but it seems something with his curses is wrong,
right?

   Klaus


;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;