Re: ignore(SIGPIPE) in gsm

[Elliot Lee <sopwith redhat com>]

On Mon, 8 Mar 1999, Miguel de Icaza wrote:

> > I implemented short patch to remove SIG_IGN handler from SIGPIPE before
> > exec. But there is also the same problem with panel, and it looks like
> > source of it is Orbit. I am going to join Orbit list, and lurk a bit
> > about CORBA, orbs, etc.
> 
> You are right.  ORBit is a source of broken SIGPIPE behaviour.
>
> For now, a simple restoration of the SIGPIPE handler to a sane default
> before executing would do it.
>
> But in the long term, ORBit should be modified to not set SIGPIPE to
> that value.  It should either provide a handler or should detect
> *properly* when a connection dies.

It does - the problem is that when it is doing that detection, SIGPIPE
gets generated. I am not doing a select() before every single write() to
triple-check that the fd has not suddenly changed state - it would slow
things down too much.

SIGPIPE does not work well with GUI programs, period.
-- Elliot
"In film you will find four basic story lines. Man versus man, man
 versus nature, nature versus nature, and dog versus vampire."
    - Steven Spielberg