Re: Race condition in handling SIGHUP

[Chet Ramey]

On 4/28/16 11:23 AM, David Kaspar [Dee'Kej] wrote:
> On Thu, Apr 28, 2016 at 2:49 PM, Chet Ramey <chet.ramey@case.edu
> <mailto:chet.ramey@case.edu>> wrote:
> 
>     Because systems using a networked password database can hang at a priority
>     that doesn't interrupt the system call when a SIGHUP arrives.
> 
> ​I'm sorry, but what exactly do you mean by 'networked password database'?
> I'm not sure I follow you correctly

YP was the original one; now you have things like LDAP and other directory
services.  Something you use when you don't want to have everyone in the
password file.  These are somewhat out of favor now.

> ​
>  
> 
>     That code is written the way it is to accommodate the much more common
>     case of users exiting a shell by hitting the `close' button on their
>     terminal window, which causes the terminal emulator to send one or more
>     SIGHUPs to the shell process, usually while readline is active.  You want
>     shell to try and save the history in this case, since that's what users
>     expect.
> 
> ​Yeah, I think this particular problem about closing the terminal window
> and sending SIGHUP to bash has been discussed in this [1] mailing list
> thread quite thoroughly, and from what I've read in the mailing list
> before, it made quite sense to me.
> 
> So, my question right now is - do you think it's feasible to the code
> setting 'terminate_immediately = 1' from bash_tilde_expand() function?

I'm going to insert the missing verb into that sentence ("remove") and
answer "yes, we're going to try it and see how it goes."


-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/