This is the mail archive of the
cygwin-apps
mailing list for the Cygwin project.
Re: do package updates automatically kill running exes?
- From: Andrew Schulman <schulman dot andrew at epa dot gov>
- To: cygwin-apps at cygwin dot com
- Date: Fri, 14 Aug 2015 07:35:58 -0400
- Subject: Re: do package updates automatically kill running exes?
- Authentication-results: sourceware.org; auth=none
- References: <qscosattddeocejq80t47mllp22muvqt3p at 4ax dot com> <qscosattddeocejq80t47mllp22muvqt3p-e09XROE/p8c at public dot gmane dot org> <55CC6CE0 dot 9040908 at dronecode dot org dot uk> <0k0psats80nbi13clrq5tmv7pat7pfevlh at 4ax dot com> <0k0psats80nbi13clrq5tmv7pat7pfevlh-e09XROE/p8c at public dot gmane dot org> <871tf7vzl1 dot fsf at Rainer dot invalid> <1bqpsat4pv5lfnql0m1srui7pkvaspsd2t at 4ax dot com> <1bqpsat4pv5lfnql0m1srui7pkvaspsd2t-e09XROE/p8c at public dot gmane dot org> <55CCED9B dot 60200 at dronecode dot org dot uk>
> On 13/08/2015 20:02, Andrew Schulman wrote:
> >>> Anyway, that answers my question. If I want to be sure a process from one of my
> >>> packages exits, I need to put a killall into its preremove script.
> >>
> >> I've put a "pkill ." in all scripts that run setup (with the requisite
> >> are you sure? y/N).
> >
> > Ugh, both psmisc (killall) and procps (pkill) aren't in base, so both of
> > those may fail.
>
> What problem are you trying to solve here?
fish 2.2, when it comes out, will throw scary-looking errors and maybe
malfunction if any exes from fish 2.1.1 are still running. In particular fish
2.1.1 runs fishd to share environment variables between instances. fish 2.2
doesn't use fishd, and will throw errors if any instances of fishd are still
running.
So IMO the best option will be to killall fish in the preremove script for fish
2.1.1. I think the cost of having to restart your shell is pretty low, and will
be better than having users puzzle through the error messages from fish 2.2.