On Tue, Oct 11, 2011 at 11:29:31AM -0500, Dan McGee wrote: > On Tue, Oct 11, 2011 at 11:13 AM, Lukas Fleischer > <[email protected]> wrote: > > There's no need to keep the lock file if we interrupt repo-{add,remove} > > using ^C (same applies to other signals). > Definite +1 here. > > > In contrast to a clean exit, we do not `rm -rf` the temporary directory > > here, so that it's still possible to analyze what went wrong. > I'm not sure what the right answer is here- I see where you're coming > from, but given that we give the user no indication we're leaving a > mess behind, I think we should either clean it up or tell the user we > did not. Also noticed that we print an error message with TERM even > when HUP and QUIT are caught.
I'm not really sure either, but I'd say that showing a message that includes "$tmpdir" might be the best thing to do here. I also submitted a separate patch that fixes the SIGTERM message issue. > > > > Signed-off-by: Lukas Fleischer <[email protected]> > > --- > > Hrm. Second time I had to manually remove > > "/srv/ftp/community/os/$arch/community.db.tar.gz.lck" on sigurd in order > > not to lock the whole database after interrupting `/arch/db-update`. > > > > I hope I didn't miss anything here but this should hopefully not cause > > any other unforeseen issues... > > > > scripts/repo-add.sh.in | 3 +++ > > 1 files changed, 3 insertions(+), 0 deletions(-) > > > > diff --git a/scripts/repo-add.sh.in b/scripts/repo-add.sh.in > > index 29b150c..753200b 100644 > > --- a/scripts/repo-add.sh.in > > +++ b/scripts/repo-add.sh.in > > @@ -522,6 +522,9 @@ remove() { > > trap_exit() { > > echo > > error "$@" > > + > > + (( CLEAN_LOCK )) && [[ -f $LOCKFILE ]] && rm -f "$LOCKFILE" > > + > > exit 1 > > } > > > > -- > > 1.7.7 > > > > > > >
