[PATCH 2/3] qf: Make qf_pull -f forcibly clean

[Lucas De Marchi]

On Thu, Mar 08, 2018 at 05:25:44PM -0800, Rodrigo Vivi wrote:
> On Thu, Mar 08, 2018 at 07:29:52AM -0800, Lucas De Marchi wrote:
> > On Wed, Mar 07, 2018 at 05:42:54PM -0800, Rodrigo Vivi wrote:
> > > qf_checkout implies that you need to have that baseline
> > > on your local repository, what it is not good for a distributed
> > > maintenance.
> > > 
> > > Let's make qf pull -f useful for the case we want to start
> > > a clean rebase from anywhere.
> > > 
> > > v2: Remove dubious comments and use -f.
> > > 
> > > Cc: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > > Cc: Michel Thierry <michel.thierry at intel.com>
> > > Cc: James Ausmus <james.ausmus at intel.com>
> > > Cc: Lucas De Marchi <lucas.demarchi at intel.com>
> > > Signed-off-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
> > > ---
> > >  qf | 23 +++++++++++++++++++++--
> > >  1 file changed, 21 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/qf b/qf
> > > index 270bcf53000a..fb04aeb71428 100755
> > > --- a/qf
> > > +++ b/qf
> > > @@ -404,11 +404,17 @@ function qf_pull
> > >  {
> > >  	cd_toplevel
> > >  
> > > -	qf fetch
> > > +	qf_fetch
> > >  	cd patches
> > >  	git pull --ff-only
> > >  
> > > -	qf co
> > > +	if [[ $FORCE ]]; then
> > 
> > if we want to force, don't we need to do this before the git pull?
> > 
> > 	if [ $FORCE ]; then
> > 	git pull --ff-only
> 
> Yes, this is probably a good idea. Although the real forced
> version is fetch & reset --hard ;)

not sure I understand. This would work iff this script was not using
'set -e'. Since it is, any command that fails will cause it to exit.

So AFAICS 'git pull --ff-only' would fail and we would not even check by
$FORCE. Am I missing anything?


Lucas De Marchi

> 
> > 
> > 	qf co
> > 
> > if we want to force, don't we need to do this before the git pull?
> 
> 
> there is a duplication here on your response that confused me...
> what do you want here besides the git pull --ff-only inside the forced block?
> 
> 
> > 
> > 
> > > +		git reset --hard HEAD
> > > +	else
> > > +		qf_co
> > > +	fi
> > > +
> > > +	cd ..
> > 
> > 
> > > +		git reset --hard HEAD
> > > +	else
> > > +		qf_co
> > > +	fi
> > > +
> > > +	cd ..
> > >  }
> > >  
> > >  function qf_stage
> > > @@ -587,6 +593,19 @@ function qf_usage
> > >  	echo "See '$qf help' for more information."
> > >  }
> > >  
> > > +FORCE=
> > > +while getopts f opt; do
> > > +	case "$opt" in
> > > +		f)
> > > +			FORCE=1
> > 
> > The way you are checking for $FORCE means that even if you assign
> > FORCE=0 the result will be true, which could be misleading (it will only
> > be false if the var is unsed). Maybe here you could do "FORCE=FORCE" or
> > assign 0 first and then in the check do a "if [ $FORCE -eq 1 ];" ?
> 
> I agree... just not sure if we should deviate from what is in use on dim.
> 
> > 
> > Lucas De Marchi
> > 
> > > +			;;
> > > +		*)
> > > +			echo "See '$qf help' for more information."
> > > +			exit
> > > +	esac
> > > +done
> > > +shift $((OPTIND - 1))
> > > +
> > >  # qf subcommand aliases (with bash 4.3+)
> > >  if ! declare -n subcmd=qf_alias_${subcommand//-/_} &> /dev/null || \
> > >  		test -z "${subcmd:-}"; then
> > > -- 
> > > 2.13.6
> > > 
> > > _______________________________________________
> > > dim-tools mailing list
> > > dim-tools at lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/dim-tools
> > _______________________________________________
> > dim-tools mailing list
> > dim-tools at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/dim-tools