Re: [PATCH 04/12] fstests: fix group list generation for whacky test names

["Darrick J. Wong" <djwong@kernel.org>]

On Fri, May 20, 2022 at 05:25:15PM +0800, Zorro Lang wrote:
> On Fri, May 20, 2022 at 04:54:45PM +0800, Zorro Lang wrote:
> > On Fri, May 20, 2022 at 04:36:47PM +0800, Zorro Lang wrote:
> > > On Tue, May 17, 2022 at 05:01:03PM +1000, Dave Chinner wrote:
> > > > From: Dave Chinner <dchinner@redhat.com>
> > > > 
> > > > Darrick noticed that tests/xfs/191-input-validation didn't get
> > > > generated properly. Fix the regex to handle this.
> > > > 
> > > > $ grep -I -R "^_begin_fstest" tests/xfs | \
> > > >   sed -e 's/^.*\/\([0-9]*\):_begin_fstest/\1/' |grep 191
> > > > tests/xfs/191-input-validation:_begin_fstest auto quick mkfs realtime
> > > > $
> > > > $ grep -I -R "^_begin_fstest" tests/xfs | \
> > > >   sed -e 's/^.*\/\([0-9]*\).*:_begin_fstest/\1/ ' |grep 191
> > > > 191 auto quick mkfs realtime
> > > > $
> > > > 
> > > > Use the regexes for matching test names defined in common/test_names
> > > > rather than trying to open code it.
> > > > 
> > > > Signed-off-by: Dave Chinner <dchinner@redhat.com>
> > > > ---
> > > 
> > > Hi Dave,
> > > 
> > > After testing, looks like this patch brings in a regression issue, it causes
> > > case number aren't sorted in group.list, then the xfstests/new program can't
> > > get a right new case number which is unused. For example:
> > 
> > Oh, this issue isn't from this patch, it's from 441606d2 ("fstests: faster group
> > file creation"). Hmm... I saw your used "sort -u" in that patch, I'm going to
> > look into what's wrong with that.
> > 
> > Thanks,
> > Zorro
> > 
> > > 
> > > # make
> > > # ./new generic
> > > Next test id is 019
> > > Append a name to the ID? Test name will be 019-$name. y,[n]: ^C
> > > # ls tests/generic/019
> > > tests/generic/019
> > > 
> > > Then comparing the tests/generic/group.list with old generic/group.list, found
> > > lots of difference, due to the number in new group.list isn't sorted from
> > > small to big (I haven't gotten chance to check if there're missing number).
> > > 
> > > Thanks,
> > > Zorro
> > > 
> > > >  tools/mkgroupfile | 5 ++++-
> > > >  1 file changed, 4 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/tools/mkgroupfile b/tools/mkgroupfile
> > > > index 24435898..414cb538 100755
> > > > --- a/tools/mkgroupfile
> > > > +++ b/tools/mkgroupfile
> > > > @@ -19,6 +19,8 @@ if [ ! -x ../../check ]; then
> > > >  	exit 1
> > > >  fi
> > > >  
> > > > +. ../../common/test_names
> > > > +
> > > >  cleanup()
> > > >  {
> > > >  	rm -f $new_groups.check
> > > > @@ -60,7 +62,8 @@ ENDL
> > > >  
> > > >  	# Aggregate the groups each test belongs to for the group file
> > > >  	grep -I -R "^_begin_fstest" $test_dir/ | \
> > > > -		sed -e 's/^.*\/\([0-9]*\):_begin_fstest/\1/' >> $new_groups
> > > > +		sed -e "s/^.*\/\($VALID_TEST_NAME\):_begin_fstest/\1/" \
> > > > +		>> $new_groups
> 
> I think add a "sort -ug" [1] will help to fix this problem.

Yes please, it's much easier for us humans to index the file if it's
already sorted numerically.

> I'm wondering
> why we use ">> $new_groups" at here, why an appending write is needed. I think
> there's not a 'loop running' for this code, right? If there's, please correct
> me, then the 'sort -ug' have to be moved to other place.

The function writes a header into the group list file warning readers
that it's an autogenerated file.

--D

> Thanks,
> Zorro
> 
> [1]
> -               sed -e "s/^.*\/\($VALID_TEST_NAME\):_begin_fstest/\1/" \
> -               >> $new_groups
> +               sed -e "s/^.*\/\($VALID_TEST_NAME\):_begin_fstest/\1/" | \
> +               sort -ug > $new_groups
> 
> I'm wondering why we use
> 
> > > >  
> > > >  	# Create the list of unique groups for existence checking
> > > >  	grep -I -R "^_begin_fstest" $test_dir/ | \
> > > > -- 
> > > > 2.35.1
> > > > 
>