On 1/21/19 6:09 PM, Dave Chinner wrote: > On Mon, Jan 21, 2019 at 11:33:16AM -0500, jeffm@xxxxxxxx wrote: >> From: Jeff Mahoney <jeffm@xxxxxxxx> >> >> In order to simplify combining excluded tests specified on the command >> line vs specified via config files, > > You mean defining excludes in configs/<hostname>.config files, > right? i.e. in config sections? Yes. > But the section code only calls _prepare_test_list if the test dev > is recreated by each section, right? So you can't really change the > expunge list from the config files without forcing a test dev > reformat, right? Right. In my testing I didn't hit this situation, but it's there. I can fix that. >> it makes sense to push the handling >> into _prepare_test_list. This means we start with a fresh $tmp.xlist >> and rebuild it each time _prepare_test_list is called. > > The patch does more than that, right? > >> Signed-off-by: Jeff Mahoney <jeffm@xxxxxxxx> >> --- >> check | 41 ++++++++++++++++++++++++----------------- >> 1 file changed, 24 insertions(+), 17 deletions(-) >> >> diff --git a/check b/check >> index 77a06b00..17073c4e 100755 >> --- a/check >> +++ b/check >> @@ -230,7 +230,28 @@ _prepare_test_list() >> done >> fi >> >> - # Specified groups to exclude >> + :> $tmp.xlist >> + >> + # Per-fstype/generic/shared file of tests to exclude (-X) >> + for xfile in $XGROUP_FILES; do > > That adds support for multiple group exclude files (i.e. multiple -X > options), right? Yes. I'll document that in the commit message. >> + for d in $SRC_GROUPS $FSTYP; do >> + [ -f $SRC_DIR/$d/$xfile ] || continue >> + for f in `sed "s/#.*$//" $SRC_DIR/$d/$xfile`; do >> + echo "$d/$f command line" >> $tmp.xlist >> + done >> + done >> + done >> + >> + # External file of tests to exclude (-E) >> + for xfile in $EXCLUDE_FILES; do >> + if [ -f $xfile ]; then >> + sed -e "s/#.*$//" \ >> + -e "s;$; file $xfile;" "$xfile" \ > > And I have no idea what problem this second expression is solving - > it wasn't in the original code that got copied here. This series is part of a larger set I've been using for a while. I reordered them and pulled out the least controversial. This was originally after a patch that adds reporting for why a test was expunged to the check output. We can ignore this part for now. Thanks for the review, -Jeff -- Jeff Mahoney SUSE Labs
