On Wed, Dec 12, 2018 at 09:14:52AM +0000, Denis OSTERLAND wrote: > Consider following situation: > > tag C > v > A---B---C topic > / \ > D---E---F---G---H master > ^ > tag E > > PTXDIST_BSP_AUTOVERSION at commit G will be E-2-gXXXX, > but at commit H it will be C-3-gXXXX.
At H I'd expect C-1-gXXXX. I think in general using --first-parent is wrong. It's only in some workflows (i.e. when there is a single person that does "mainline merges") where the first parent is special. > With --first-parent it will be E-3-gXXXX. > This is more intuitive result, > especially when more than just one branch gets merged. > > Signed-off-by: Denis Osterland <[email protected]> > --- > scripts/kernel/setlocalversion | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/scripts/kernel/setlocalversion b/scripts/kernel/setlocalversion > index 3feb87dce..7eaaaf413 100755 > --- a/scripts/kernel/setlocalversion > +++ b/scripts/kernel/setlocalversion > @@ -55,12 +55,12 @@ if head=`git ${git_abbrev} rev-parse --verify --short > HEAD 2>/dev/null`; then > > # If we are at a tagged commit (like "ptxdist-2010.05.0"), we > # print it here > - if atag="`git describe --exact-match 2>/dev/null`"; then > + if atag="`git describe --first-parent --exact-match 2>/dev/null`"; then With --exact-match --first-parent is a noop, isn't it? > echo "$atag" | awk -F- '{printf("%s", $(NF))}' unrelated to this patch: if atag is "ELWMS-1.0-rc17" the result here is "rc17", isn't it? Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | http://www.pengutronix.de/ | _______________________________________________ ptxdist mailing list [email protected]
