Hi, On Thu, Apr 18, 2024 at 7:36 PM Sean Anderson <[email protected]> wrote: > > When a patch is added to a series after the initial version, there are no > changes to note except that it is new. This is typically done to suppress > the "(no changes in vN)" message. It's also nice to add a change to the > cover letter so reviewers know there is an additional patch. Add a tag to > automate this process a bit. > > There are two nits with the current approach: > > - It favors '-' as a bullet point, but some people may prefer '*' (or > something else) > - Tags (e.g. 'patman: ' in 'patman: foo bar') are not stripped. They are > probably just noise in most series, but they may be useful for treewide > series to distinguish 'gpio: frobnicate' from 'reset: frobnicate', so > I've left them in. > > Suggestions for the above appreciated. > > Suggested-by: Douglas Anderson <[email protected]> > Signed-off-by: Sean Anderson <[email protected]> > --- > > tools/patman/func_test.py | 2 ++ > tools/patman/patchstream.py | 5 +++++ > tools/patman/patman.rst | 13 +++++++++++++ > ...t-cast-for-sandbox-in-fdtdec_setup_mem_siz.patch | 1 + > tools/patman/test/test01.txt | 1 + > 5 files changed, 22 insertions(+) > > diff --git a/tools/patman/func_test.py b/tools/patman/func_test.py > index 3b4c9448882..af6c025a441 100644 > --- a/tools/patman/func_test.py > +++ b/tools/patman/func_test.py > @@ -293,6 +293,7 @@ Changes in v4: > change > - Some changes > - Some notes for the cover letter > +- fdt: Correct cast for sandbox in fdtdec_setup_mem_size_base() > > Simon Glass (2): > pci: Correct cast for sandbox > @@ -342,6 +343,7 @@ Changes in v4: > - Multi > line > change > +- New > - Some changes > > Changes in v2: > diff --git a/tools/patman/patchstream.py b/tools/patman/patchstream.py > index ec1ca874fb2..a09ae9c7371 100644 > --- a/tools/patman/patchstream.py > +++ b/tools/patman/patchstream.py > @@ -477,6 +477,11 @@ class PatchStream: > self.change_version = self._parse_version(value, line) > elif name == 'cc': > self.commit.add_cc(value.split(',')) > + elif name == 'added-in': > + version = self._parse_version(value, line) > + self.commit.add_change(version, '- New') > + self.series.AddChange(version, None, '- %s' % > + self.commit.subject)
I won't have a chance to track this down right now, but I figured I'd at least report it. Maybe it's already been looked at? Once before when I used this I saw a case where using "Commit-added-in" was doubling tags. AKA I end up seeing: - New - New When I came back to try to look at it later I couldn't reproduce it. I'm seeing it again now. I did the bare minimum to debug and I see that the code in process_line() is being run twice for the same line. The first seems to be in get_metadata_for_list() and the second in process_stream(). I can do a quick-n-dirty fix for this, but presumably it would be better to dig into the real root cause? -Doug
