On Mon, 12 May 2025 at 16:18, Cédric Le Goater <c...@kaod.org> wrote:
>
> On 5/12/25 15:52, Cédric Le Goater wrote:
> > On 5/8/25 12:42, Peter Maydell wrote:
> >> The logic goes:
> >> * when we see a "new file" line, we set expect_spdx to 1
> >> * when we see an SPDX tag we set expect_spdx to 0
> >> * at a later point, if we still have expect_spdx 1 we complain
> >>
> >> The problem is that the "later point" here is "we saw another
> >> 'new file' line", not "we got to the end of all the parts of
> >> the patch that touch this file". So if the patch adds two
> >> new files then we'll warn for the first one (when we see the
> >> "new file" line for the secand new file), but we won't warn
> >> for a patch which adds only one new file (there's never a second
> >> "new file" line) or for the last new file added in a patch
> >> that adds multiple files.
> >>
> >> I'm not sure where the "complain if expect_spdx is 1" check
> >> should go, but I don't think it should be here...
> >
> > yes. I just made the same observation on this patch :
> >
> >
> > https://lore.kernel.org/qemu-devel/20250509163645.33050-7-rre...@linux.ibm.com/
> >
>
> This seems to work well enough.
Dan sent out a patchset that overhauls the spdx detection
logic last week ("
scripts/checkpatch: fix
SPDX-License-Identifier
detection"):
https://www.mail-archive.com/qemu-devel@nongnu.org/msg1113638.html
(Unfortunately lore.kernel.org seems to be missing it,
and patchew was unsubscribed to qemu-devel emails over the
weekend so it doesn't have it either.)
thanks
-- PMM
