Re: [PATCH] tags: add a space regex to DECLARE_BITMAP

From: Marc Zyngier
Date: Tue Nov 02 2021 - 12:35:39 EST


On Tue, 02 Nov 2021 14:25:19 +0000,
Zackary Liu <zackary.liu.pro@xxxxxxxxx> wrote:
>
> Hi Marc,
>
> On Tue, Nov 2, 2021 at 12:53 AM Marc Zyngier <maz@xxxxxxxxxx> wrote:
> >
> > On 2021-11-01 15:59, Zhaoyu Liu wrote:
> > > When "make tags", it prompts a warning:
> > >
> > > ctags: Warning: drivers/pci/controller/pcie-apple.c:150:
> > > null expansion of name pattern "\1"
> > >
> > > The reason is that there is an indentation beside arguments of
> > > DECLARE_BITMAP, but it can parsed normally by gtags. It's also
> > > allowed in C.
> > >
> > > So fix this regex temporarily, and wait for better solutions
> > > applied to other regexs.
> > >
> > > Reviewed-by: Marc Zyngier <maz@xxxxxxxxxx>
> >
> > No, please! I never reviewed this patch. I *suggested*i it [1]
> > as a potential avenue for improvement.
> >
> > You cannot put this tag on a patch unless someone explicitly
> > grants it.
> >
>
> I'm really sorry that I used the wrong "Reviewed-by" tag,
> Perhaps "Suggested-by" tag would be more appropriate.

Yes, that'd be the right one.

>
> Sorry again.
>
> > > Signed-off-by: Zhaoyu Liu <zackary.liu.pro@xxxxxxxxx>
> > > ---
> > > scripts/tags.sh | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/scripts/tags.sh b/scripts/tags.sh
> > > index b24bfaec6290..7e5f19391f20 100755
> > > --- a/scripts/tags.sh
> > > +++ b/scripts/tags.sh
> > > @@ -186,7 +186,7 @@ regex_c=(
> > >
> > >
> '/\<DEFINE_\(RT_MUTEX\|MUTEX\|SEMAPHORE\|SPINLOCK\)(\([[:alnum:]_]*\)/\2/v/'
> > >
> '/\<DEFINE_\(RAW_SPINLOCK\|RWLOCK\|SEQLOCK\)(\([[:alnum:]_]*\)/\2/v/'
> > > '/\<DECLARE_\(RWSEM\|COMPLETION\)(\([[:alnum:]_]\+\)/\2/v/'
> > > - '/\<DECLARE_BITMAP(\([[:alnum:]_]*\)/\1/v/'
> > > + '/\<DECLARE_BITMAP([[:space:]]*\([[:alnum:]_]*\)/\1/v/'
> > > '/\(^\|\s\)\(\|L\|H\)LIST_HEAD(\([[:alnum:]_]*\)/\3/v/'
> > > '/\(^\|\s\)RADIX_TREE(\([[:alnum:]_]*\)/\2/v/'
> > > '/\<DEFINE_PER_CPU([^,]*, *\([[:alnum:]_]*\)/\1/v/'
> >
> > More importantly, and assuming this is the correct approach,
> > why should we limit this to DECLARE_BITMAP()?
>
> I will try to come up with a better approach, and commit [ patch v2
> ] later.

You could try and apply the same approach consistently for all
macros. Since you are using ctags, you're in a good position to test
it and work out whether this is a sensible approach.

Thanks,

M.

--
Without deviation from the norm, progress is not possible.