Re: [PATCH] btrfs-progs: enable -Wmissing-prototypes for debug builds

[David Sterba <dsterba@suse.cz>]

On Tue, May 02, 2023 at 08:00:57PM +0800, Qu Wenruo wrote:
> 
> 
> On 2023/5/2 19:41, David Sterba wrote:
> > On Tue, May 02, 2023 at 09:29:30AM +0800, Qu Wenruo wrote:
> >> During development I'm a little surpirsed we don't have
> >> -Wmissing-prototypes enabled even for debug builds.
> > 
> > The build supports W=levels like kernel and -Wmissing-prototypes is in
> > level 1. In some cases we may want to add the warnings to be on by
> > default for debugging builds though.
> 
> I just did a quick search for the word "missing" of progs Makefile, 
> unforunately no hit, thus I doubt if it's even in debug level 1.
> 
> And the missing prototypes warning is by default enabled for kernel.

$ grep missing Makefile.extrawarn
warning-1 += -Wmissing-declarations
warning-1 += -Wmissing-format-attribute
warning-1 += $(call cc-option, -Wmissing-prototypes)
warning-1 += $(call cc-option, -Wmissing-include-dirs)
warning-1 += $(call cc-disable-warning, missing-field-initializers)
warning-2 += $(call cc-option, -Wmissing-field-initializers)

> [...]
> >>   
> >> +#include "sha.h"
> > 
> > This does not seem necessary, include whole file just for one prototype.
> 
> This is necessary as we would define a global function, without 
> including the header we got the missing prototype warning.
> 
> All the other comments make sense and I would update the patchset.
> 
> 
> The only other concern is, would this extra warning causing more hassles 
> for Josef to sync the kernel and progs code?

I don't know yet but it is possible that it would.  In that case the separate
patch would make it easier to enable once all warnings are fixed, I can keep it
at the top of the branch.