Add "-Wimplicit-fallthrough" to default flags (was Re: pgsql:Support FETCH FIRST WITH TIES) - Mailing list pgsql-hackers

From Julien Rouhaud
Subject Add "-Wimplicit-fallthrough" to default flags (was Re: pgsql:Support FETCH FIRST WITH TIES)
Date
Msg-id 20200412081825.qyo5vwwco3fv4gdo@nol
Whole thread Raw
Responses Re: Add "-Wimplicit-fallthrough" to default flags (was Re: pgsql: Support FETCH FIRST WITH TIES)  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Sat, Apr 11, 2020 at 02:47:34PM -0400, Tom Lane wrote:
> Julien Rouhaud <rjuju123@gmail.com> writes:
> > On Tue, Apr 7, 2020 at 10:28 PM Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> >> Support FETCH FIRST WITH TIES
> 
> > FTR I now get the following when compiling with -Wimplicit-fallthrough -Werror:
> 
> Yeah, assorted buildfarm animals are mentioning that too.  I wonder
> if we should add that to the default warning options selected by
> configure?  I don't remember if that's been discussed before.
> 

I'm all for it.  It seems like a trap easy to catch up early, and we do want to
enforce it anyway.  I'm attaching a simple patch for that if needed, hopefully
with the correct autoconf version.

> > It seems that this fall-through comment:
> >                         /* fall-through */
> > Is not recognized by my compiler (gcc (Gentoo 9.3.0 p1) 9.3.0).    If
> > that's something we should fix, PFA a naive patch for that.
> 
> Hmm, I feel like this logic is baroque enough to need more of a rewrite
> than that :-(.  But not sure exactly what would be better, so your
> patch seems reasonable for now.  The comments could use some help too.

Yes I just checked the state machine to make sure that the fallthrough was
expected, but the comments are now way better, thanks!

Attachment

pgsql-hackers by date:

Previous
From: Noah Misch
Date:
Subject: Re: cleaning perl code
Next
From: Julien Rouhaud
Date:
Subject: Re: pg_stat_statements issue with parallel maintenance (Was Re: WALusage calculation patch)