Обсуждение: gcc -Wimplicit-fallthrough and pg_unreachable

I noticed in CI builds of PL/Java with PG 13 that -Wimplicit-fallthrough=3
in pg_config's CFLAGS was causing some switch case fallthrough warnings
after an elog(ERROR. [1]

I added my own pg_unreachable() after the elog(ERROR, ...) calls [2]
and that did away with the warnings.

But it looks odd, and shouldn't it be unnecessary if elog is supplying
its own pg_unreachable() in the elevel >= ERROR case? [3]

Has anyone else seen this behavior? Am I doing something that stops gcc
from recognizing ERROR as a compile-time constant? Is it just a gcc bug?

(Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0 seems to be the gcc version on the
Travis runner.

Regards,
-Chap




[1] https://travis-ci.com/github/tada/pljava/jobs/447208803#L694

    AFAICT, the #L694 doesn't really take you to line 694 in a Travis log
    because they collapse lines so the browser will not see the anchor,
    so one must click to expand the containing line range first. Irksome.
    Copied below.

[2]
https://github.com/tada/pljava/blame/2c8d992/pljava-so/src/main/c/type/Type.c#L246

[3]
https://git.postgresql.org/gitweb/?p=postgresql.git;a=blob;f=src/include/utils/elog.h;h=a16f925#l104





In file included from /usr/include/postgresql/13/server/postgres.h:47:0,
                 from
/home/travis/build/tada/pljava/pljava-so/src/main/c/type/Type.c:13:
/usr/include/postgresql/13/server/utils/elog.h:125:5: warning: this
statement may fall through [-Wimplicit-fallthrough=]
  do { \
     ^
/usr/include/postgresql/13/server/utils/elog.h:145:2: note: in expansion of
macro ‘ereport_domain’
  ereport_domain(elevel, TEXTDOMAIN, __VA_ARGS__)
  ^~~~~~~~~~~~~~
/usr/include/postgresql/13/server/utils/elog.h:215:2: note: in expansion of
macro ‘ereport’
  ereport(elevel, errmsg_internal(__VA_ARGS__))
  ^~~~~~~
/home/travis/build/tada/pljava/pljava-so/src/main/c/type/Type.c:256:3: note:
in expansion of macro ‘elog’
   elog(ERROR, "COERCEVIAIO not implemented from (regtype) %d to %d",
   ^~~~
/home/travis/build/tada/pljava/pljava-so/src/main/c/type/Type.c:258:2: note:
here
  case COERCION_PATH_ARRAYCOERCE:
  ^~~~

Chapman Flack <chap@anastigmatix.net> writes:
> I noticed in CI builds of PL/Java with PG 13 that -Wimplicit-fallthrough=3
> in pg_config's CFLAGS was causing some switch case fallthrough warnings
> after an elog(ERROR. [1]

Yeah, I can replicate this here (gcc 8.3.1 on RHEL8).  My recollection
is that we saw this when trialling -Wimplicit-fallthrough, and determined
that the hack used to teach the compiler that elog(ERROR) doesn't return
fails to prevent -Wimplicit-fallthrough warnings even though it does work
for other purposes.  Using this test case:

int
foo(int p)
{
    int x;

    switch (p)
    {
        case 0:
            x = 1;
            break;
        case 1:
            elog(ERROR, "bogus");
            break;
        case 2:
            x = 2;
            break;
        default:
            x = 3;
    }

    return x;
}

I do not get a warning about x being possibly uninitialized (so it
knows elog(ERROR) doesn't return?), but without the "break" after
elog() I do get a fallthrough warning (so it doesn't know that?).

Seems like a minor gcc bug; no idea if anyone's complained to them.
I think we've found other weak spots in -Wimplicit-fallthrough's
coverage though, so it's not one of gcc's best areas.

As illustrated here, I'd just add a "break" rather than
"pg_unreachable()", but that's a matter of taste.

            regards, tom lane