Re: compiler warnings with GCC 4.5

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> read.c: In function nodeRead:
> read.c:370:3: warning: case value 1000001 not in enumerated type
> NodeTag

> This can be fixed by changing
> switch (type)
> to
> switch ((int) type)

No objection from here. We don't attempt to cover all possible NodeTags
in that switch anyway, so I don't see that we're losing any error
detection capability by adding the cast.

> preproc.y: In function base_yyparse:
> preproc.y:8043:19: warning: operation on yyval.str may be undefined

> The code in question looks like

> | SETOF SimpleTypename opt_array_bounds
> { $$ = $$ = cat_str(3, make_str("setof"), $2, $3.str); }
> ^^^^^^^

> It is converted from the main grammar, so it looks like a bug in the
> conversion routine.

The double assignment looks inefficient, but I bet what it's really unhappy
about is that one of the cat_str arguments has a .str suffix and the
other doesn't --- why is that?

regards, tom lane