Re: Avoid unecessary MemSet call (src/backend/utils/cache/relcache.c)
| От | Ranier Vilela |
|---|---|
| Тема | Re: Avoid unecessary MemSet call (src/backend/utils/cache/relcache.c) |
| Дата | |
| Msg-id | CAEudQApjWmVbhgzeKnCVnodrYr-c6n-tn+Bx=8BuY0O5GZ5LUA@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: Avoid unecessary MemSet call (src/backend/utils/cache/relcache.c) (Julien Rouhaud <rjuju123@gmail.com>) |
| Список | pgsql-hackers |
Em qui., 11 de ago. de 2022 às 09:23, Julien Rouhaud <rjuju123@gmail.com> escreveu:
On Thu, Aug 11, 2022 at 08:51:53AM -0300, Ranier Vilela wrote:
> Em qui., 11 de ago. de 2022 às 08:48, Alvaro Herrera <
> alvherre@alvh.no-ip.org> escreveu:
>
> > On 2022-Aug-11, Ranier Vilela wrote:
> >
> > > According to:
> > > https://interrupt.memfault.com/blog/c-struct-padding-initialization
> >
> Did you see the Strategy 3 table, { 0 } ?
It explicitly shows that at least Ubuntu clang version 13.0.0-2 with -01
doesn't do anything about the padding bytes (and that's after testing only 2
different compilers). Even if those compilers didn't show any problem, we
still couldn't rely on an undefined behavior and assume that no other compilers
behave differently.
Yeah, although not a problem in the main current compilers clang, gcc and msvc,
it seems that this cannot be changed.
it seems that this cannot be changed.
Being an undefined behavior, filling structures with holes, it seems to me that you should always use MemSet or memset.
Since even a current structure without holes could be changed in the future and become a bug.
Since even a current structure without holes could be changed in the future and become a bug.
regards,
Ranier Vilela
В списке pgsql-hackers по дате отправления: