Re: arrayfuncs: fix read of uninitialized mem
| От | Tom Lane |
|---|---|
| Тема | Re: arrayfuncs: fix read of uninitialized mem |
| Дата | |
| Msg-id | 3328.1095257892@sss.pgh.pa.us обсуждение |
| Ответ на | arrayfuncs: fix read of uninitialized mem (Neil Conway <neilc@samurai.com>) |
| Ответы |
Re: arrayfuncs: fix read of uninitialized mem
|
| Список | pgsql-patches |
Neil Conway <neilc@samurai.com> writes:
> I fixed it by initializing dims_str[0] to '\0' circa line 1018 in
> current sources. While doing so I couldn't resist the temptation to fix
> a few of arrayfunc.c's crimes against good programming practise, so the
> attached patch includes some additional cosmetic improvements.
I dislike what you did at lines 983-1012 (replace a local variable by
possibly-many stores into an array). Also, as long as we're fixing
unreadable code, how about (line 1019)
for (i = j = 0, k = 1; i < ndim; k *= dims[i++], j += k);
becomes
for (i = j = 0, k = 1; i < ndim; i++)
k *= dims[i], j += k;
or some such. The empty loop body is a mistake waiting to happen.
Looks fine to me otherwise.
regards, tom lane
В списке pgsql-patches по дате отправления: