Re: Warnings building 8.5alpha3 on Solaris 10

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: Warnings building 8.5alpha3 on Solaris 10
Дата
Msg-id 1846.1265148760@sss.pgh.pa.us
обсуждение исходный текст
Ответ на Warnings building 8.5alpha3 on Solaris 10  (Isaac Morland <ijmorlan@uwaterloo.ca>)
Список pgsql-bugs
Isaac Morland <ijmorlan@uwaterloo.ca> writes:
> I get the following warnings building 8.5alpha3 on Solaris 10.
> I do not get these warnings building 8.4 exactly the same way.

> standby.c:239: warning: int format, pid_t arg (arg 3)
> standby.c:247: warning: int format, pid_t arg (arg 5)

That looks like a real problem, but it seems to be cleaned up already
in HEAD.

> descriptor.c:23: warning: missing braces around initializer
> descriptor.c:23: warning: (near initialization for `descriptor_once.__pthread_once_pad')
> memory.c:73: warning: missing braces around initializer
> memory.c:73: warning: (near initialization for `auto_mem_once.__pthread_once_pad')
> connect.c:16: warning: missing braces around initializer
> connect.c:16: warning: (near initialization for `actual_connection_key_once.__pthread_once_pad')
> misc.c:61: warning: missing braces around initializer
> misc.c:61: warning: (near initialization for `sqlca_key_once.__pthread_once_pad')

These all seem to relate to uses of PTHREAD_ONCE_INIT, which is a
system-provided macro --- so I think you need to complain to Sun
that their headers provide warning-inducing declarations.  It should
be just cosmetic as far as functionality goes, though.

> auth.c:82: warning: initialization from incompatible pointer type
> (except in 8.4 it's at line 80)

This one probably is because of a discrepancy in "const" decorations of
parameters for the PAM callback proc.  Again, it's cosmetic, and seems
more likely to indicate obsolete Solaris headers than anything we should
change on our end.

            regards, tom lane

В списке pgsql-bugs по дате отправления:

Предыдущее
От: Jasen Betts
Дата:
Сообщение: whole-row functional index?
Следующее
От: Tom Lane
Дата:
Сообщение: Re: whole-row functional index?