[HACKERS] Missing SIZE_MAX

Поиск
Список
Период
Сортировка
Искать
От
Tom Lane
Тема
[HACKERS] Missing SIZE_MAX
Дата
Msg-id
15883.1504278595@sss.pgh.pa.us
Список
Дерево обсуждения
[HACKERS] Missing SIZE_MAX Tom Lane <tgl@sss.pgh.pa.us>
Re: [HACKERS] Missing SIZE_MAX Alvaro Herrera <alvherre@alvh.no-ip.org>
Re: [HACKERS] Missing SIZE_MAX Tom Lane <tgl@sss.pgh.pa.us>
Re: [HACKERS] Missing SIZE_MAX Robert Haas <robertmhaas@gmail.com>
Re: [HACKERS] Missing SIZE_MAX Tom Lane <tgl@sss.pgh.pa.us>
Re: [HACKERS] Missing SIZE_MAX Tom Lane <tgl@sss.pgh.pa.us>
Re: [HACKERS] Missing SIZE_MAX Robert Haas <robertmhaas@gmail.com>
Commit 2e70d6b5e added a dependency on SIZE_MAX to libpq/fe_exec.c.
According to C99 and recent POSIX, that symbol should be provided
by , but SUS v2 (POSIX 2001) doesn't require 
to exist at all ... and I now notice that gaur/pademelon doesn't
have it, and unsurprisingly is failing to compile fe_exec.c.

We have a workaround for that symbol in timezone/private.h:

#ifndef SIZE_MAX
#define SIZE_MAX ((size_t) -1)
#endif

and a bit of grepping finds other places that are using the (size_t) -1
trick explicitly.  So what I'm tempted to do is move the above stanza
into c.h.  Any objections?
		regards, tom lane


В списке pgsql-hackers по дате отправления
От: Greg Stark
Дата:
От: Tom Lane
Дата:
FAQ