From: Tom Lane <tgl@sss.pgh.pa.us>
To: Robert Haas <robertmhaas@gmail.com>
cc: "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Subject: Re: constants for tar header offsets
In-reply-to: <CA+TgmoZniAFyu-UOt-OtbGnVnc00BhGbN6sMjH_tX+XNoj1a_w@mail.gmail.com>
References: <CA+TgmoZNbLwhmCrNtkJAvi8FLkwFdMeVU3myV2HQQpA5bvbRZg@mail.gmail.com> <2658038.1681832336@sss.pgh.pa.us> <CA+TgmoZniAFyu-UOt-OtbGnVnc00BhGbN6sMjH_tX+XNoj1a_w@mail.gmail.com>
Comments: In-reply-to Robert Haas <robertmhaas@gmail.com>
	message dated "Tue, 18 Apr 2023 11:53:08 -0400"
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-ID: <2661234.1681833969.1@sss.pgh.pa.us>
Content-Transfer-Encoding: 8bit
Date: Tue, 18 Apr 2023 12:06:09 -0400
Message-ID: <2661235.1681833969@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/2661235.1681833969%40sss.pgh.pa.us>
Precedence: bulk

Robert Haas <robertmhaas@gmail.com> writes:
> On Tue, Apr 18, 2023 at 11:38 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> 2. The header size is defined as 512 bytes, but this doesn't sum to 512:
>> +       TAR_OFFSET_PREFIX = 345         /* 155 byte string */

> I think that what happened is that whoever designed the original tar
> format decided on 512 byte blocks. And the header did not take up the
> whole block. The USTAR format is an extension of the original format
> which uses more of the block, but still not all of it.

Hmm, you're right: I checked the POSIX.1-2018 spec as well, and
it agrees that the prefix field is 155 bytes long.  Perhaps just
add another comment line indicating that 12 bytes remain unassigned?

			regards, tom lane