[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-tar] Year-2242 bug in GNU tar 1.15.1 (also, year-1969 bug)
From: |
Joerg Schilling |
Subject: |
Re: [Bug-tar] Year-2242 bug in GNU tar 1.15.1 (also, year-1969 bug) |
Date: |
Wed, 18 May 2005 01:04:57 +0200 |
User-agent: |
nail 11.2 8/15/04 |
Paul Eggert <address@hidden> wrote:
> Here's a problem I encountered when stress testing GNU tar 1.15.1
> on Solaris 8 (compiled with gcc -m64, using GCC 4.0.0):
>
> $ touch -d @-1 early
> $ touch -d @8589934592 late
> $ TZ=UTC0 ls -l --full-time early late
> -rw-rw-r-- 1 eggert faculty 0 1969-12-31 23:59:59.000000000 +0000 early
> -rw-rw-r-- 1 eggert faculty 0 2242-03-16 12:56:32.000000000 +0000 late
> $ tar -cf tar early late
> $ TZ=UTC0 tar -tvf tar
> -rw-rw-r-- eggert/faculty 0 1969-12-31 23:59:59 early
> -rw-rw-r-- eggert/faculty 0 2242-03-16 12:56:32 late
> $ tar -H posix -cf tar early late
> tar: value -1 out of time_t range 0..8589934591
> tar: value 8589934592 out of time_t range 0..8589934591
> tar: Error exit delayed from previous errors
The POSIX pax format allows any number of digits for time_t and
I would asume that negative time stamps are also possible.
> Unless I'm missing something, the error messages at the end seem
> incorrect. POSIX 1003.1-2001 allows for time stamps that exceed the
> 33-bit unsigned-int limit of traditional 'tar'.
Could you tell me where you believe to read that POSIX.1-1990 litits
the time stamp to unsigned int? Wouldn't it be possible that
33 bit two's complement may also meet the requirements?
Jörg
--
EMail:address@hidden (home) Jörg Schilling D-13353 Berlin
address@hidden (uni)
address@hidden (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily