dash.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jilles Tjoelker <jilles@stack.nl>
To: Aleksandar Ristovski <aleksandar.ristovski@gmail.com>
Cc: dash@vger.kernel.org
Subject: Re: shell: dash - large file support
Date: Sun, 26 Jul 2015 21:21:41 +0200	[thread overview]
Message-ID: <20150726192141.GC4346@stack.nl> (raw)
In-Reply-To: <CAPaOzWwJkwozmDrk2BA5xHF+E6iM5mDrV9ORz=4RiFCdsKYbZA@mail.gmail.com>

On Fri, Jul 24, 2015 at 02:50:48PM -0400, Aleksandar Ristovski wrote:
> For builds that build 32-bit dash, configure misses to setup large
> file support resulting in issues with large files.

> For example:
> ...dp/dash-0.5.8/build$ ls -l /tmp/largefile.sh
> -rw-rw-r-- 1 aristovski aristovski 3794653588 Jul 24 14:12 /tmp/largefile.sh
> ...dp/dash-0.5.8/build$ ./src/dash /tmp/largefile.sh
> ./src/dash: 0: Can't open /tmp/largefile.sh

> Where dash was configured (on a 64-bit platform) like so:

> ...dp/dash-0.5.8/build$ CFLAGS=-m32 ../configure && make -j8

> Now, if I make the change in configure.ac and reconfigure the project,
> I get proper operation:

> ...0.5.8.patched/build$ ./src/dash /tmp/largefile.sh
> Running...
> ...

> (the actual working of the 'largefile.sh' is removed for brevity)

> Patch:

> --- dash-0.5.8/configure.ac     2014-09-28 04:19:32.000000000 -0400
> +++ dash-0.5.8.patched/configure.ac     2015-07-24 14:41:05.855055430 -0400
> @@ -4,6 +4,9 @@ AC_CONFIG_SRCDIR([src/main.c])
>  AC_CONFIG_HEADERS(config.h)
> +dnl On 32-bit builds, check for large file support
> +
>  dnl Checks for programs.

If I were the maintainer, I would use AC_SYS_LARGEFILE and remove all
complexity related to using largefile interfaces in some situations and
non-largefile interfaces in other situations.

However, the maintainer seems to prefer using largefile interfaces only
when needed, apparently to reduce the size of the binary. In the case of
large shell scripts, this is POSIXly correct (see XCU 1.5 Considerations
for Utilities in Support of Files of Arbitrary Size).

However, using a non-largefile stat() or lstat() is always wrong, since
the inode number may not fit in a non-largefile ino_t. This happens in
various cases: test -ef/-nt/-ot, finding a dot script (. special
builtin) and some related to the current directory.

On another note, test -nt/-ot should really compare the nanoseconds
parts as well.

Jilles Tjoelker

  reply	other threads:[~2015-07-26 19:21 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-07-24 18:50 shell: dash - large file support Aleksandar Ristovski
2015-07-26 19:21 ` Jilles Tjoelker [this message]
2015-07-28  2:25   ` Herbert Xu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150726192141.GC4346@stack.nl \
    --to=jilles@stack.nl \
    --cc=aleksandar.ristovski@gmail.com \
    --cc=dash@vger.kernel.org \


* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).