Re: [Bug-tar] Tar short read

[Kevin Fox]

On Tue, 2011-09-27 at 12:53 -0700, Paul Eggert wrote:
> On 09/27/11 12:18, Kevin Fox wrote:
> > No, what could happen is a user could stop tar mid stream, then continue
> > it, and a read might show up partial and tar would freak out.
> 
> In practice real systems don't do partial reads after a start/continue,
> when you're reading such small buffers from regular files.

Agreed. I think its unlikely to trigger the bug. That doesn't mean its
not a bug.

> As far as tar goes, it's low priority to fix this,

Thats ok. You can just call it a known bug until then. :)

> as the problem
> occurs only in file systems that are so incompatible with common
> behavior that they will break lots of applications.

I think things using the FILE* variants will work fine. Anything
following the GNU recommendations will also be fine. The ironic thing
seems to be that the GNU utilities look to be the hardest hit by not
following the GNU recommendation. :)

>   (This is true
> regardless of whether POSIX allows this file-system behavior.)

Sure. I definitely see the pragmatic, short term solution.

> It wouldn't hurt to change tar, I suppose, so long as the change
> doesn't overly hurt performance or clarity, but this is really more
> of a case where it's better to fix

Work around, not fix. ;)

>  the file system than to fix the
> many applications that the file-system behavior would break.

Agreed. I'll say again, I'll work around in my file system. I'd just
like to see the problem start to be addressed where appropriate instead
of continuing for every new file system to have to apply their own
workarounds to "fix" buggy programs. If things are started now, maybe in
the next decade or two it will be resolved. :)

Thanks,
Kevin