Re: [PATCH] Add an option not to use link(src, dest) && unlink(src) when that is unreliable

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Apr 23, 2009, 19:33 UTC
Message-ID: <alpine.DEB.1.00.0904232132380.10279@pacific.mpi-cbg.de>
In-Reply-To: <200904232116.10769.j6t@kdbg.org>

Hi,

On Thu, 23 Apr 2009, Johannes Sixt wrote:

Show 11 quoted lines

> On Donnerstag, 23. April 2009, Johannes Schindelin wrote:
> > It seems that accessing NTFS partitions with ufsd (at least on my EeePC)
> > has an unnerving bug: if you link() a file and unlink() it right away,
> > the target of the link() will have the correct size, but consist of NULs.
> >
> > It seems as if the calls are simply not serialized correctly, as
> > single-stepping through the function move_temp_to_file() works flawlessly.
> >
> > As ufsd is "Commertial software", I cannot fix it, and have to work
> 
> "commercial software"

I just quoted the license string of that wonderfully high-quality kernel module.

Maybe I should have added the beloved "[sic!]".

Show 17 quoted lines

> > At the same time, it seems that this fixes msysGit issues 222 and 229 to
> > assume that Windows cannot handle link() && unlink().
> >
> > Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
> ...
> > @@ -2225,7 +2225,9 @@ int move_temp_to_file(const char *tmpfile, const char
> > *filename) {
> >  	int ret = 0;
> >
> > -	if (link(tmpfile, filename))
> > +	if (unreliable_hardlinks)
> > +		ret = ~EEXIST;
> 
> It took me a while to see why we need a tilde here, but it's ok. Perhaps this 
> helps others:
> 
> +		ret = ~EEXIST;	/* anything but EEXIST */

Will do.

> Nevertheless:
> 
> Acked-by: Johannes Sixt <j6t@kdbg.org>

Thanks.

But it will have to wait for Saturday.

Ciao, Dscho