On Thu, Jun 05, 2014 at 07:23:48AM +0000, Mike Gabriel wrote:
> On Mi 04 Jun 2014 01:07:21 CEST, Adam Borowski wrote:
> >I'm afraid this patch doesn't work. It makes the code compile, but if you
> >try to execute it, it will assume any symlinks form a loop.
> >
> >The relevant snippet is:
> > if (++num_links > MAXSYMLINKS)
> > {
> > errno = ELOOP;
> > goto error;
> > }
> >which fails to execute if the returned value is -1. Linux' headers use an
> >arbitrary bogus value of MAXSYMLINKS 20 to let old code work, Hurd guys
> >decided that it's better for wrong code to fail at compilation stage.
> >It's the same story as MAX_PATH.
> >
> >sysconf(_SC_SYMLOOP_MAX) returns -1 on Linux as well.
>
> with so much background knowledge on this, do you think you can
> provide a patch for the patch?
>
> Should we simply set MAXSYMLINKS to this value of 20 instead?
I don't see much gain in trying to exhaust the system limit on pathological
symlink-to-symlink scenarios, so yeah, just setting this to 20 sounds a good
idea to me.
--
Gnome 3, Windows 8, Slashdot Beta, now Firefox Ribbon^WAustralis. WTF is going
on with replacing usable interfaces with tabletized ones?
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
