[PATCH] Treat symbolic links to directories like normal files when building tarballs

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

[PATCH] Treat symbolic links to directories like normal files when building tarballs

Samba - samba-technical mailing list
Treating symbolic links to directories as if they are directories causes
a couple of problems when building tarballs.  The attached patches fix
the the problems and appear not to break anything.  The commit messages
are quite long because it is somewhat difficult to explain the subtlety
in fewer words.   ;-)

The 1st commit fixes an actual bug.  However, I haven't opened a bug
because the problem is only in test code and is only seen when it is run
from a tarball, so definitely not in autobuild.  I'm happy to open a bug
and backport if people think this is important enough.

The 2nd commit fixes a theoretical problem that I hit in practice when
trying to work around the bug for the first commit by using
DIST_FILE() instead.  Thanks to Amitay for saying "don't do that,
fix the bug"...  :-)

I'm planning on using some more symlinks in the CTDB tests and the
first bug needs to be fixed for the new symlinks to work.

Please review and maybe push...

peace & happiness,
martin

samba.patch (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [PATCH] Treat symbolic links to directories like normal files when building tarballs

Samba - samba-technical mailing list
On Tue, Aug 8, 2017 at 9:47 PM, Martin Schwenke via samba-technical <
[hidden email]> wrote:

> Treating symbolic links to directories as if they are directories causes
> a couple of problems when building tarballs.  The attached patches fix
> the the problems and appear not to break anything.  The commit messages
> are quite long because it is somewhat difficult to explain the subtlety
> in fewer words.   ;-)
>
> The 1st commit fixes an actual bug.  However, I haven't opened a bug
> because the problem is only in test code and is only seen when it is run
> from a tarball, so definitely not in autobuild.  I'm happy to open a bug
> and backport if people think this is important enough.
>
> The 2nd commit fixes a theoretical problem that I hit in practice when
> trying to work around the bug for the first commit by using
> DIST_FILE() instead.  Thanks to Amitay for saying "don't do that,
> fix the bug"...  :-)
>
> I'm planning on using some more symlinks in the CTDB tests and the
> first bug needs to be fixed for the new symlinks to work.
>
> Please review and maybe push...
>
> peace & happiness,
> martin
>

Pushed to autobuild after fixing the comment in the first patch.

If there is diff output in comment, then it needs to be indented, otherwise
"git am" gets confused.

Amitay.
Loading...