[SSSD] [PATCHES] python bindings for managing users in local domains

Jakub Hrozek jhrozek at redhat.com
Mon Sep 14 20:51:03 UTC 2009


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 09/14/2009 08:53 PM, Simo Sorce wrote:
> this is the code I see for the function
> init_sss_tools()
> 
> +    octx = talloc_zero(tctx, struct ops_ctx);
> +    if (octx == NULL) {
> +        DEBUG(1, ("Could not allocate memory for data context\n"));
> +        ERROR("Out of memory\n");
> +        ret = ENOMEM;
> +        goto fini;
> +    }
> +    octx->ctx = tctx;
> +
> +    *_octx = octx;
> 
> This means that the 'octx' memory context is the child of 'tctx'.
> Needs to be the other way around otherwise if you free octx, tctx will
> not be automatically freed.
> 

Ahh, I know what happened. I fixed the behavior after the first review,
but squashed the fix into the second patch (see the changes in
tools_util.c), not the first one. And because I was looking at code my
HEAD, I saw only the fixed behavior..

Should I redo this? The patches will most probably be applied together,
but maybe it's better if every patch is self-contained..

	Jakub
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAkqurJ8ACgkQHsardTLnvCWUdQCg2z3+O2v5M1rrVKSdYY4WfH3+
r4UAoInwT3pZeljigRN5al0kF2QP+oSB
=mUza
-----END PGP SIGNATURE-----



More information about the sssd-devel mailing list