[SSSD] [PATCH] Fix broken client in master

Jakub Hrozek jhrozek at redhat.com
Mon Mar 18 10:09:30 UTC 2013


On Mon, Mar 18, 2013 at 09:47:13AM +0100, Lukas Slebodnik wrote:
> On (16/03/13 15:46), Simo Sorce wrote:
> >On Sat, 2013-03-16 at 15:40 +0100, Lukas Slebodnik wrote:
> >> On (16/03/13 14:38), Lukas Slebodnik wrote:
> >> >On (15/03/13 21:35), Simo Sorce wrote:
> >> >>Fixes https://fedorahosted.org/sssd/ticket/1838
> >> >>
> >> >>Simo.
> >> >>
> >> >>-- 
> >> >>Simo Sorce * Red Hat, Inc * New York
> >> >
> >> >I tried to grep sources in src/sss_client directory
> >> >(with reverted commit 22d381367c27910fe82f476a76b9f4ede555e35a)
> >> >
> >> >$ grep -Rn '#include "util/util.h"'
> >> >sudo/sss_sudo.c:28:#include "util/util.h"
> >> >ssh/sss_ssh_knownhostsproxy.c:33:#include "util/util.h"
> >> >ssh/sss_ssh_client.c:33:#include "util/util.h"
> >> >ssh/sss_ssh_authorizedkeys.c:25:#include "util/util.h"
> >> >autofs/sss_autofs_private.h:22:#include "util/util.h"
> >> >autofs/autofs_test_client.c:31:#include "util/util.h"
> >> >
> >> >Is the difference among pam, nss and others clients?
> >> >
> >> >LS
> >> 
> >> Oh,
> >> I see the problem. But wouldn't be better solution to add another
> >> dependency (src/util/io.c) for libnss_sss instead of code duplication
> >> Because there is also another dependency to file src/util/murmurhash3.c,
> >> which is outside src/sss_client. Patch attached.
> >
> >Yes murmurhash is a necessary dependency due to the mmap code.
> >
> >As for the patch if you want to keep using common code we need to do 2
> >things:
> >
> >1) create a header file just for this file util/io.h for example.
> >Include this file in util/util.h and include just io.h in sss_client/
> >code.
> >
> >2) add a comment on top of io.c that this file is also used by
> >sss_client so it cannot include things like DEBUG macros and it should
> >not include util.h
> >
> >Simo.
> >
> >-- 
> >Simo Sorce * Red Hat, Inc * New York
> >
> Thak you very much for explanation in both mails.
> Updated patch attached.
> 
> LS

These patches fix the problem. I was wondering for a while if
introducing a new header file etc is worth saving the code duplication
of a couple of lines, but getting file IO right is important I think, so
we should be trying to reuse already proved patterns.

In other words, Ack.



More information about the sssd-devel mailing list